University of Southern California Dissertations and Theses

Reproducing inequity in organizations: gendered and racialized emotional labor in pubic organizations

(USC Thesis Other)

Reproducing inequity in organizations: gendered and racialized emotional labor in pubic organizations

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

REPRODUCING INEQUITY IN ORGANIZATIONS:
GENDERED AND RACIALIZED EMOTIONAL LABOR IN PUBLIC ORGANIZATIONS

By

Cynthia J. Barboza-Wilkes

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PUBLIC POLICY AND MANAGEMENT)

August 2022
ii

TABLE OF CONTENTS
LIST OF TABLES ..................................................................................................................................... V
LIST OF FIGURES ................................................................................................................................. VI
ABSTRACT ............................................................................................................................................ VII
INTRODUCTION ..................................................................................................................................... 1
CHAPTER I. HIDDEN EMOTIONAL BURDENS: MIXED-METHODS MEASUREMENT OF
UNACKNOWLEDGED INTERPERSONAL DYNAMICS ................................................................ 12
ABSTRACT ............................................................................................................................................................ 12
INTRODUCTION ................................................................................................................................................. 13
Emotional Labor as Emotion Regulation .................................................................................................. 15
Multilevel Empirical Challenges in the Study of Emotions ................................................................ 17
DATA, METHODS, AND MEASURES ............................................................................................................ 19
Data Collection ..................................................................................................................................................... 20
Diary Sample & Attrition .................................................................................................................................. 21
Timing of Data Collection ................................................................................................................................ 24
RESULTS ................................................................................................................................................................ 26
Qualitative Diary Prompts & Content Analysis ...................................................................................... 26
External Interactions With the Public ........................................................................................................ 29
Internal Interactions With Colleagues ....................................................................................................... 32
Statistical Modeling ............................................................................................................................................ 35
Measures .................................................................................................................................................................. 35
Model Specification ............................................................................................................................................. 36
Exploratory Semi-structured Exit Interviews ......................................................................................... 43
DISCUSSION ......................................................................................................................................................... 44
CONCLUSIONS/POLICY IMPLICATIONS .................................................................................................. 47
CHAPTER II. EMOTIONAL LABOR AT THE INTERSECTIONS OF IDENTITIES: AN
ASSESSMENT OF DIFFERENTIAL EMOTIONAL LABOR EXPERIENCES IN LOCAL PUBLIC
SERVICE ................................................................................................................................................. 50
ABSTRACT ............................................................................................................................................................ 50
INTRODUCTION ................................................................................................................................................. 50
Organizations as Gendered and Racialized Structures ...................................................................... 52
Emotional Implications of Gender and Racial Power Dynamics .................................................... 54
Differences in Black Men and Black Women’s Emotional Labor .................................................... 55
Research Questions & Hypotheses ................................................................................................................ 57
DATA, METHODS, AND MEASURES ............................................................................................................ 58
Diary Design ........................................................................................................................................................... 58
Dependent Variable—Emotional Labor .................................................................................................... 60
Independent Variables of Interest—Intersectional Identity ............................................................ 60
Control Variables ................................................................................................................................................. 60
QUANTITATIVE RESULTS .............................................................................................................................. 62
QUALITATIVE RESULTS ................................................................................................................................. 72
iii
Emotional Events ................................................................................................................................................. 74
Experiences of Negative & Positive Emotions ......................................................................................... 74
Emotional Labor .................................................................................................................................................. 75
DISCUSSION ......................................................................................................................................................... 82
Limitations & Future Directions ................................................................................................................... 86
CHAPTER III. THE EMOTIONAL TOLL OF COVID-19: A LONGITUDINAL STUDY OF
BURNOUT IN LOCAL GOVERNMENT AT THE ONSET OF THE PANDEMIC ........................ 89
ABSTRACT ............................................................................................................................................................ 89
INTRODUCTION ................................................................................................................................................. 89
EMOTIONAL LABOR & BURNOUT .............................................................................................................. 92
Crisis, Compassion, and Increased Emotional Labor ........................................................................... 94
The Dissonance–Burnout Relationship Through Conservation of Resources (COR) Theory
..................................................................................................................................................................................... 95
DATA, METHODS, AND MEASURES ............................................................................................................ 97
Quantitative Analysis ......................................................................................................................................... 99
DEPENDENT VARIABLE—BURNOUT .................................................................................................................. 100
PREDICTORS OF INTEREST—EMOTIONAL LABOR & INTERSECTIONAL IDENTITY ......................................... 101
ADDITIONAL CONTROL VARIABLES ..................................................................................................................... 102
RESULTS ............................................................................................................................................................. 104
Empirical Tests .................................................................................................................................................. 104
Qualitative Content Analysis ....................................................................................................................... 108
Reduced Personal Accomplishment & Job-Related Confidence .................................................... 110
Cynicism & Disengagement .......................................................................................................................... 112
Emotional Exhaustion .................................................................................................................................... 113
DISCUSSION/CONCLUSION ........................................................................................................................ 117
Recommendations ............................................................................................................................................ 119
Limitations & Future Directions ................................................................................................................ 121
CONCLUSION ...................................................................................................................................... 123
FUTURE DIRECTIONS ........................................................................................................................................... 125
BIBLIOGRAPHY ................................................................................................................................. 129
APPENDICES ...................................................................................................................................... 141
APPENDIX 1A: FULL DIARY PROTOCOL ............................................................................................................ 141
Survey Items and Factor Loading ............................................................................................................. 141
Cadence & Entries for Open-Ended Diary Prompts ........................................................................... 142
APPENDIX 1B: OLS MODELS PREDICTING ATTRITION .................................................................................. 144
APPENDIX 1C: PANDEMIC-SPECIFIC CONSIDERATIONS FOR COMPASSION, POSITIONALITY, AND
REFLEXIVITY ......................................................................................................................................................... 145
APPENDIX 1D: PRIMARY AND SECONDARY CODING SCHEME EXAMPLES ................................................... 146
APPENDIX 1E: FULL OLS MODELS FOR DISAGGREGATED FORMS OF EMOTIONAL LABOR ...................... 147
APPENDIX 1F: FULL ML MODELS FOR WITHIN-PERSON VARIATION IN EMOTIONAL LABOR ................ 149
APPENDIX 2A: OLS MODELS PREDICTING ATTRITION .................................................................................. 150
APPENDIX 2B: FULL DIARY PROTOCOL ............................................................................................................ 151
Survey Items and Factor Loadings ............................................................................................................ 151
iv
Cadence for Open-Ended Diary Prompts ................................................................................................ 152
APPENDIX 2C: OVERALL FREQUENCY OF EMOTIONAL LABOR & DIFFERENCES BY GROUP ..................... 153
Frequency of Emotional Labor Across All Participants ................................................................... 153
Chi-Square Tests for Differences in Emotional Labor Across Groups ........................................ 153
APPENDIX 2D: GROUP MEANS FOR GENUINELY EXPRESSED EMOTIONS .................................................... 154
APPENDIX 2E: FULL RESULTS OF SINGLE-AXIS VS. DOUBLE-AXIS MODELS OF EMOTIONAL LABOR . 155
APPENDIX 2F: FULL RESULTS OF DOUBLE-AXIS OLS MODELS OF DISCRETE EMOTIONS ................... 156
APPENDIX 2G: MLM MODELS PREDICTING SUPPRESSED EMOTIONS WITH KEY INTERACTIONS .......... 158
APPENDIX 2H: PRIMARY AND SECONDARY CODING SCHEME EXAMPLES ................................................... 161
Appendix 2I: Antecedents and Consequences of Emotional Events ............................................ 162
Antecedents of Emotional Events .............................................................................................................. 162
Consequences of Emotional Events ........................................................................................................... 168
APPENDIX 3A: FULL DIARY PROTOCOL ............................................................................................................ 171
Survey Items and Factor Loadings ............................................................................................................ 171
Cadence for Open-Ended Diary Prompts ................................................................................................ 173
APPENDIX 3B: EIGENVALUES FOR BURNOUT FACTORS OVER TIME ............................................................ 175
APPENDIX 3C: UNCONDITIONAL MODELS PREDICTING VARIATION IN BURNOUT .................................... 176
APPENDIX 3D: INTERSECTIONALITY & EL INTERACTIONS FOR BURNOUT ................................................. 177
APPENDIX 3E: PRIMARY AND SECONDARY CODING SCHEME EXAMPLES .................................................... 179

v

List of Tables

Table 1.1: Co-occurrence of Emotional Experience by Work-related Interaction 29
Table 1.2: Examples of Emotions Evoked by External Interactions 29
Table 1.3: OLS Models for Disaggregated Forms of Emotional Labor 37
Table 1.4: ML Models for Within-person Variation in Emotional Labor 42
Table 2.1: Single-Axis vs. Double-Axis OLS Models of Emotional Labor 66
Table 2.2: Double-Axis OLS Models Predicting Discrete Emotions 68
Table 2.3: Key Interactions in MLM Predicting Suppressed Emotions 70
Table 2.4: Key Terms for Establishing Secondary Codes 74
Table 2.5: Emotional Events 74
Table 2.6: Qualitative Themes by Group 82
Table 3.1: Coefficients for Intersectional Identity & Emotional Labor 106

vi
List of Figures

Figure 1.1: Multilevel Dynamics in Emotional Labor 7
Figure 1.2: Research Design 20
Figure 1.3: Demographic Group Response Rates (%) by Day 23
Figure 1.4: Diary Sample Distribution by Intersectional Identity 24
Figure 1.5: Interview Sample Distribution by Intersectional Identity
Figure 1.6: Qualitative Coding Scheme
26
28
Figure 2.1: Group Means for Amplified Emotions 63
Figure 2.2: Group Means for Suppressed Emotions 65
Figure 2.3: Percentage of Total Diary Entries by Intersectional Identity 73
Figure 3.1: Hypothesized Relationship between Crisis, Emotions, and Burnout 96
Figure 3.2: Hypothesized Relationships between Cultural Resources, Surface Acting,
and Burnout
97
Figure 3.3: Sequencing of Data Collection 102
Figure 3.4: Type of Burnout 109
Figure 3.5: Percent of Secondary Codes Applied to Burnout Dimensions 110

vii
Abstract
Emotional labor research in public administration lags behind other fields, is often
omitted from discussions of representative bureaucracy, and rarely looks at its gendered and
racialized dimensions. The existing scholarship fails to consider the dynamic nature of emotions
and that different emotions (e.g., happiness versus anger) might warrant different emotional
labor techniques for different groups. Meanwhile, scholars from sociology, applied psychology,
and organizational behavior widely recognize the importance of emotional labor, but few have
used an intersectional lens to study the well-recognized phenomenon.
This dissertation uses an intersectional approach to codify the difficult-to-measure and
often unobserved emotional labor that can institutionalize inequity within public organizations.
An intersectional approach is essential to make visible the experiences of those at the intersection
of multiple marginalized identities, and this dissertation describes in detail how the antecedents,
experiences, and consequences of emotional labor differ based on the employee’s combination of
gender and racial identity. Using a mixed-methods research design that combines daily diary
entries and semi-structured interviews, this work (1) describes and measures the emotional labor
embedded in both service encounters with the public and internal interactions among colleagues,
(2) looks at subgroup differences in the emotional effort at the intersection of race and gender,
and (3) assesses the relationship between emotional labor and burnout to inform our
understanding of the well-being of a diverse public sector workforce.
I find meaningful differences within and between individuals in the emotions needed to
effectively engage the public and navigate public institutions. The results reveal that, compared
to their peers, women of color engage in more taxing forms of emotional labor, feel more
emotionally constrained by organizational rules, are more cognizant of managing gendered and
viii
racialized stereotypes, and are more sensitive to whether the climate allows for authentic
expression. I also show that public employees experienced heightened burnout during the
pandemic, and the suppression of emotion contributed to that burnout, but in different ways for
different groups. In particular, women of color who suppressed negative emotions were more
likely to experience a reduced sense of personal accomplishment, increased cynicism,
disengagement from their work, and more emotional exhaustion.
This project reveals important distinctions in the type of emotional labor demanded of
public employees and how those emotional demands differ across gender and racial identities.
The results make visible the experiences of those at the margins of multiple lived experiences of
oppression, allowing women of color to articulate their own emotional experiences in ways that
center their voices. Importantly, this work highlights the importance of factoring emotional labor
into the experience of burnout at work while emphasizing that the relationship between the two
varies for individuals of different backgrounds. I provide concrete proof that there is an uneven
distribution of emotional labor in public organizations, and it falls predominantly on women of
color.
Measuring a construct as complex and dynamic as emotional labor lays the groundwork
for important reform. By codifying, measuring, and describing the differential emotional burdens
embedded in public organizations, I quantitatively demonstrate the need for equitable human
resource management practices that address how organizations structurally reinforce inequity.
1

Introduction
Problem
Representation in local government matters, but women and people of color are
underrepresented, especially in leadership and management. The murder of George Floyd and
devastation in communities of color from the pandemic have highlighted the longstanding
problems in local public services. Public services do not serve communities equally. Among the
many reasons for these disparities is the challenge of advocating effectively for underserved
communities from within public bureaucracies. Local governments struggle to recruit and retain
minorities, particularly women of color. This project therefore addresses an essential yet
unexplored question: What prevents women of color from thriving in local governments? One
explanation for problems with recruitment and retention may be that women of color have such
high emotional labor burdens on the job. Emotional labor includes coping with prejudice and
discrimination within the workplace, educating White colleagues on the lived experience of
systemic oppression, and being one of few contacts between public agencies and communities of
color desperate to have someone understand their circumstances. Uncompensated emotional
work exhausts the emotional reserves of the people doing it. Despite the clear consequences of
emotional labor, it is understudied in public administration.
This dissertation integrates feminist theory, critical race theory, organizational behavior,
and emotional labor scholarship to rigorously measure and detail the disproportionate burdens in
relational work for women of color in the public sector workforce. This work is designed to
unpack how we can foster sustained engagement and minimize burnout among multiply
marginalized members of public organizations, particularly those engaged in active
2
representation. I use active representation to be inclusive of advocacy work done by minoritized
individuals within the public sector workforce on behalf of minoritized constituents through a
representative bureaucracy lens. I pursue a deeper understanding of engagement and burnout in
an increasingly diverse public sector by systematically exploring the use of emotions as a social
and psychological resource on the job.
Scholarly Status Quo & Highlighting the Gap
Diversity in the public sector workforce is often described as an asset, given that the job
involves understanding the diverse needs and values of the communities served. Representative
bureaucracy scholarship makes the case that a government workforce that demographically
reflects the constituents it aims to serve will be more democratic and produce more equitable
outcomes (Ding et al., 2021). Recent meta-analytic work shows that more representative
organizations can improve performance, but increasing the number of women and BIPOC in the
workplace is not sufficient to enhance equity and fairness within the organization (Ding et al.,
2021; Hoang et al., 2022). It is not enough to focus on diversity efforts in hiring. We also need to
look at the implicit burdens associated with being a member of an underrepresented group and
how those burdens impact an employee’s ability to advocate for their community from within the
organization. We know little about how efforts to diversify the workforce have impacted the
internal culture, norms, and implicit expectations within public organizations. Public
administration scholars are calling for more research that explicitly acknowledges how
Whiteness and masculinity shape our understanding of public organizations (Portillo et al.,
2022), and this dissertation aims to measure systemic inequities in public organizations to lay the
foundation for building more equitable workplaces moving forward.
3
Humphrey (2021b) argued that more research is needed to unpack the workplace
challenges experienced by public servants of color, who may experience conflicting demands
and pressures from organizational socialization and their racial identity. On top of standard job
demands, employees of color are expected to address racially charged issues on behalf of their
organizations, creating expectations for racialized emotional labor (Humphrey, 2021b).
Emotional labor involves regulating the emotions experienced and expressed on the job to
conform to organizational norms and expectations (Hochschild, 1983). Prior emotional labor
research has predominantly explored service jobs in the private sector, with less attention paid to
the nuances of public sector work (Humphrey, 2021a). The need to negotiate shared public
values and incorporate the (often conflicting) goals of diverse constituents presents novel
emotional challenges internally and externally for public employees.
When emotions are studied, scholarship has largely focused on gender disparities, with
less attention paid to racial disparities (Wingfield, 2010). Meanwhile, the existing research on
racial stratification in organizations does not spend enough time explicitly defining or discussing
emotional labor (Wilkins & Pace, 2014). Scholars from sociology, applied psychology, and
organizational behavior widely recognize the importance of emotional labor, but few have used
an intersectional lens to study the well-recognized phenomenon. Dealing with diversity
constitutes a source of affective events that often involve some amount of emotional labor, and
understanding the interactions between these topics helps inform how organizations can create
positive and healthy environments (Ashkanasy et al., 2002). Emotions (e.g., anger, resentment,
despair, sympathy, etc.) are central to experiences of injustice and inequity. Emotions scholars
offer that negative experiences (via race or gender) cause negative emotions, and those who are
disadvantaged have fewer material, social, and psychological resources with which to cope
4
effectively with such experiences, but these theories do not adequately address why gender and
racial disadvantage do not seem to impact emotions in the same ways (Grandey et al., 2013;
Wilkins & Pace, 2014; Wingfield, 2010).
There are several key opportunities to improve the methodological and theoretical rigor
of the study of emotional labor in public sector organizations. Methodologically, Guy et al.
(2008) noted that a lot of the interpersonal skills necessary in local government have not been
quantified, so they have not made it into job descriptions or performance reviews, nor are they
considered as factors justifying compensation. Emotional labor in public organizations is under-
measured and not adequately grappled with in all of its complexity. The existing scholarship on
emotional labor in public administration research largely uses cross-sectional surveys and/or
interviews to understand whether or not individuals are suppressing or amplifying emotions. This
approach fails to account for the dynamic nature of emotions and the fact that different emotions
(e.g., happiness vs. anger) might warrant different emotional labor techniques for different
groups. By integrating best practices for the measurement of emotional labor from sociology,
applied psychology, and organizational behavior, I help bridge this methodological gap and bring
much-needed attention to the study of emotional labor in public organizations.
Theoretically, there are two opportunities to expand the existing work on emotional labor.
First, scholarship to date has largely looked at emotional labor in service encounters, with less
attention paid to the internal workplace dynamics that also require emotional effort. Second,
while past work has looked at gendered and (to a lesser extent) racialized emotional labor, an
intersectional lens is necessary to appropriately theorize the inequitable distribution of emotional
labor in public organizations. Using feminist and critical race methodologies that center the
voices of understudied groups, I explore the impact of differences in race, gender, and power on
5
emotional experiences among the public workforce, with a particular emphasis on the
relationship between emotional labor and burnout. Taken together, the goals of this project are to
leverage a mixed-methods research design to (1) measure and describe the nature of emotional
labor within public organizations, (2) reveal how that labor differs based on individual identity,
and (3) outline the well-being outcomes of that emotional labor for different groups of
employees.
My research adds intersectional layers to Wingfield and Alston’s (2014) theory of racial
tasks by empirically interrogating the differential emotional burdens experienced by women of
color in the workforce. Expanding upon racialization, I specifically highlight how the
simultaneous experience of multiple systems of oppression demands unacknowledged emotional
labor of minoritized groups. Identifying differential emotional burdens informs how more
equitable human resource management policies and practices can be developed and implemented
to ensure organizations proactively address how they structurally reinforce inequality.
Argument
Measuring and describing the nature of emotional labor within public organizations helps
to broaden the existing emotions scholarship by shifting the focus to a new context—the public
sector. Within public administration and representative bureaucracy research, this work is the
first to explore how different sets of discrete emotions are expressed and suppressed by distinct
groups of employees as a tool to manage the way they are perceived by others. Doing so
provides a more nuanced measurement and understanding of the inequitable distribution of
emotional labor across the public workforce. Applying an intersectional lens to the study of
emotional labor also expands previous work by effectively theorizing nuanced differences in
gendered and racialized emotional expectations on the job. Importantly, measuring the
6
inequitable distribution of emotional labor across demographic groups in the workforce sheds
light on important outcomes for employee engagement and well-being.
The link between emotional labor and burnout has been well established, but less is
known about how social identity might complicate the relationship between the two. In this
dissertation, I argue the study of emotional labor and burnout in public organizations must adopt
an intersectional framework rather than unidimensional classifications to account for the fact that
identity is a multidimensional construct that manifests in nuanced power dynamics in
organizations. The emotional burdens of engaging in advocacy work from within public
organizations are compounded in their complexity when the intersection of gender, race, and
other markers of difference force advocates to code-switch across different audiences and
stakeholders to bring about change. Accounting for the emotional toll this takes is consequential
for the retention of a diverse workforce. Accounting for the intersection of racialized and
gendered norms for specific emotions allows for a more nuanced understanding of who is unduly
burdened by emotional demands on the job. Thus, this work helps organizations understand the
equity implications embedded in the way diverse teams function as the public sector becomes
increasingly diverse, collaborative, and interdependent. Using intersectionality as the research
paradigm helps to sharpen theorization on the relationship between emotional labor and burnout,
which inevitably will help scholars and practitioners develop more culturally responsive human
resource management practices.
Approach
Studying emotional labor effectively requires a longitudinal approach with multilevel
theorization to understand dynamics at the person, event, and organization levels (Ashkanasy et
al., 2002; Grandey & Melloy, 2017; Imose & Finkelstein, 2019). At the person level, social
7
status, values, traits, and abilities are variables that influence the emotion regulation processes
and their outcomes (Grandey & Melloy, 2017). At the event level, the outcome of one emotional
event becomes the employee’s next emotional event, creating a feedback loop (Côté, 2005; Côté
et al., 2013). At the organization level, the characteristics of interpersonal interactions and
relationships as well as the overall work environment are influential in selecting emotional labor
strategies and subsequent outcomes (Grandey & Melloy, 2017; Imose & Finkelstein, 2019).
Relationship dynamics (e.g., power asymmetry and warmth between people) contribute to
within-person variations in emotional expectations, meaning an individual’s emotional
experience will vary based on both event-level feelings and the person they are engaging with
(Grandey & Melloy, 2017). As shown in Figure 1.1, there are variables and dynamics at each
level that inform the emotional labor required and how that impacts subsequent interactions.
Figure 1.1: Multilevel Dynamics in Emotional Labor

To account for temporal and multilevel complexity, I adopt a mixed-methods approach
combining surveys, interviews, and diary entries, which allows me to find generalizable trends
8
and patterns from longitudinal data while also contextualizing the numbers with in-depth
qualitative findings that center the voices of participants in their natural work environment. The
integration of these methods is particularly helpful in addressing the dynamic and highly
contextual nature of emotions. The longitudinal measurement of emotional labor through
multiple modalities allows for a within-person approach to studying variation over time while
controlling for individual differences, which has not previously been done in the context of
public organizations. When examining behaviors, diaries account for the temporal sequencing of
events and can control for omitted variables by using participants as their own controls and
studying them in their natural setting (Bolger et al., 2003). Experiences of emotional labor may
be transient, differing widely between days and depending on situations, and through daily
diaries, the dynamic nature of emotion regulation can be studied much more precisely than is
possible using cross-sectional, static methods. Facilitating semi-structured interviews after diary
data collection allows for a more in-depth exploration of the themes surfaced in the diaries.
These interviews serve the function of both expanding on existing findings and checking for
understanding between myself and the participants. Together, these approaches can support
robust statistical modeling alongside rich qualitative content analysis, helping public
administration scholars see concretely how much additional emotional effort is required of
women, people of color, and especially women of color in the public sector workforce.
Chapter Structure
Public administration needs more focus on emotional labor. Moreover, we need to be
thinking about emotional labor beyond the service context to explore how internal organizational
dynamics create additional interpersonal demands. In Chapter 1, I describe the nature of
emotional labor within public organizations by comparing the emotional norms and effort
9
associated with service encounters with residents to those associated with internal interactions
among colleagues. This chapter provides the most nuanced and robust measurement of emotional
labor in public organizations to date and adds to the emotional labor literature by comparing and
contrasting the type of emotional labor required to engage in service encounters with the public
versus internal team dynamics.
Thinking about internal power dynamics requires a theoretical lens that extends beyond
an analysis of gender or race as separate sources of emotional labor and instead examining how
these simultaneous identities mutually constitute different sets of emotional expectations.
Chapter 2 describes how emotional labor differs based on individual identity, showcasing how
women of color in particular are tasked with unique gendered and racialized emotional burdens
on the job. While past work by Wingfield (2010) has qualitatively explored the emotional labor
experiences of women of color in organizations, this chapter is the first to take a quantitative
approach to measure the disproportionate amount of emotional labor women of color take on
relative to White women, men of color, and White men.
Lastly, while existing research suggests some emotional labor techniques are costly and
contribute to burnout, we know little about whether some groups are more susceptible to burnout
due to engaging in different forms of emotional labor. Chapter 3 focuses on the outcomes of
emotional labor for the individual employee, looking at how the relationship between emotional
labor and burnout changes for different demographic groups. With data collected before and
during the COVID-19 pandemic, this chapter is also able to shed light on the emotional demands
of public service work during times of crisis and uncertainty. Importantly, this chapter takes an
intersectional approach to explore the emotional labor embedded in crisis response work, an
understudied context where inequities abound. Taken together, these chapters commit to a more
10
robust measurement of emotional labor through the theoretical lens of intersectionality, which
gives organizational scholars much deeper insight into the emotional burdens inequitably
distributed across employees.
Importance
For public administration scholars, this research expands what we know about the
emotional labor embedded in public service work. For scholars in related fields where emotional
labor scholarship is further along, the intersectional approach adds much-needed theoretical and
methodological rigor to the way we think about the distribution of emotional labor. Emotional
labor is a concept that many of us intuitively understand, but this dissertation quantitatively and
qualitatively proves the emotional effort we feel is very real. With the data presented here, we no
longer have to speculate about whether emotional labor is real, costly, or contributes to burnout.
Nor should we speculate whether that labor is shared equally. Importantly, my findings show that
the toll of emotional labor is not evenly distributed. In essence, this project does the legwork of
measuring a construct that is difficult to measure and analyzing it in a way that is sensitive to
power asymmetry in organizations by showcasing how it manifests differently across groups.
The insights from this dissertation are designed as a first step toward creating better
experiences for the employees who serve the public. This research has direct implications for all
facets of human resource management: recruitment, selection, compensation, training,
development, and retention (Guy et al., 2008). I show in concrete terms that emotional labor
takes different forms based on the intersection of gender and racial identity, and I highlight how
different groups choose which emotions to display based on a combination of job demands,
implicit organizational norms, and identity-based stereotypes. With a better understanding of the
inequitable division of emotional labor, we can potentially build culturally competent strategies
11
to foster sustained engagement and minimize burnout to support employees who are also
members of minoritized communities as they advocate for change. Academically, this research
agenda contributes to the fields of public administration, organizational behavior, and human
resource management by highlighting the emotional dimensions of representative bureaucracy.
12
CHAPTER I. Hidden Emotional Burdens: Mixed-Methods Measurement of
Unacknowledged Interpersonal Dynamics
ABSTRACT
Many of the interpersonal skills necessary in local government are difficult to quantify,
so they don’t make it into job descriptions or performance reviews, nor are they considered as
factors justifying compensation. The inadequate measurement of key relational skills makes it
difficult to know if emotional burdens are falling on some employees more than others across
contexts. This paper codifies the difficult-to-measure and often unobserved emotional labor that
can institutionalize inequity within organizations. Using a mixed-methods research design that
combines daily diary entries and semi-structured interviews, this paper describes and measures
the emotional labor embedded in both service encounters with the public and internal
interactions among colleagues. Content analysis and multilevel modeling reveal meaningful
differences within and between individuals in the emotions needed to effectively engage the
public and navigate public institutions. Insights from the content analysis show that internal
interactions among colleagues seem to be more prevalent examples of emotional events in the
memories of participants. By contrast, statistical modeling shows that only external interactions
with members of the public are predictive of both faking positive and suppressing negative
emotions on the job. This study demonstrates that the integration of multiple methods creates
insights that are more comprehensive than what would have been gained using only separate
methods. Substantively, the paper also reveals important distinctions in the type of emotional
labor demanded in interactions between colleagues compared to public engagement.
13
INTRODUCTION
Whether it involves “dealing with a senior who was rude and condescending when all I
did was call and check in on her, just as she had asked me to do last week” (4:718) or being
“happy to see one of the residents who I checked in last night looking better” (5:2018),
interactions with members of the public often evoke emotional responses among local
government employees. Public management scholars have historically emphasized technical and
managerial competencies, with less attention paid to interpersonal and emotional skills at work
(Awasthi & Mastracci, 2021; Humphrey, 2021). Emerging research suggests that emotions and
professionalism in local public service are interrelated, meaning government employees must be
skilled in emotional labor to be professional (Humphrey, 2021). Emotions are core to
interpersonal relationships. In public organizations, the ability to regulate emotions is key to
connecting with the community, building and maintaining collaborative relationships with
colleagues, and contributing to organizational objectives.
Hochschild (1983) first introduced “emotional labor” to describe how workers manage
feelings on the job. Emotional labor refers to the effort it takes to display the emotions that align
with the norms and requirements of the job. More specifically, Guy et al. (2008) described
emotional labor as “work which requires the engagement, suppression, and/or evocation of the
worker’s emotions in order to get the job done” and suggested it is systematically underestimated
in most public organizations (p. xii). Social workers, schoolteachers, emergency call-takers, and
other frontline public servants often must manage their own emotions while simultaneously
managing the emotions of the person(s) seeking service (Lu & Guy, 2019).
The impact emotional labor has on the well-being of the employee depends on how it is
done. Hochschild (1983) identified two strategies by which employees can achieve these
14
emotional requirements: surface acting, where employees mask their feelings and display the
expected expressions, creating dissonance between feelings and expressions, and deep acting,
where employees alter their internal feelings to appear more genuine in their performance.
Feminist organizational theorists have argued that authentic expression of emotions at work
supports the psychological well-being of organizational members (Grandey et al., 2013; Martin,
2000; Martin et al., 1998). Surface acting definitionally suggests employees’ emotional
expressions are inauthentic, and empirical studies have shown the dissonance between what is
felt and what is expressed has harmful consequences for the individual (Brotheridge & Lee,
2002; Hülsheger & Schewe, 2011).
Current scholarship tells us we should expect external interactions with the public to be
emotionally taxing, but public administration scholarship to date largely ignores the reality that it
can also be emotionally effortful for employees to manage relationships with their coworkers,
staff, and supervisors. Examining how emotional labor manifests in both external encounters
with the public and internal team dynamics answers the call among emotional labor scholars to
expand beyond the service employee context to non-service jobs and interactions among
colleagues, painting a much more complete picture of the overall employee experience (Grandey
et al., 2013; Humphrey et al., 2015; Wang et al., 2011).
This manuscript provides a literature review of emotional labor as a form of emotion
regulation and the multilevel empirical challenges associated with studying emotions. I then
describe the data, methods, and measures designed to address the theoretical and empirical gaps
in public administration scholarship. Next, I review the results, which demonstrated that
emotional labor differed significantly by context. Qualitatively, internal interactions among
colleagues were more salient in the minds of respondents in interviews and open text responses,
15
but statistically, external interactions with members of the public (residents) were more
commonly associated with emotional labor, particularly surface acting. I discuss the implications
for practice and future research and specify the limitations of this study.
Emotional Labor as Emotion Regulation
Over the past 20 years, the “affective revolution” in organizational behavior research has
led to an increase in empirical work that looks at emotional labor as emotion regulation,
including several meta-analyses (Grandey, 2000; Grandey & Gabriel, 2015; Hülsheger &
Schewe, 2011; Kammeyer-Mueller et al., 2013). Emotion regulation (ER) is defined by Gross
(1998) as an individual process for influencing which emotions are felt, when they are felt, how
they are experienced, and how they are expressed. Antecedent-focused
1
strategies are used to
change a situation or modify feelings, and response-focused strategies are used to change
expressions and behavior after emotion is felt. Grandey (2000) argued that Hochschild’s (1983)
deep and surface acting concepts can be mapped onto antecedent- and response-focused
strategies, respectively. Grandey (2000) linked deep acting to the antecedent-focused strategy of
cognitive change (or reappraisal), which involves altering the meaning one makes of a situation
in a way that influences the emotions that the situation will produce. As an example, you might
interact with someone rude or condescending, but rather than taking it personally, you reframe
the interaction in your mind by telling yourself they must be having a bad day, then move on
without experiencing much negative emotion.
Response-focused strategies involve changing one or more of the experiential, behavioral,
or physiological components of an emotional response (Grandey, 2000; Gross, 1998). As an

1
Antecedent-focused emotion regulation strategies include (1) situation selection, (2) situation
modification, (3) attention deployment, and (4) cognitive change or reappraisal (Grandey, 2000;
Gross, 1998).
16
example, a coworker may have made an insensitive comment about you in a meeting, but rather
than showing you were hurt, you smiled and laughed it off to avoid making a scene. Meta-
analytic work has shown inauthenticity experienced during response-focused surface acting is
negatively associated with feelings of personal accomplishment and positively associated with
emotional exhaustion, psychological strain, psychosomatic complaints, depressed mood, stress,
burnout, and job dissatisfaction (Grandey & Gabriel, 2015; Hülsheger & Schewe, 2011;
Kammeyer-Mueller et al., 2013). Among response-focused emotion regulation techniques,
suppression of emotion has been shown to require more attentional monitoring from the actor to
ensure feelings do not slip out and is believed to increase feelings of inauthenticity and
sympathetic activation (Grandey & Melloy, 2017; Grandey et al., 2013; Gross, 1998). By
contrast, reappraisal strategies, while more effortful than no regulation at all, ultimately led to
few, if any, cognitive and social costs (Grandey et al., 2013). This suggests that modifying a
situation or our thoughts about it (deep acting) may be preferable to suppressing or faking
emotions (surface acting) as a route to psychological well-being (Grandey, 2000; Gross, 1998).
Emotional labor scholarship in public administration has focused primarily on street-level
bureaucrats and the customer service context, with more limited exploration of emotional labor
dynamics within teams.
2
Street-level bureaucrats interact directly with the public and are
typically under large caseloads, forced to navigate ambiguous agency goals with inadequate
resources (Lipsky, 2010). Part of the challenge for frontline employees is managing sometimes
conflicting sets of obligations to and expectations from the public, their peers, the organization,
and society. These employees are often required to interpret policy on a case-by-case basis but

2
Throughout this paper, service encounters with members of the public will be called external
interactions, and encounters with colleagues within the organization will be called internal
interactions.
17
are constrained by internal rules and processes that can frustrate both the employee and the
individuals seeking service, creating tension and potential for inequitable service outcomes based
on ad-hoc policy adaptations (Lipsky, 2010). But focusing exclusively on the external
interactions of street-level bureaucrats provides little insight into the emotional labor experienced
within the organization across employees in a range of roles that vary in terms of their
responsibilities to engage the public directly.
This paper first asks, (RQ1) is there a difference in the way employees describe
emotional labor in service encounters with residents versus internal interactions with colleagues
within the organization? Importantly, Leiter and Maslach (2000) argued that interactions internal
to the organization, among coworkers, are important sources of job stress and burnout. While we
can anticipate that interactions with the public are likely to be different from interactions with
subordinates, peers, and supervisors, to date we know little about the extent to which internal
versus external interactions drive the majority of emotional labor. Thus, the second research
question asks (RQ2) whether it is the frequency of (a) internal interactions among colleagues or
(b) external interactions with the public that is most tightly correlated with surface acting
behaviors.
Multilevel Empirical Challenges in the Study of Emotions
The antecedents of emotional labor are complex and dynamic. Grandey and Melloy’s
(2017) multilevel model of emotional labor as emotion regulation integrates feedback loops at
the event, person, relational, and organizational levels. At the event level, the difficult
conversation we have with our boss might influence the way we behave in the team meeting that
follows. At the person level, we each come to the table with different social status, values,
personalities, and abilities that guide our behaviors across contexts. At the relational level, we
18
think about our own individual differences and how those stack up with the characteristics of the
person we interact with, meaning each person we interact with might come with a different set of
expectations for how we should behave. Work groups may have people of different backgrounds,
who hold different social identities, have different life experiences, and hold different status
within the organizational hierarchy, all of which influences how we might interact with them. At
the organization level, different work teams and professions will also have norms for how to
behave, which will influence individual decisions to regulate emotions.
With emotion regulation conceptualized as a dynamic process subject to influences at the
event, person, relational, and organization levels, there are several methodological challenges.
Existing scholarship on emotional labor in public organizations has predominantly relied upon
cross-sectional data from surveys, with more recent calls to expand the methodological range in
the study of emotions (Côté, 2005; Grandey et al., 2013). Because emotions and affective states
change rapidly from moment to moment, cross-sectional data is often inadequate for the study of
antecedents and consequences of emotional labor. The complex, transient, and dynamic nature of
emotion research also necessitates measures for capturing the less conscious and more
momentary nature of emotion regulation while situating the research within the social-
organizational context (Grandey & Melloy, 2017).
Many emotional labor scholars advocate for some combination of longitudinal in-situ
methods such as the experience sampling method (ESM) to capture emotional labor and its
effects as they unfold over time along with interviews to unpack the meaning people make of
their emotional experiences (Ashkanasy, 2003; Bono & Vey, 2005; Diefendorff et al., 2021;
Grandey & Gabriel, 2015; Grandey et al., 2013). Additionally, collecting data in a naturalistic
19
setting incorporates the organizational context in which the experience and regulation of emotion
take place.
DATA, METHODS, AND MEASURES
Substantively, this paper explores (RQ1) whether or not emotional labor manifests
differently in external versus internal interactions, and (RQ2) if the frequency of internal versus
external interactions influences surface acting behaviors. I address the first question by using
qualitative content analysis of open-ended diary entries to differentiate the experience of
emotional labor across interaction types. I address the second question by presenting a series of
OLS regressions using data that averages responses across a 14-day panel of diary data to
uncover the variables relevant to predicting emotional labor. I then introduce multilevel
modeling to look at between- and within-person differences in emotional labor over the 3-week
span of the study. In these multilevel analyses, observations on the day (or event) level are nested
within persons, meaning daily observations constitute level-1 data, and the stable person or
situation characteristics constitute level-2 data (Bolger et al., 2003; Myin-Germeys et al., 2021;
Ohly et al., 2010). A content analysis of qualitative exit interviews is then used to add depth and
context to our findings from the diaries (see Figure 1.2).

20
Figure 1.2: Research Design

Data Collection
The data for this study combines qualitative and quantitative diary prompts with semi-
structured interviews. Experience sampling methods, including diary designs, are considered
effective research modalities for studying self-regulation at work (Johnson et al., 2018; Koopman
et al., 2020; Myin-Germeys et al., 2021). In diary designs, participants respond to repeated
assessments at moments over the course of time while functioning in their natural settings
(Scollon et al., 2003). Diary methods are ideal for studying thoughts, feelings, and behaviors
within the natural work context as well as characteristics of the work situation, which may
fluctuate on a daily basis.
In a complementary approach, the day reconstruction method (DRM; Kahneman et al.,
2004) is designed to capture the positive and negative feelings that accompany daily activities by
facilitating accurate emotional recall (Bakker & Oerlemans, 2011). In the DRM, respondents are
asked to describe each episode of the day by including details of timing, tasks, location, and
21
participants to draw on episodic memory and retrieve information about the pleasure and
intensity of emotions that occurred at the time (Bakker & Oerlemans, 2011). The DRM inspires
the prompts selected within the diary approach (see Appendix 1A for full diary protocol).
For the data to be reliable, and inferences made from the data valid, diary studies need a
high level of participant commitment and dedication, which often requires detailed training to
ensure that participants fully understand the protocol (Bolger et al., 2003; Myin-Germeys et al.,
2021). Because of the additional effort required of diary participants compared to cross-sectional
surveys and other lower-touch longitudinal designs, Ohly et al. (2010) argued that “special
attention has to be paid to recruitment of participants and anticipation of attrition over the course
of data collection” (p. 85). To minimize these burdens, the daily instruments were designed to be
short, only taking 5 to 7 minutes to complete.
Diary Sample & Attrition
In their review, Ohly et al. (2010) found previous diary studies in high-ranking journals
had sampled at least 100 persons, focusing on predictors at the person level, and at least 5 days
per person, focusing on predictors at the day level. Roughly 20% attrition is common in diary
studies (Ohly et al., 2010), so I initially aimed to recruit a sample larger than 100 as a buffer
against dropouts. Recruitment outreach included a detailed explanation of the aims of the study
and the utility of responding accurately, and participants were personally emailed daily to
provide the next survey link and enhance participants’ feeling of involvement.
Participants were recruited from a list of local government employees across Los Angeles
County. Baseline participant surveys captured demographic data, attitudinal variables (e.g.,
prosocial motivation, attitudes towards the public), and job characteristics (e.g., autonomy,
discretion, perceived red tape, perceived fairness, frequency and type of citizen engagement). An
22
additional onboarding survey was used to introduce participants to the research project,
providing instructions and guidance to set expectations for the diaries, while also collecting some
additional perceptions of workload and workplace climate.
In total, 181 employees from over 10 distinct municipal governments
3
spanning 39
departments in Los Angeles County filled in at least one of the daily diary questionnaires,
resulting in a total of 1,563 completed daily surveys. Only two members of the sample were
elected officials, and 10 were appointed, meaning the vast majority of the sample represents
careerists who were hired into their positions. Not all respondents filled in the questionnaire
every day; on average, the response rate of the daily questionnaire varied between 38% and 77%
depending on the day, and Figure 1.3 illustrates daily response rates as a percentage broken
down by demographic groups.
4
These percentages are consistent with previous scholarship using
a diary method, in which typically between 40 and 120 respondents are included, filling in daily
questionnaires for a period of 5 to 10 days (Ohly et al., 2010). Systematic attrition by race and
gender was ruled out using a series of logistic regression models (see Appendix 1B).

3
Atwater, Bev Hills, Calabasas, Gardena, Glendale, Lakewood, Long Beach, Los Angeles, Malibu,
Norwalk.
4
Note that the lowest response rates occurred on Fridays (i.e., Day 5 and Day 10).
23
Figure 1.3: Demographic Group Response Rates (%) by Day

The sample consisted of 94 (51.93%) women and 87 (48.07%) men, which is fairly
consistent with the general population for Los Angeles County (51% women and 49% men).
While it was not possible to obtain the demographic breakdown of the workforces for all 10
municipalities included in the sample, based on the workforce of the City of Los Angeles, which
accounts for the highest number of study participants, women were over-sampled (only 28% of
LA City workforce). In terms of race and ethnicity, the sample was 52% White and 48%
24
employees of color, which is an under-sampling of people of color relative to the City of Los
Angeles workforce and the Los Angeles County population generally. Among the employees of
color, 44.8% identified as Mixed/Other, 23% Hispanic or Latinx, 14.9% Black, 10.3% Asian,
3.4% Native American or Alaskan, and 3.4% Pacific Islander or Native Hawaiian. As points of
comparison, the City of LA workforce is 38% Hispanic, 29% White, 16% Black, 11% Asian, 5%
Filipino, and <1% American Indian, with no data on those who identify as Mixed or Other, and
the general population of Los Angeles County is 48% Hispanic, 26% White, 15% Asian, 8%
Black, and 3% two or more races/ethnicities (U.S. Census Bureau, 2020). Lastly, the mean age
of the respondents was 50, with a standard deviation of 10 years.
Figure 1.4: Diary Sample Distribution by Intersectional Identity

Timing of Data Collection
I employed an interval-contingent protocol, which is a time-based design that requires
participants to report on their experiences at regular predetermined intervals (Bolger et al., 2003).
Because the sample included working professionals and emotional events are difficult to define
25
for participants, signal-contingent and event-contingent
5
protocols were avoided. I instead used a
fixed-interval design to avoid additional participant burden with a more intrusive frequency and
set a time lag of 24 hours within which respondents could answer about a particular workday
without introducing excessive lag time that would blur the experiences between days (Ohly et al.,
2010). Scollon et al. (2003) cautioned that the quality of the data typically declines after 2–4
weeks of continuous data collection, so diaries were capped at 14 workdays over the course of 3
work weeks from April 13–30, 2020.
The daily diary entries were designed to capture the emotional experience of employees
for that particular workday. On each day, participants received a set of prompts asking about
their interpersonal interactions with professional colleagues as well as interactions with the
public (see Appendix 1A). Emotional episodes
6
described in open-ended diary responses were
the unit of analytical interest.
Diary respondents were also given the opportunity to opt into a qualitative exit interview.
The interviews were 30-minute semi-structured conversations. In total, 60 participants from the
sample of diary respondents opted into interviews. The sample of interviewees were fairly
representative of the larger diary sample, though White women responded at slightly higher rates
and men of color responded at slightly lower rates than they did for the diary entries (see Figure
1.5). The interview protocol was informed by diary responses and designed to help develop

5
Signal-contingent protocols rely on some signaling device to prompt participants to provide
diary reports at fixed, random, or a combination of fixed and random intervals (Bolger et al.,
2003, p. 588). Event-contingent protocols require participants to provide a self-report each time
the event in question occurs, but any ambiguity in the definition of the triggering event(s) may
lead participants to omit relevant examples (Bolger et al., 2003).
6
Grandey et al. (2013) defined emotion regulation episodes as “segments of experience that are
thematically organized around goals for regulating our emotional experiences and expressions”
(p. 34).
26
additional context around emotional episodes for employees on the job. I took a semi-structured
approach following the DRM guidance to allow participants to first structure their own response,
then probe for details about event sequencing, people involved, consequences, and the meaning
participants made of the experience in their own minds. To verify understanding, this approach
requires asking participants about their interpretation of emotional events and checking my own
interpretations to see if I misunderstood or only partially understood their perspective. Additional
considerations for how I addressed my own positionality and reflexivity when working with
essential workers during the pandemic are available in Appendix 1C.
Figure 1.5: Interview Sample Distribution by Intersectional Identity

RESULTS
Qualitative Diary Prompts & Content Analysis
The first research question asks, (RQ1) is there a difference in the way employees
describe emotional labor in a service encounter with residents versus internal interactions within
the organization? Given that all the concepts pertaining to emotional labor in public service have
not been identified or fully developed, this initial question is exploratory in nature, lending itself
to a qualitative research design (Corbin & Strauss, 2014).
27
For diary coding, I used the entire free response entry as the unit of analysis. Some
participants responded with a single word or sentence, while others wrote multiple paragraphs
containing several interactions. Selecting the entire entry was a way of standardizing the unit of
analysis across respondents and approximates incident-by-incident coding (Charmaz, 2014). I
developed a coding scheme based on code families (Campbell et al., 2013). As recommended by
Charmaz (2014), coding of diary entries began immediately on the first day of data collection,
which enabled me to study emerging data to identify which codes to explore as tentative
categories and how I could best leverage qualitative interviews to fill the conceptual gaps in the
diary data.
Using ATLAS.ti software, new codes were created on a continuous basis, eventually
reusing more and more of the existing codes until reaching the first saturation point where there
was no further need to create new codes (Friese, 2019). This initial round of coding produced
over 100 unique codes, which were then sorted into higher-level concepts to synthesize how the
concepts fit together. In the second phase, I identified all entries that used language invoking
actions/interactions on the job. Among the diary entries involving work-related interactions,
secondary codes were developed through a focused and selective approach that used the most
frequent and significant initial codes to sort and integrate large amounts of data into thematic
clusters
7
(Charmaz, 2014). This round of coding sorted entries into external versus internal
interactions. For example, a diary entry like “responding to upset constituents” would have been
assigned a primary code of interaction and a secondary code of external (for more examples, see
Appendix 1D). Internal interactions were further sorted into mentions of (1) peers/coworkers, (2)

7
A random sample of 50 diary entries was given to a graduate research assistant to confirm
coding was applied consistently. Eighty-two percent of codes from the independent coder were
consistent with my original coding.
28
subordinates/supervisees, and (3) management/supervisors/leadership. These categories meet the
recommended criteria of being clearly distinguishable from each other, allowing for a thematic
analysis within each (Friese, 2019). By constantly comparing incidents within each of the three
categories, different properties began to emerge (Glaser & Strauss, 1999). In particular, entries
across internal and external interactions were further coded by mention of negative emotions and
appraisals (e.g., anxiety, anger, frustration, bad day, etc.); positive emotions and appraisals (e.g.,
happiness, hope, contentment, good day, etc.); and emotional labor (e.g., changing expression
through surface acting, changing emotional experience through deep acting), which co-occurs
with both positive and negative emotions.
Figure 1.6: Qualitative Coding Scheme

There were a total of 6,897 responses to 10 distinct open-ended diary prompts (see
Appendix 1A). Of those responses, 1,944 (28%) made reference to both a negative emotional
experience, and 1,172 (17%) made reference to a positive emotional experience. Looking at the
subset of 3,116 total diary entries with mention of either a positive or negative emotional
experience, participants most often shared work-related interpersonal interactions as the root
source of their emotional experience. Among entries that mentioned both work-related
interpersonal encounters and emotions, internal interactions accounted for 79% of all entries, and
external interactions with the public accounted for just 21% (see Table 1.1). The columns in
External
Interactions
Internal
Interactions
Emotional
Labor
Negative
Positive
Emotional
Labor
Negative
Positive
Subordinates Coworkers Supervisors
29
Table 1.1 show the percentage of workplace interactions that evoked negative emotions, positive
emotions, and emotional effort.
Table 1.1: Co-occurrence of Emotional Experience by Work-Related Interaction
External
(21%)
Internal
(79%)
Coworkers
(51%)
Staff
(14%)
Supervisors
(14%)
Negative Emotions 90% 76% 87% 78%
Positive Emotions 10% 24% 13% 22%
Emotional Labor 32% 18% 41% 18%
External Interactions With the Public
Three themes emerged specific to external interactions: (1) the challenge of managing
residents’ emotions, (2) difficulty navigating internal constraints and departmental silos to meet
the needs of the public, and (3) how rewarding it can be to help the community (see Table 1.2).
Table 1.2: Examples of Emotions Evoked by External Interactions
Managing
Residents’
Emotions
“Continuing to be polite and nice to someone, even though they are yelling at you for no
reason.”

“Interacting with people wearing facial coverings is really difficult. Tough to read emotions
and humor, nearly impossible to connect on a deeper level emotionally. Even though I am no
longer isolated all day, the coverings make it feel isolating.”
Difficulty
Navigating
Internal
Constraints
“Dealing with angry people with no authority to actually help them.”

“We are trying to help a resident, and my department tells me it’s not my job to find housing
for a severely ill person.”

“Unfortunately, what he needed help with was something I was not familiar with. I directed
him to the correct department to call. He became really angry because he said he didn’t want
to wait on hold and wanted me to find him someone to speak to. I explained that that
department is completely separate from my unit and I wouldn’t even know who to transfer him
to.”
Rewarding
Experiences
“I was able to speak with an owner of an apartment to see if a few minor violations had been
corrected. He was in a good mood and pleasant to speak with.”

“I discovered some kids were here at the park last night and they took some sidewalk chalk and
created a huge mural on our basketball courts … that made me feel good.”
30
Managing the Emotions of the Public
Ninety percent of the external interactions with the public were described as negative,
with feelings of anxiety and anger being referenced at a ratio of 9:7. Employees described
members of the public getting angry, frustrated, and irritated, which was a source of concern,
anxiety, and irritation in employees themselves. Many employees described their encounters
becoming “hostile,” with one employee sharing that a resident “proceeded to yell and scream at
me in front of his two boys—he was rude, threatening, and condescending” (5:547). In most
interactions with residents, there was a lack of familiarity, with one employee sharing it can be
stressful that “there is always an element of the unknown interacting with the public” (17:192).
Thirty-two percent of the diary entries describing external interactions made explicit reference to
the emotions of residents. Employees shared that they believed their jobs required them to
endure mistreatment, and that they had developed strategies for coping with frustrated residents
who misdirect their anger towards them. One employee described a resident who called and
“started out simply attacking government in general, so I basically let him rant in my ear for
several minutes” (5:1109). Another employee shared, “I listened to the frustration and the fear
and spoke calmly about what we had done and are doing” (5:492). Employees seemed attuned to
the emotional effort that goes into these interactions and shared, “dealing with clients who are
extremely stressed out … while remaining calm and collected myself does result in some ‘stress
transference,’ if you will” (17:145).
Difficulty Navigating Constraints & Departmental Silos
Beyond managing the emotions of residents, employees shared that it was often
characteristics of the work environment that drove their negative emotions. In particular, 25% of
entries referencing external interactions also mentioned irritation stemming from navigating the
31
burdensome nature of policies and procedures, inadequate technology/resources, and/or the
siloed nature of departments. Employees shared, “talking to the public makes me sad because we
can’t fix every problem” (5:1526) or even answer every question (4:722). Employees described a
sense of helplessness, highlighting the challenge of “dealing with angry people with no authority
to actually help them” (4:718) or communicating that their department was unable to assist them
(4:1389). These findings are consistent with street-level bureaucracy theory, which asserts that a
major source of tension in local government is trying address residents on a case-by-case basis
while adhering to rules and regulations internally that restrict employee discretion (Lipsky,
2010). On rare occasions, employees described their ability to work around rules and “red tape”
to help the public as a source of pride and accomplishment:
There was a homeless old lady with sweet eyes. She looked at us worriedly, and I
am sure she was wondering if we were going to ask her to leave. Before I could
say anything, my partner said, “I'm not going to ask that lady to leave. Why, so
she can go onto the street and be victimized?” That made my heart smile for an
instant. As the train pulled away, the sweet old lady smiled and gave a thankful
nod as she waved. My heart breaks a little bit every time I see someone like her.
It’s in these moments on the job where a flood of different emotions hit me all at
once. I was sad for her, I was happy to serve her by leaving her be, and I was
proud of my partner and my job. (5:1253)
Rewarding Feelings From Helping the Community
Despite the challenges, 10% of diary entries mentioning interactions with the public
described positive emotions, most often a sense of gratification from the encounter. One
employee shared, “I like seeing the residents calm down and look more comfortable than when
they first walked in” (16:2186). This was particularly true when employees were able to build
recurring relationships with residents. Having the opportunity to follow up with a resident they
had helped in the past was described as a source of joy and purpose. One employee shared, “I got
a call from a crying woman who said she was having a crisis and couldn’t reach her therapist, but
32
then she thought of me. … She started out crying and ended up laughing, and I felt good about
this” (16:286).
Taken together, the themes that emerged from entries involving external interactions
suggest most encounters with residents are emotionally taxing, and employees feel they must
display positive emotions and suppress negative emotions despite mistreatment. However, when
employees are able to build relationships and help residents resolve their issues, there is more
potential for positive encounters and meaningful connections.
Internal Interactions With Colleagues
Interestingly, the vast majority (79%) of the interactions described by participants were
internal to the organization. While external interactions with residents were mostly described as
one-time encounters, the overwhelming majority of internal interactions with colleagues were
described as recurring relationships. Participants described internal interactions in terms of
relational context variables such as familiarity, power dynamics, and warmth. In particular,
negative emotional experiences frequently co-occurred with mentions of colleagues where there
was an expectation of tension rooted in fundamental incompatibility or tension surrounding
differences between colleagues at different levels of the organizational hierarchy.
Familiarity—Irreconcilable Differences
Respondents described some of their coworkers as consistent sources of frustration, with
one employee sharing, “one particular coworker just frustrates me every day” (5:1420) and
another adding that their primary source of stress came from “dealing with negative attitudes
from negative colleagues” (4:1141). These frustrations seemed to stem from differences in both
work style and personality. One respondent shared that he and a coworker “share responsibility
for communicating information to the community, and we have very different styles and
33
thoughts about how best to do that” (4:932). Another referenced different emotional tendencies
between colleagues, sharing, “I know everyone has their days, and I can do pretty good masking
my feelings; others do not have the same ability, so their emotions can rub off and cause stress in
the work environment” (17:1118). Respondents described needing to be “patient” and “allowing
feelings of frustration to pass” in order to manage their relationships with colleagues. These
descriptions suggest that as employees become familiar with their colleagues, they develop
expectations for how the interactions will go, and have strategies for managing their emotions
around difficult colleagues.
Power & Responsibility
In addition to coworker relationships, hierarchical relationships with staff and supervisors
each accounted for another 14% of work-related interactions respectively. For supervisors, it was
particularly challenging to navigate the emotional experiences of their staff. Eighty-seven
percent of all staff interactions were described as negative. One supervisor described how
difficult it can be to support employees who are “not prepared to deal with the stress of being a
frontline provider to the public” (5:1969). Across most departments and job functions in the
sample, leaders shared the challenge of attuning to the emotional coping mechanisms of their
staff. One leader described their biggest source of stress coming from “dealing with 109 different
employees with different ways of handling anxiety and fear” (4:959). Beyond supporting the
needs of their teams, supervisors also had their own emotions to manage and often described the
need to maintain “professionalism.” One respondent shared that a major difficulty was “wanting
to tell my team how I really feel but needing to keep it professional” (4:1159).
The stress of leader–staff interactions often ran in both directions, with some staff
complaining that leadership “isn’t great with making staff feel better or uplifted” (5:1436).
34
Seventy-eight percent of interactions with supervisors were described as negative. One employee
shared that the most difficult part of their day was “being reminded how rude my coworker is
and how my boss doesn’t want to address it” (4:1630). Other employees shared the sentiment
that they felt their supervisors were responsible for setting the emotional tone for the group and
intervening. A very consistent source of frustration was supervisor/management
(mis)communication. When interacting with leadership, respondents described feeling
constrained in their attitudes and behaviors because they were in a subordinate position.
Respondents also shared how emotionally challenging it can be to interact with supervisors who
do not notice or value their contributions.
Warmth
When negative emotions emerged among colleagues, 15% of diary entries noted they
confided in trusted coworkers as a coping mechanism or release of tension. As an example, one
employee described venting to a colleague about shared frustrations, saying, “it actually felt
good to talk about my frustration and even more to hear I wasn’t alone—it didn’t solve the
problem, but it still had some resolution” (5:809). On the receiving end, being a confidant for
others was described as both rewarding and challenging. One respondent shared, “it felt good to
help someone just by being a listening ear when she needed to talk” (16:911). In relationships
characterized by less warmth, employees described suppressing their negative emotions to avoid
burdening their colleagues with their feelings. As examples, one employee shared that they
focused on “keeping negative emotions private so as not to affect those around me, coworkers
and subordinates” (5:1101), while another explicitly named the inauthenticity they felt when
manipulating their emotions, sharing, “I try to keep positive when talking with coworkers, but
sometimes I feel as if I’m putting on a mask” (14:1307).
35
Taken together, the themes that emerged from internal interactions suggest employees
were cognizant of the need to present positive emotions for the sake of maintaining relationships
and group morale, which often required suppressing negative emotions. The extent to which
employees felt like they could be honest with their emotions depended in part on familiarity,
warmth, and power dynamics in individual relationships, including the emotional tone set by
leadership.
Statistical Modeling
Building from the initial qualitative analysis, I turn to statistical modeling to understand if
the frequency of either internal or external interactions is a predictor of emotional labor. A series
of closed-ended survey items in the daily diaries were used to test these relationships through
both ordinary least squares (OLS) and multilevel statistical modeling techniques. Both the
literature and preliminary content analysis suggest we should anticipate that more frequent
internal and external interactions should drive surface acting—specifically amplified positive
emotions (H1) and suppressed negative emotions (H2). Less is known about the relative strength
of the relationship between internal versus external interactions and surface acting behaviors.
Measures
The dependent variable of interest was emotional labor. Glomb and Tews (2004)
developed a conceptually grounded, psychometrically sound instrument to measure emotional
labor with an emphasis on the experience of discrete emotions—the Discrete Emotions
Emotional Labor Scale (DEELS). Their conceptualization and operationalization of emotional
labor encompasses genuine, faked, and suppressed positive and negative emotional displays, and
initial results provide evidence of the convergent, discriminant, and criterion-related validity of
the DEELS instrument (Glomb & Tews, 2004). Exploratory factor analysis showed that discrete
36
emotions clustered into “positive” and “negative” groupings across all three types of emotional
labor (i.e., genuine expression, amplification, and suppression). The “positive” clusters for
genuine, amplified, and suppressed emotions included (1) enthusiasm, (2) happiness, and (3)
contentment. The “negative” clusters for genuine, amplified, and suppressed emotions included
(1) irritation, (2) anxiety, (3) sadness, (4) concern, (5) fear, and (6) anger.
At the event level, the models included separate controls for the frequency of interaction
employees had externally with members of the public and internally with members of their
organization to account for variations in social interactions on a given day. At the person level,
demographic differences in gender, race, and age were controlled. Additionally, the models
accounted for differences in the motivation employees had to help others (prosocial motivation),
as well as perceptions of the workplace climate surrounding emotional expression and social
support among colleagues, which were measured at baseline. Survey items for all control
variables are shown in Appendix 1A.
Model Specification
Six OLS models with dependent variables for the genuine expression, amplification, and
suppression of positive and negative emotions were specified to assess whether or not the
frequency of interactions with colleagues or residents were predictors of both positive and
negative clusters of emotions. The equation for the first OLS model is listed below, and
subsequent models switch out the dependent variable and adapt the controls to include the five
additional emotional labor factors beyond the dependent variable. To account for potential
differences between street-level bureaucrats and those in roles that are predominantly internal,
the results in Table 1.3B include an additional control for the percentage of time the employee
spends engaging the public directly as a function of their job. This variable was collected using a
37
single survey item prior to the pandemic, which creates some challenges for interpretation given
that some employees saw the nature of their role change significantly with the shift to
teleworking and assignments to serve in temporary disaster service worker (DSW) roles at the
time of the diary study. All models included department fixed effects and robust standard errors
clustered at the individual level.
Genuine Positive = !
!
+ !
"
(genuine neg) + !
#
(amplified pos) + !
$
(amplified neg) + !
%
(suppressed pos) +
!
&
(suppressed neg) + !
'
(avg interaction colleagues) + !
(
(avg interaction residents) + !
)
(intersectional
identity) + !
*
(generation) + !
"!
(prosocial motivation) + !
""
(climate of authenticity) + !
"#
(social support)
+ ɛ
Table 1.3: OLS Models for Disaggregated Forms of Emotional Labor
8

1.3A Modeled Without % Public Engagement
VARIABLES (1)
AVG
Genuine Pos
(2)
AVG
Genuine Neg
(3)
AVG
Amplified Pos
(4)
AVG
Amplified Neg
(5)
AVG
Suppressed Pos
(6)
AVG
Suppressed Neg
AVG Freq coworkers 0.0671
(0.100)
-0.0122
(0.0824)
-0.332***
(0.108)
0.170
(0.104)
-0.458***
(0.0989)
0.491***
(0.0835)
AVG Freq residents

0.244***
(0.0644)
-0.520***
(0.0753)
0.229***
(0.0713)
0.153*
(0.0844)
0.0788
(0.0677)
0.141***
(0.0515)
Observations
R-squared
1,026
0.657
1,026
0.697
1,026
0.693
1,026
0.664
1,026
0.674
1,026
0.781
1.3B Modeled With % Public Engagement
(1)
AVG
Genuine Pos
(2)
AVG
Genuine Neg
(3)
AVG
Amplified Pos
(4)
AVG
Amplified Neg
(5)
AVG
Suppressed Pos
(6)
AVG
Suppressed Neg
% Public Engagement -0.00176
(0.0119)
-0.00647
(0.00757)
0.0393***
(0.00885)
0.105***
(0.0120)
0.0140
(0.00950)
-0.0414***
(0.00806)
Avg Freq coworkers 0.0652
(0.102)
-0.0191
(0.0843)
-0.284***
(0.105)
0.258***
(0.0964)
-0.442***
(0.0964)
0.431***
(0.0798)
Avg Freq residents 0.244***
(0.0642)
-0.518***
(0.0759)
0.211***
(0.0704)
0.0922
(0.0789)
0.0736
(0.0677)
0.152***
(0.0511)
Observations
R-squared
1,026
0.657
1,026
0.697
1,026
0.699
1,026
0.715
1,026
0.675
1,026
0.788
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

8
Full model results including controls are available in Appendix 1E.
38
Table 1.3 shows the coefficients for the average frequency of interactions with coworkers
as well as the average frequency of interactions with residents across days modeled without the
percentage of public engagement time pre-pandemic (1.3A) and with the percentage of public
engagement time pre-pandemic (1.3B). The first model in Table 1.3A shows a positive
association between more frequent interactions with residents and the genuine expression of
positive emotions
9
but no statistically significant relationship for more frequent interactions with
colleagues. The same signs and significance hold in the first model of Table 1.3B, where the
main effects of the percentage of public engagement time are not significant. The second model
in Table 1.3A shows a negative association between more frequent interactions with residents
and the genuine expression of negative emotions
10
but no statistically significant relationship for
more frequent interactions with colleagues. The same is true in the first model of Table 1.3B,
where again the main effects for percentage of public engagement time are not significant. The
key takeaway from the genuine emotion models is that frequent external interactions seem
tightly related to more genuine positive emotional experiences and fewer opportunities to express
negative feelings.
In the third model of Table 1.3A, there continues to be a positive association between the
frequency of interactions with residents and amplified positive emotions,
11
but there is also a
negative association between the frequency of interactions with colleagues and the amplification
of positive emotions. Table 1.3B also shows that jobs with a higher percentage of public
engagement tend to see higher levels of amplified positive emotions. In the fourth model of

9
The dependent variable in the first model is a factor variable for genuinely expressed positive
emotions (eigenvalue = 1.76) ranging from -0.95 to 2.05.
10
The dependent variable in the second model is a factor variable for genuinely expressed
negative emotions (eigenvalue = 1.92) ranging from -0.77 to 3.72.
11
The dependent variable in the third model is a factor variable for amplified positive emotions
(eigenvalue = 1.83) ranging from -0.45 to 3.83.
39
Table 1.3A, there is no statistically significant relationship between the frequency of interaction
with either residents or colleagues and the amplification of negative emotions.
12
Interestingly, in
Table 1.3B, jobs with a higher percentage of public engagement time tend to see more amplified
negativity, and the signs and significance for coworker interactions are not positively associated
with more amplified negativity, but more frequent resident interactions are not significant. The
key takeaway from the amplified emotion models is that external interactions contribute to the
pressure to fake feelings, whereas there is less manufactured positivity between colleagues.
In the fifth model of Table 1.3A, there is no statistically significant relationship between
frequency of interaction with residents and suppressed positive emotions,
13
but there is a
negative association between more frequent interactions with colleagues and suppressed positive
emotions. The same signs and significance hold in the fifth model of Table 1.3B, where the main
effects for percentage of public engagement time are not significant. In the sixth model of Table
1.3A, there is a positive association between the frequency of interaction with both residents and
colleagues and the suppression of negative emotions.
14
Table 1.3B shows consistent signs and
significance for internal and external interactions, but surprisingly, the main effects for jobs with
a higher percentage of public engagement time are associated with less suppressed negativity,
warranting further exploration. Taken together, the suppression models reveal that there’s
pressure to keep negative feelings bottled up internally and externally, but for those whose jobs
require routine public engagement, that pressure to suppress negativity may be weaker.

12
The dependent variable in the fourth model is a factor variable for amplified negative emotions
(eigenvalue = 3.77) ranging from -0.22 to 9.45.
13
The dependent variable in the fifth model is a factor variable for suppressed positive emotions
(eigenvalue = 1.83) ranging from -0.37 to 4.64.
14
The dependent variable in the sixth model is a factor variable for suppressed negative emotions
(eigenvalue = 3.46) ranging from -0.85 to 2.70.
40
To summarize, the models confirm there are statistically significant relationships between
our positive and negative emotional labor factors and the frequency of interaction with both
colleagues and residents. Column 3 in Tables 1.3A and 1.3B confirms a positive relationship
between external encounters and pressure to fake positivity but does not support the same for
internal interactions (H1). Column 6 across both tables confirms the qualitative insight that
frequent interactions with both residents and colleagues are related to suppressing negative
emotions on the job, supporting H2. Taken together, these results suggest that internally and
externally, there’s a persistent pressure to downplay negativity, but in interactions with the
public, it is especially important to put on a happy face. However, open questions still remain
about the extent to which street-level bureaucrats who have a higher percentage of their time
dedicated to public engagement may have a different set of strategies for managing their
emotions, given that they seem to have a positive association with amplifying both positive and
negative emotions but suppress less negativity than those in roles with less time dedicated to
public engagement.
Next, I ran a series of multilevel models that focus on multiple observations for
individual employees, which allow for an in-depth understanding of within-person differences,
using individuals as their own controls. First, I fit several unconditional models predicting the six
emotional labor factors to estimate intraclass correlation and assess the extent to which variation
in emotional labor was occurring within versus between individuals. These preliminary models
reveal the bulk of emotional labor variation has to do with differences between people based on
either their personal attributes (e.g., personality, emotional intelligence, etc.) or their job
characteristics, while 25–46% of variation comes from differences in the events experienced by a
single individual. Within-person variation reflects how experiences for one individual can differ
41
depending on time or context; a within-person modeling approach allows me to isolate the
relationships between interactions and emotional labor with more precision by using individuals
as their own controls.
The dependent variables in the six multilevel models were time-varying factors for (1)
genuinely expressed positive emotions, (2) genuinely expressed negative emotions, (3) amplified
positive emotions, (4) amplified negative emotions, (5) suppressed positive emotions, and (6)
suppressed negative emotions, which were analyzed on a daily basis for fluctuations in emotional
experience (level-1 variables). The independent variables of interest were level-1 time-varying
measures of frequency of interactions with residents and coworkers for a given individual by
day. Additional level-1 variables controlled in the model included the daily factors for the five
other types of emotional labor (i.e., positive/negative genuine expression factors,
positive/negative amplification factors, and positive/negative suppression factors). The
remaining controls constitute level-2 data, meaning they are subject-specific time-invariant
variables. The level-2 controls for all six models include gender, race, age, prosocial motivation,
perceived climate of authenticity, and social support, with fixed effects for person (ጊ
+
) and day of
study ("
,
). The equation for the first model is listed below, and subsequent models swap out the
dependent variable, adjusting the level-1 controls to reflect the other five emotional labor factors
beyond the new dependent variable.
Genuine Pos#
,+
= #
!!
+ #
"!
(interaction colleagues) + #
#!
(interaction residents) + #
$!
(genuine neg) +
#
%!
(amplified pos) + ##
&!
(amplified neg) + #
'!
(suppressed pos) + #
(!
(suppressed neg) +
#
!"
(intersectional identity) + #
!#
(generation) + #
!$
(red tape) + #
!%
(climate of authenticity) + #
!&
(social
support) + ጊ
+
+ "
,
+ %
!+
+ &
,+

42
Table 1.4: ML Models for Within-Person Variation in Emotional Labor
15

(1) (2) (3) (4) (5) (6)
VARIABLES Daily
Genuine Pos
Daily
Genuine Neg
Daily
Amplified Pos
Daily
Amplified Neg
Daily
Suppressed Pos
Daily
Suppressed Neg
Daily Freq coworkers

0.224*
(0.117)
-0.0656
(0.133)
-0.0128
(0.131)
-0.0500
(0.110)
-0.223
(0.144)
0.128
(0.121)
Daily Freq residents 0.0973
(0.0873)
0.0691
(0.0980)
0.256***
(0.0958)
-0.0686
(0.0805)
-0.128
(0.106)
0.0260
(0.0899)
Observations 336 336 336 336 336 336
Number of groups 95 95 95 95 95 95
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
I dropped individuals with fewer than three diary entries, leaving the total number of
subjects at 95 with 336 observations across individuals. Interestingly, the frequency of
interaction with coworkers was statistically insignificant at the 5% level across all six multilevel
models, and the frequency of interaction with residents was only significant in the case of
amplified positive emotions, where we see a positive association between more frequent
interactions with residents in a given day and the frequency of amplifying positive emotions that
same day for an individual. It might be inferred that public service with a smile is a consistent
display rule correlated with surface acting across individuals and job types, and when controlling
for all other individual characteristics, more frequent interactions with residents on a given day
are associated with more forced positivity. It could be that there is no real relationship between
internal interactions and surface acting behaviors, or there could be either limitations on the
measures used to capture the relationship or issues of the study being underpowered, with only
95 in the sample.

15
Full model results including controls are available in Appendix 1F.
43
Exploratory Semi-structured Exit Interviews
It is curious that the original content analysis found internal encounters with colleagues
were more often described as a source of emotional labor among employees, but statistical
modeling suggests frequency of interaction with the public is predictive of more forms of
emotional labor. The discrepancy between qualitative and quantitative findings among the same
sample helps to raise constructive doubts that are generative in a triangulated research process
(Locke et al., 2008). The statistical analysis is limited by the measures in the models, which only
describe the frequency of interactions and emotional labor but not the intensity of either. Perhaps
this indicates that emotional labor in relationships with colleagues carries more emotional weight
in memory.
To understand why internal encounters originally appeared more salient in the
employees’ recollections in their open-ended diary entries, exit interviews were conducted to
expand upon preliminary findings and complement research participants’ written accounts
(Charmaz, 2014). In a phased approach, interview coding progressed from tagging internal and
external interactions to inductively developing secondary codes that would eventually be sorted
into thematic concepts. In comparing the descriptions of internal versus external encounters, it
emerged that engaging in the research process itself made participants more aware of patterns of
emotional labor in recurring relationships, which tended to be internal.
Interviewees most often reflected upon recurring relationships rather than encounters,
whether that was to express appreciation for their colleagues or challenges between them. When
encounters with the public were described, interviewees shared that they had less personal
information about the individual to help them anticipate how the interaction might go and more
formal organizational expectations about how service encounters should be handled. With
44
respect to internal dynamics, respondents described anticipating either a positive or a negative
emotional experience based on the nature of their relationship with the colleague involved.
Interestingly, interviewees shared that they were affected by their own relationships and the
relationships of their peers, often engaging in mediation to resolve issues between coworkers.
Many interviewees described experiences where they were not directly engaged but observed an
interaction with a colleague they cared about that evoked emotion. Interviewees described a
more ambiguous and informal set of emotional expectations between colleagues. While service
encounters provide clear-cut examples of emotional labor performed to meet organizational
expectations, internal interactions were described as layered with much more complexity given
the range of relationship dynamics and informal emotional expectations in groups. Interviewees
were also asked to reflect on the research process itself. Their responses revealed that the nature
of a daily diary encouraged them to notice patterns, and patterns were more often detected when
participants were able to reflect on multiple interactions with the same individual (i.e., a
colleague).
DISCUSSION
This paper set out with two empirical questions: (RQ1) whether there is a difference in
the way employees describe emotional labor in external encounters with residents versus internal
interactions with colleagues and (RQ2) whether it is the frequency of (a) internal interactions or
(b) external interactions that is most tightly correlated with surface acting behaviors. The results
of this paper make it clear that public administration scholarship needs to focus on the
interpersonal competencies demanded in both internal and external interactions. It is also clear
that we are really underestimating emotional labor in public organizations if we are overly reliant
on a single method.
45
While it is tempting to interpret the findings from the qualitative and quantitative
components of this study as contradictions, the unexpected and counterintuitive results stem
from the inherent features of the different data collection time frames and methods (Mele &
Belardinelli, 2019). Collecting data at the onset of the COVID-19 pandemic may have coincided
with fewer public interactions and put unusual emphasis on new remote and hybrid forms of
interactions that had embedded challenges related to the new mediums of communication. In the
diary entries, it is made clear that employees were still interacting with others often, with roughly
one-third of the sample continuing to work in person, one-third hybrid, and one-third remote.
Thus, I feel confident that the emphasis on internal interactions in the qualitative content analysis
is not an artifact of the time when this data was collected. More likely, it is the case that the
inherent features of this triangulated research design reveal different dimensions of emotional
labor. The contradictions justify the rationale for preferring a mixed-methods over a single-
method design.
Allowing respondents to structure their own open-ended responses revealed that
interactions internal to the organization among colleagues are often emotionally charged, based
on how frequently they were described as sources of emotional labor. Beyond the frequency of
interaction, the initial qualitative work revealed that characteristics of internal relationships such
as familiarity, power, and warmth seemed to influence employees’ emotional experiences, and
the semi-structured exit interviews allowed for a deeper exploration of these complex and
emotionally evocative relationships among coworkers. This study makes a methodological
contribution by testing qualitative findings through a dynamic statistical approach to study
emotional labor through the use of diaries to uncover the between- and within-person variation in
experiences of emotional events and to accommodate day-to-day differences. This variation has
46
been ignored in previous research using aggregated measures of emotional labor. Statistical
modeling revealed inconsistencies from initial qualitative findings, which is exciting because it
points to the need for a deeper understanding. Modeling revealed that interactions with residents
were predictive of multiple forms of surface acting, whereas internal interactions were only
predictive of suppressing negative emotions in averaged models and statistically insignificant
when looking at within-person differences. The use of a diary study, which is sensitive to such
variation and has high internal and ecological validity, and complementary qualitative interviews
constitutes a major improvement over current research into emotional labor in public
administration.
This doesn’t mean the qualitative findings were wrong, but perhaps incomplete.
Returning to a content analysis of interviews allowed for more nuance to be added. In conducting
the second round of exploratory content analysis, it becomes clearer that recurring relationships
seem to carry more emotional weight in memory, and a diary design may add to those insights by
emphasizing emotional pattern recognition throughout the research process.
When asked to reflect on their emotions, employees gravitated towards emotional
encounters and relationships inside the organization. The qualitative results make clear that
navigating interactions with colleagues consumes a lot of mental and emotional energy. These
interactions are varied and demand emotional sophistication. Quantitatively, this study shows
that there is pressure in most daily interactions to suppress negativity, but especially with the
public, there is an expectation of service with a smile. This is made more apparent in within-
person models of surface acting variation, where we see the biggest increase in forced positivity
coming from frequent interactions with residents. While these external interactions seem to carry
more weight statistically, they seem to be less meaningful and less memorable to employees.
47
This warrants additional research that explores the relationship characteristics among colleagues
and how relational- and group-level factors might influence emotional labor. In particular, future
scholars should aim to recreate the models here while disaggregating internal interactions to
reflect the organizational hierarchy. Specifically, disentangling how lateral interactions with
coworkers versus vertical ones with subordinates or supervisors might influence emotional labor
will be important in deconstructing the ways in which organizations reinforce inequity.
Codifying familiarity, warmth, and power asymmetry in internal dynamics is an important next
step in emotional labor scholarship in the public sector. In particular, a more in-depth exploration
of status and power asymmetry as a driver of emotional labor within public organizations will
help unpack important diversity management considerations.
CONCLUSIONS/POLICY IMPLICATIONS
What is clear from this study is that public service work is emotion-laden, and emotional
labor should be treated as a competency in local government. Awasthi and Mastracci (2021)
proposed a set of integrated emotive competencies that should be adopted in public affairs
education to help make visible the interpersonal skills necessary for public organizations to
thrive. Future work should explore how enhancing competencies for emotional labor, emotional
intelligence, and empathy might positively impact public administrators, public organizations,
and the communities they serve. In recent years, scholars have called for recruitment, training,
development, and support for employees that address the whole person, which includes
implementing organizational practices
16
to help build emotional self-awareness to support
ongoing emotion management (Mastracci & Sawbridge, 2019).

16
See Mastracci and Sawbridge (2019) for a comprehensive review of organizational
interventions designed to help build capacity for emotional labor, including Creating Learning
Environments for Compassionate Care (CLECC), Relation-Centered Leadership (RCL), Group
Supervision (Smojikis; ), Mindfulness, Restorative supervision, Samaritans Volunteers Support
48
The emotional burdens experienced on the job have important implications for employee
burnout, mental health, and retention. Systematically describing and measuring emotional
experiences and labor highlights the opportunities for public managers to develop, train, and
reward employees for important interpersonal skills while also implementing practices to help
employees recover from particularly taxing forms of it (i.e., surface acting). Importantly, this
paper has only scratched the surface, and future research should aim to explore the lagged effects
of emotional labor both on the job and at home.
Emotional events are likely part of a feedback loop, where the outcome of one’s
emotional labor becomes the next emotional event. This “double interact” (Rafaeli & Sutton,
1987) is seldom tested due to the complexity of such analysis (Grandey & Melloy, 2017). Based
on the meaning one makes, emotions can hold positive and negative connotations simultaneously
or in quick succession, and these emotional ambivalences are in need of study (Bagozzi, 2003).
Emotional ambivalences may call for more attention to be paid to the dynamic social context of
organizations (Fischer & Van Kleef, 2010). Walter and Bruch (2008) suggested that other
contextual variables, such as perceived organizational support and social network structures
within work groups, may further influence the spiraling process and highlight the importance of
longitudinal study designs when attempting to empirically test the interactive dynamics of
collective affect emergence (Diefendorff et al., 2021).
For Guy et al. (2008), “we live in an era when greater responsiveness is required of
public workers, yet there is a blind spot in employee performance appraisals that makes emotion
work invisible” (p. 122). Additional multilevel research examining the interplay across the team
and organizational levels is needed to understand emotional labor and potential implications for

System, Schwartz Rounds (Point of Care Foundation (PoCF)), Critical Incident Stress
Debriefings (CISDs), and Self-Care Plans.
49
burnout in local government (Humphrey et al., 2015; Maslach et al., 2012; van Kleef & Fisher,
2016).

50
CHAPTER II. Emotional Labor at the Intersections of Identities: An Assessment of
Differential Emotional Labor Experiences in Local Public Service
ABSTRACT
An intersectional approach to the study of emotional labor is essential to make visible the
experiences of those with multiple marginalized identities. This paper explores the frequency and
variety of emotional labor experienced by local government employees and asks whether the
experience of emotional labor differs in intersectional analyses of women of color compared to
analyses that look at race and gender as separate independent variables. Specifically, this
analysis leverages open-ended prompts and measures of discrete emotions in a series of daily
diaries to unpack how different groups engage in emotional labor. The results reveal that women
of color engage in more emotional labor than their peers, feel more emotionally constrained by
organizational rules, are cognizant of managing gendered and racialized stereotypes, and are
more sensitive to whether the climate allows for authentic expression. Thus, this paper makes
visible the experiences of those at the margins of multiple lived experiences of oppression,
allowing women of color to articulate their own emotional experience in a way that centers their
voices.
INTRODUCTION
More representative organizations can improve performance, but increasing the number
of women and BIPOC in the workplace is not sufficient to enhance equity and fairness within the
organization (Ding et al., 2021; Hoang et al., 2022). Public administration scholars are calling for
more research that explicitly acknowledges how Whiteness and masculinity shape our
understanding of public organizations (Portillo et al., 2022). Racial socialization and the benefits
of aligning with Whiteness complicate using representation as a proxy for equality in
51
organizations (Agho, 2022). For Ray et al. (2022), “racialized burdens emerge via intentional
(but plausibly deniable) rules, or facially neutral rules that disproportionately harm marginalized
racial groups” (p. 3). Humphrey (2021a) added that more research is needed to unpack the
workplace challenges experienced by public servants of color, who may experience conflicting
demands and pressures from organizational socialization and their racial identity. On top of
standard job demands, employees of color are expected to address racially charged issues on
behalf of their organizations, creating expectations for racialized emotional labor (Humphrey,
2021a).
Hochschild (1983) defined emotional labor as “the management of feeling to create a
publicly observable facial or bodily display for a wage” (p. 7). Grandey (2000) proposed
managing observable expressions (surface acting), managing feelings (deep acting), and
regulating the expression of genuine emotions in organizational settings as the working
definition of emotional labor. Grandey and Gabriel (2015) argued women and people of color
may be culturally expected to regulate their emotions to manage negative group stereotypes and
conform to organizational expectations, perpetuating emotional dissonance and inequity in
organizations among the underrepresented. This paper demonstrates that emotional labor differs
significantly for women of color, who engage in more taxing forms of emotional labor than
White women, men of color, and White men.
I begin by presenting a literature review of organizations as gendered and racialized
structures where stereotypes and controlling images constrain the behaviors of women and
people of color. Next, I review the emotional implications of gender and racial power dynamics
in organizations, highlighting that intersectional theorizing about emotional labor among women
of color has been missing from diversity management scholarship. I then introduce the mixed-
52
methods research design that combines qualitative interviews with a 3-week diary design among
local government employees in a large metropolitan area in California. Lastly, I present the
results of the quantitative and qualitative analyses, discuss the implications, and identify avenues
for future intersectional emotional labor scholarship. I argue throughout that the failure to
account for the implicit emotional labor embedded in organizational life creates management
blind spots with respect to supporting an increasingly diverse workforce.
Organizations as Gendered and Racialized Structures
Critical race and feminist theories emphasize race and gender as structures of oppression
for the way relative status can constrain behavioral freedom in organizations (Acker 2006,
20012; Martin, 2000; Ray, 2019; Ray et al., 2022). Research on conformity explores how women
and racial/ethnic minorities are encouraged to adopt the same values, informal norms, and formal
practices as their White male colleagues, which leaves unanswered what advantages of diversity
are lost through assimilation practices (Martin, 2000; Mastracci & Arreola, 2016). Researchers
have shown that individuals respond to diversity among team members through emotion
regulation, typically surface acting (Grandey & Melloy, 2017; Kim et al., 2013). As a response to
assimilation pressures, Ozcelik (2013) argues surface acting occurs in workplace relationships to
either maintain a sense of acceptance and belonging at work or secure key resources.
For Ray (2019), organizations are racial structures, where cognitive schemas and
stereotypes operate as the link between formal and informal organizational rules and disparities
in social and material resources. Stereotypes in this sense can be defined as impressions that
“help people identify what members of various categories are believed to do (descriptions), are
encouraged to do (prescriptions), or are discouraged from doing (proscriptions)” (Hall et al.,
2018, pp. 644–645). Negative stereotypes place an additional burden on members of stereotyped
53
groups, who feel their personal failure would reflect negatively on the larger group to which they
belong (Roberson & Kulik, 2007). Research has shown that social group stereotypes can have a
harmful effect on employee well-being and make it difficult for them to perform to their true
potential (Roberson & Kulik, 2007).
Humphrey (2021a) noted that some racial groups may have stereotypes that are
incongruent with role expectations in organizations, meaning members of these groups are
evaluated more harshly. As an example, roles that demand relationship-building require
interpersonal warmth are often incongruent with the stereotype of the angry Black man, leading
to harsher evaluations of Black men’s performance (Grandey et al., 2019; Humphrey, 2021a).
Past research has shown that employees of color respond to negative stereotypes and controlling
images by cultivating non-threatening dispositions and appearances (Durr & Wingfield, 2011;
Wingfield 2007, 2010). Many employees of color engage in impression management
17
and
“strategic passing” to navigate organizations so they are not mischaracterized along stereotypical
lines (Agho, 2022).
Emotional labor is one avenue employees of color have for counteracting negative group
stereotypes (Grandey et al., 2019; Humphrey, 2021a). Prior research has also suggested that
emotional labor is a mechanism for building relationships to enhance social capital (Humphrey,
2020, 2021a). In everyday interactions, women and employees of color are often forced to
choose between assimilating to achieve belonging and bringing their full selves to work to
maintain authenticity (Kanter, 1977; Smith et al., 2019). For employees of color, presenting as
palatable to White people and White institutions through emotion management is a strategy for

17
Chawla et al. (2020) defined impression management as a set of strategies leveraged by
employees to create, maintain, or alter a desired image towards others.
54
maintaining favorable impressions, and Cartwright (2022) argued, “these cultivated dispositions
and projections of palatability function as cultural capital for nonwhite people” (p. 7).
Emotional Implications of Gender and Racial Power Dynamics
The most frequently studied marker of difference in the emotional labor literature is
gender, but empirical findings have been mixed. Hochschild (1983) originally argued that
women in general have less independent access to key resources (e.g., money, power, and
authority) in society, and this subordinated status has several implications: (1) Women use
emotion as a resource to offer men in exchange for the material resources they lack; (2) women
are socialized in childhood to specialize in accommodating forms of emotional labor; (3) general
subordinated status leaves individual women with weaker protections from verbal and emotional
abuse. In their meta-analyses, Bono and Vey (2005) found gender to be unrelated to emotional
labor, while Wang et al. (2011) found gender has a small but significant relationship with surface
and deep acting. I propose these conflicting findings may be caused in part by meaningful
differences in the emotional expectations, constraints, and freedoms placed on White women
compared to women of color.
Wingfield and Alston (2014) have argued that workers of color in predominantly White
workplaces take on unacknowledged self-presentation and emotional work in social interactions
on the job. They characterize these interpersonal burdens as racial tasks
18
that ultimately
preserve organizational hierarchies and reproduce inequity. Employees of color are often
informally tasked with ingratiating themselves to White colleagues and managing their own
impression with peers to counter negative race-based stereotypes. Undertaking the additional

18
Racial tasks in this context refer to the ways in which implicit responsibilities at work have
racialized meanings that affect the processes and behaviors within the organization (Wingfield &
Alston, 2014).
55
work of racialized emotional labor means there is an inequitable distribution of interpersonal
effort in most organizations. Whether it’s laughing at a racist joke or restraining one’s own
frustration to make those around them comfortable, racial tasks in organizations imply
employees of color are not permitted the same opportunities for genuine emotional expression as
their White colleagues (Wingfield, 2010; Wingfield & Alston, 2014).
There are problematic consequences of what critical race scholar Kimberlé Crenshaw
(1989) termed a single-axis framework, or the tendency to treat race and gender as mutually
exclusive categories of experience and analysis. Despite the wealth of studies of gender
differences in emotional labor, surprisingly little attention has been paid to racial differences
within gender categories, homogenizing the experiences of racially diverse groups of women
(Grandey et al., 2013; Wharton, 1999). Because the self is composed of overlapping, nested
identities that become more or less pronounced in different contexts, an intersectional approach
to the study of emotional labor is more informative than focusing on any one social category at a
time (Grandey & Melloy, 2017; Martin et al., 1998). Intersectional scholars approach lived
identities as an interlaced “matrix” of mutually reinforcing systems of oppression where
individual aspects of identity cannot be treated as separable or as superordinate (Collins, 2008;
May, 2015; McCall, 2005). To date, there are few intersectional studies of emotional labor, with
those that exist predominantly exploring gender differences among Black employees (Wingfield,
2007, 2009, 2010, 2020, 2021).
Differences in Black Men and Black Women’s Emotional Labor
Wingfield’s (2007, 2009, 2010, 2020, 2021) work has shown how Black employees learn
to repress their emotions and downplay their experiences of racism to combat negative
controlling images and make the White people they work with comfortable, making their
56
emotional labor a tool for yielding institutional rewards and building cultural capital (Cartwright,
2022). When Black employees experience microaggressions on the job, emotional norms to be
pleasant put pressure on Black workers to mask their feelings of annoyance and frustration
(Wingfield, 2021). Black employees have revealed that emotions like anger are unavailable to
them in the work context because of the pressure to appear poised and avoid being seen as
threatening (Wingfield, 2007, 2010, 2021).
While Black men and women generally attempt to avoid displaying anger, research has
shown that Black women are more likely to strategically display anger to be taken more
seriously (Wingfield, 2021). Wingfield (2007) found that Black women often had more agency
in vocalizing their mistreatment because they were perceived as less threatening, whereas Black
men, who have little recourse for mistreatment, tended to repress any emotions that could be
construed as angry or confrontational. Black men perceived greater risk in drawing attention to
their coworkers’ and employers’ racism, so they frequently became emotionally detached at
work (Wingfield, 2007). For Purdie-Vaughns and Eibach (2008), non-prototypical members of a
social group experience social invisibility, which has advantages and disadvantages. The more
representative you are of a social category, the more tightly the group stereotype will be applied,
creating important nuances in the expectations society holds for people at the intersections of
identities (Hall et al., 2018; Purdie-Vaughns & Eibach, 2008; Smith et al., 2019). Additionally,
certain situations may activate different dimensions of our identities and create opportunities for
either strategic assimilation or differentiation based on the context (Hall et al., 2018; Razzante et
al., 2021). Wingfield (2021) argued, “for a Black woman attorney, race, gender, and class
overlap in ways that will necessitate different kinds of emotional labor depending on whether her
57
workplace interactions involve White women colleagues, White men judges, or Black men
clients” (p. 200).
Research Questions & Hypotheses
Existing research suggests we should expect to see meaningful differences between
groups in the frequency and variety of emotions expressed, amplified, and suppressed on the job
based on the stereotypes activated in different contexts. To test these assumptions, I ask (RQ1)
whether or not there are statistically significant differences in the experience of emotional labor
across groups of employees at the intersection of different gender and racial identities. It is
expected that the experience of emotional labor should differ as a function of one’s identity, with
those socially located at the intersection of multiple marginalized identities (i.e., women of color)
being more likely to engage in taxing forms of emotional labor on the job given their relative
power/status disadvantage. Next, I explore (RQ2) whether or not structural elements of the
workplace, including job characteristics, influence the amount of emotional labor experienced
by different groups. Lastly, 60 qualitative interviews and over 6,000 open-ended diary entries
were analyzed using a grounded approach to uncover (RQ3) how different groups describe
emotional labor on the job.
While specific hypotheses for racial subgroups are beyond the scope of this paper, prior
scholarship on gendered and racialized emotional labor in organizations suggests we should
anticipate:
H1 White women, men of color, and women of color will engage in more emotional labor
(amplification and suppression) than White men.
H2 Employees of color will engage in more suppression of emotions than their White
peers.
H3 Women of color will strategically amplify negative emotions more than men of color.
58
H4 Men of color will suppress emotions more than other groups.
H5 White women, men of color, and women of color will be more sensitive to structural
elements of the workplace and job characteristics than White men and will leverage
more emotional labor when perceptions of procedural and interpersonal constraints are
high.
H6 White men, White women, men of color, and women of color will describe different
experiences of emotional labor.
DATA, METHODS, AND MEASURES
Diary Design
A diary method was used to collect daily accounts of emotional labor among a sample of
181 local government employees from a large metropolitan area in California. The initial sample
consisted of 28% White men, 24% White women, 20% men of color, and 28% women of color.
Workforce population data was only available for the City of Los Angeles, where men accounted
for 71% and women accounted for 29% of the workforce, which is skewed relative to the general
population of Los Angeles County, which is 51% women and 49% men (U.S. Census Bureau,
2020). Additionally, those who identify as Hispanic make up 48% of the population of Los
Angeles County, followed by 26% White, 15% Asian, 8% Black, and 3% identifying as two or
more races (U.S. Census Bureau, 2020). Consistent with previous diary method scholarship with
a sample of this size, response rates to the daily questionnaire varied between 38% and 77%
(Myin-Germeys et al., 2021; Ohly et al., 2010). A series of logistic regression models predicting
missed days confirmed there was no systematic attrition based on gender or race (see Appendix
2A). Daily diaries were administered through an online survey tool and were designed to capture
the emotional labor of employees for each particular workday. The data was collected
59
successively in the period April 13–30, 2020,
19
during which participants received a daily set of
prompts asking about their interpersonal interactions and emotional labor.
The majority of research on race and emotional labor has been qualitative, and Wingfield
(2021) called for additional quantitative work at a broader scale to document how gender and
race simultaneously shape emotion management. To establish a baseline and explore the
frequency and variety of emotional labor experienced by local government employees, I provide
descriptive statistics showing the frequency with which nine discrete emotions were either (1)
expressed when genuinely felt, (2) suppressed, or (3) amplified over the course of the 3-week
diary. These descriptive statistics were broken down by four demographic subgroups (i.e., White
men, White women, men of color, and women of color).
20
Next, I compared the single-axis
model predicting emotional labor, which treated race and gender as separate independent
variables, to a double-axis model that included an interaction term to account for the intersection
of race and gender. A series of multilevel models were run separately to identify if perceived
workplace red tape, the emotional climate, and the frequency of interpersonal interactions were
predictors of emotional labor for White men, White women, men of color, and women of color.
Lastly, I used systematic content analysis of semi-structured interviews and open-ended diaries
to inductively derive what was different about the nature of emotional labor across groups in
respondents’ own words.

19
Note: This data was collected at the onset of the COVID-19 pandemic, which may limit the
generalizability of findings.
20
A limitation of the data is the small sample size for specific race-gender intersections, which
necessitates aggregating racial minorities into a White–non-White binary to preserve statistical
sampling power.
60
Dependent Variable—Emotional Labor
A modified version of the Discrete Emotions Emotional Labor Scale (DEELS; Glomb &
Tews, 2004) was used to measure emotional labor. The original scale measures 14 discrete
emotions, with questions about the frequency with which those emotions are (1) genuinely
expressed, (2) amplified, or (3) suppressed. Because this battery of questions was administered
daily, imposing considerable burden on participants, the number of discrete emotions was
reduced to nine (i.e., contentment, concern, happiness, enthusiasm, fear, anxiety, sadness,
irritation, and anger). The frequency of each emotion was coded on a four-point scale from “zero
times” to “many times” each day.
Independent Variables of Interest—Intersectional Identity
Gender and racial/ethnic identities were initially treated as separate independent variables
in a single-axis model. Subsequent double-axis models used an interaction term to highlight the
intersection of race and gender as a simultaneous identity. The interaction term combined
dummy variables for women
21
(vs. men) and employees of color (vs. White), yielding four total
combinations.
Control Variables
At the individual level, all models controlled for age, which was broken down into three
generational cohorts: (1) baby boomers, (2) Generation X, and (3) millennials or younger.
22

21
While a non-binary gender identity option was available to participants in the survey
instrument, this particular sample only identified as either women or men.
22
Baby boomers were born in 1946–64, Generation X in 1965–80, millennials in 1981–96, and
Generation Z in 1997–2012. Given the small sample size of Generation Z employees in the
workforce, they were combined into the millennials or younger category.
61
Additional controls for prosocial motivation
23
and work–home boundary conditions
24
did not
improve model fit, so were left out of the final models. Agency fixed effects were included to
account for within-agency differences. To get a sense of the rules in the work environment, the
model incorporated a red tape factor (eigenvalue = 1.22) derived from Borry’s (2016) Three-Item
Red Tape (TIRT) Scale. The red tape factor helps to control for the perceived rigidity of rules in
the organization and accounts for (the lack of) job autonomy, a known antecedent of emotional
labor (Grandey et al., 2013). To understand the extent to which employees felt comfortable
expressing emotion authentically, the model included a climate of authenticity factor (eigenvalue
= 4.17) composed of seven-items from Grandey et al. (2012). The authenticity factor helps to
control for the implicit and informal organizational expectations surrounding emotional
expression. Two variables were collected daily and were included to account for the volume of
interactions participants had both externally with the public and internally with team members.
Each day, participants were asked how often they interacted with both groups. An additional
item from the Areas of Worklife Survey was used to measure the extent to which employees felt
supported by their colleagues (Leiter & Maslach, 2011). Lastly, two items asking the extent to

23
Prosocial Motivation Controls: The models were run with and without a prosocial motivation
factor (eigenvalue = 2.02), composed of four items that assessed the extent to which an employee
identified with their role, making them more likely to experience emotions congruent with
explicit and implicit role expectations (Humphrey et al., 2015). Results were not sensitive to the
inclusion of prosocial motivation, so this factor variable was omitted from the final model.
24
Work–Home Boundary Controls: To account for the potential spillover effects of personal
and family life obligations as a source of emotional labor, the models were run with a series of
items combined into a factor representing preferences surrounding either the segmentation or
integration of work and family obligations (eigenvalue = 2.17). The factor comprised three items
from Kreiner’s (2006) measuring segmentation preferences (Kreiner, 2006), three items
measuring work/family role integration-blurring (Desrochers et al., 2005), four items measuring
work–family and family–work conflict (Netemeyer et al., 1996), and four items measuring
integration/segmentation enactment behaviors (Kossek et al., 2006). This work–home factor
variable was not statistically significant in any of the models and did not improve model fit, so it
was not included in the final models presented here.
62
which employees felt their gender and ethnic identities were well represented in their department
were included to account for endogeneity derived from certain demographic groups self-
selecting into departments with more representation of their identities.
QUANTITATIVE RESULTS
Across all participants, positively valenced emotions (e.g., happiness and enthusiasm) were most
frequently amplified, negatively valenced emotions (e.g., irritation and anxiety) were most
frequently suppressed, and a mix of happiness and irritation were genuinely expressed most often
(see Appendix 2C). A series of Pearson’s chi-square tests and unconditional models clustered at
the individual were run to disentangle whether White men, White women, men of color, and
women of color have different tendencies to amplify, suppress, and genuinely express emotions.
The results reveal statistically significant differences across all forms of emotional labor at the
5% level, with all but three having p-values<0.001.
25

The mean values for emotional amplification (Figure 2.1) and emotional suppression
(2.2) for nine discrete emotions are broken down by intersectional identity. Figure 2.1 shows all
groups had a tendency to amplify positive emotions (i.e., enthusiasm, happiness, and
contentment). In particular, men of color had the highest mean values for amplified enthusiasm,
happiness, and contentment. It is also noteworthy that White women had the lowest mean values
for amplifying most negative emotions and did not report any amplification of anxiety, fear, or
anger. Women of color also amplified enthusiasm more often than White men. The takeaway is
that only employees of color seemed to amplify their emotions more often than White men, and
only when it came to projecting positivity in the workplace. Higher overall mean values for

25
Amplification of anger (p<0.053), amplification of sadness (p<0.054), and amplification of
anxiety (p<0.002) were likely less significant because they occurred less frequently overall than
most other forms of emotional labor.
63
amplified enthusiasm, happiness, and contentment imply that men of color may feel more
pressure to project positivity than other groups, and for women of color, this seems to be the case
for enthusiasm. While the mean values for amplifying negative emotions were quite low for all
groups across the board, women of color appeared to amplify concern more than any other
group. This provides some partial support to H3, that women of color may strategically leverage
amplification of certain negative emotions, whereas we do not see the same patterns among
White women or men of color.
Figure 2.1: Group Means for Amplified Emotions
Unconditional Models Predicting Likelihood of Amplifying Emotions (ref: White Men)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women - - - - - - ns ns ns
men of color ns ns ns ns ns ns ns ns ns
women of color ns ns ns ns ns ns + ns ns

64

All groups reported higher means for suppressing emotions than amplifying emotions,
suggesting that emotional restraint is a common technique, especially for emotions like irritation.
Figure 2.2 shows that women of color had the highest mean values for suppressing negative
emotions, but all groups reported high levels of suppressing irritation. It is also noteworthy that
men of color had the highest means for suppressing positive emotions, while White women
appeared to suppress positivity least often. Taken together it is surprising to see that the mean
values for emotional suppression are not higher for men of color compared to the other groups.
This suggests that there is only weak support for H2, H3, and H4, since it is only women of color
with higher mean values for suppression, and only in the case of negative emotions. While there
were no explicit hypotheses about the genuine expression of emotion, Appendix 2D provides a
window into the specific emotions that different groups felt it was safe to authentically express.
Taken together, these results signal that emotional tendencies vary by discrete emotions and by
intersectional identities.

65
Figure 2.2: Group Means for Suppressed Emotions
Unconditional Models Predicting Likelihood of Suppressing Emotions (ref: White Men)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women ns ns ns ns ns ns - ns ns
men of color ns ns ns ns ns ns ns ns ns
women of color ns + + + + + ns ns ns

To explore between-group differences in more depth, 12 OLS models predicting
amplification, suppression, and genuine expression of positive and negative emotion were
analyzed. Table 2.1 shows the differences between a single-axis (2.1A) and a double-axis (2.1B)
approach to modeling emotional labor. The main difference between models is the inclusion of
66
an intersectional identity interaction in the double-axis approach. Controls for interpersonal
interactions, age, red tape, climate of authenticity, social support, ethnic and gender
representation, and agency fixed effects (ጊ) were included in all single- and double-axis models.
Single-Axis
Emotional Labor = !
!
+ !
"
(gender) + !
#
(race) + !
$
(avg colleague interactions) + !
%
(avg
resident interactions) + !
&
(generation) + !
'
(red tape) + !
(
(climate of authenticity) + !
)
(social
support) + !
)
(ethnic rep) + !
*
(gender rep) + ጊ + ɛ
Double-Axis
Emotional Labor = !
!
+ !
"
(intersectional identity) + !
#
(avg colleague interactions) + !
$
(avg
resident interactions) + !
%
(generation) + !
&
(red tape) + !
'
(climate of authenticity) + !
(
(social
support) + !
)
(ethnic rep) + !
*
(gender rep) + ጊ + ɛ

Table 2.1: Single-Axis vs. Double-Axis OLS Models of Emotional Labor
Amplified Pos Amplified Neg Suppressed Pos Suppressed Neg Genuine Pos Genuine Neg
Single-Axis
women -0.172 0.152 -0.151 -0.353** -0.120 -0.128
(0.125) (0.227) (0.159) (0.172) (0.194) (0.200)
employees of color 0.508** -0.220 0.334 0.112 0.203 0.216
(0.220) (0.269) (0.257) (0.242) (0.197) (0.298)
Obs 937 937 937 937 937 937
R-squared 0.613 0.462 0.546 0.597 0.611 0.495
Double-Axis
White women -0.00590 -0.128 -0.201 -0.473*** -0.0641 -0.522***
(0.151) (0.172) (0.186) (0.154) (0.210) (0.176)
men of color 0.851** -0.799* 0.231 -0.134 0.317 -0.596*
(0.409) (0.473) (0.347) (0.338) (0.404) (0.349)
women of color 0.353 -0.0962 0.178 -0.253 0.0889 0.0480
(0.241) (0.264) (0.317) (0.286) (0.271) (0.347)
Obs 937 937 937 937 937 937
R-squared 0.622 0.486 0.547 0.602 0.612 0.548
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
The single-axis approach masks within-group differences, which means the significance
we see among women or employees of color may be misleading. When we look at racialized
emotional labor, the single-axis models in Table 2.1 show that employees of color appear to
engage in more amplification of positive emotions than their White peers. By contrast, the
double-axis models show the prevalence of emotional labor among White women, men of color,
and women of color, with non-Hispanic White men as the omitted reference group. In the
double-axis model, it was only men of color, not women of color, who amplified positive
emotions more often. Men of color also were also less likely to amplify or genuinely express
negativity, but the results for women of color were not statistically significant. Men and women
67
of color appear to have important differences when it comes to how they handle their emotions,
findings that were obscured in the single-axis models.
When looking at gendered emotional labor, the single-axis models suggest women as a
group appear to engage in less suppression of negativity compared to their male counterparts.
The consistencies for gender across the single- and double-axis models only apply to the
suppression of negative emotions for White women. The double-axis model shows that women
of color do not follow the same trends for suppressing negativity, and White women are also less
likely to genuinely express negativity. These double-axis results confirm that treating race and
gender as separate categories of analysis will obscure important differences among women and
employees of color.
While the original models looked at positive and negative emotions, I disaggregated
emotional labor into discrete emotions for a more nuanced analysis to understand whether the
between-group differences in amplification and suppression were uniform across discrete
emotions. Table 2.2 takes the same double-axis approach combining gender and race to explore
discrete emotions that are either amplified or suppressed as the dependent variables, with White
men omitted as the reference group. For ease of interpretation, Table 2.2 only shows covariates
for intersectional identity but accounts for the same controls as previous models.
26

Emotional Labor = !
!
+ !
"
(intersectional identity) + !
#
(avg colleague interactions) + !
$
(avg
resident interactions) + !
%
(generation) + !
&
(red tape) + !
'
(climate of authenticity) + !
(
(social
support) + !
)
(ethnic rep) + !
*
(gender rep) + ጊ + ɛ

26
Full model results for double-axis OLS models predicting discrete emotion, including controls,
are available in Appendix 2F.
68
Table 2.2: Double-Axis OLS Models Predicting Discrete Emotions
Amplified Emotions
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.115 -0.0266 -0.0141 -0.107 -0.0271 -0.0263 -0.00505 -0.0736 0.102
(0.0763) (0.0576) (0.0468) (0.0746) (0.0541) (0.0447) (0.114) (0.113) (0.101)
men of color -0.403** -0.262 -0.202 -0.140 -0.244 -0.193 0.565* 0.582* 0.597**
(0.178) (0.162) (0.131) (0.175) (0.154) (0.120) (0.302) (0.307) (0.253)
women of color -0.118 -0.0269 -0.000847 0.104 -0.0302 -0.0387 0.402** 0.167 0.210
(0.121) (0.0874) (0.0746) (0.145) (0.0821) (0.0652) (0.200) (0.178) (0.147)
Obs 937 937 937 937 937 937 937 937 937
R-squared 0.491 0.481 0.529 0.559 0.488 0.495 0.630 0.599 0.684
Suppressed Emotions
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.302* -0.497*** -0.439*** -0.315** -0.378*** -0.416** -0.0767 -0.137 -0.189
(0.156) (0.131) (0.159) (0.126) (0.142) (0.162) (0.112) (0.113) (0.131)
men of color -0.547* -0.229 -0.189 -0.0517 -0.00455 -0.0343 0.241 0.129 -0.0525
(0.301) (0.317) (0.304) (0.297) (0.315) (0.305) (0.199) (0.206) (0.263)
women of color -0.418* -0.214 -0.208 -0.00253 -0.0681 -0.502* 0.145 0.0545 0.157
(0.246) (0.268) (0.293) (0.228) (0.259) (0.269) (0.186) (0.187) (0.218)
Obs 937 937 937 937 937 937 937 937 937
R-squared 0.628 0.673 0.559 0.632 0.574 0.554 0.571 0.556 0.542
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
For amplified emotions, Table 2.2 shows that employees of color, and men of color in
particular, were more likely to amplify positive emotions than White men. Men of color
amplified enthusiasm, happiness, and contentment more than White men and were less likely to
amplify irritation. Among women, only women of color appeared to amplify enthusiasm more
often than White men. Taken together, these results suggest men of color, and to a lesser extent
women of color, were amplifying positive emotions more frequently than White men, which
suggests they may be compensating for negative racialized stereotypes by emphasizing
positivity.
Counter to expectations, all groups were less likely than White men to suppress
negativity. Across all six negative emotions measured (i.e., irritation, anxiety, sadness, concern,
fear, and anger), White women were less likely than White men to suppress their emotions. It
69
was anticipated that there would be added pressure to suppress negative emotions to manage
racialized stereotypes, but there were no statistically significant relationships suggesting negative
emotions were more often suppressed by employees of color. Among employees of color,
women of color were less likely to suppress irritation and fear, and men of color were less likely
to suppress irritation. These findings also run counter to the group means revealed previously,
where women of color saw higher mean values for suppression of negative emotion. The
unanticipated results may stem from the possibility that different groups may have different
baselines for experiencing emotions in the first place, which would affect their emotional labor,
and which the instrument used was unable to capture. Additionally, these results potentially
signal that different job- and organization-level variables may be more prevalent among men and
women of color, warranting further exploration.
A series of multilevel models were run to better understand what contributes to variation
in the suppression of emotions among White women, men of color, and women of color. At
level 1, the models included time-variant variables for daily interaction with colleagues and
residents. At level 2, the models controlled for intersectional identity, generation, and individual
perceptions of red tape, a climate of authenticity, and social support. The dependent variables for
the models were interchanged to reflect the suppression of nine discrete emotions, and all models
included agency fixed effects (ጊ
!
) and day fixed effects ("
"
). Table 2.3 presents the coefficients
for three sets of interactions between intersectional identity and perceived (a) climate of
emotional authenticity, (b) social support, and (c) red tape to assess which groups were most
sensitive to these organizational and job-related variables. Each interaction term was modeled
separately for parsimony, and full model results including all controls are available in Appendix
2G.
70
'%((&)**)+ ,-./0.1
,+
= #
!!
+ #
"!
(interaction colleagues) + #
#!
(interaction residents) +
#
!"
(intersectional identity) + #
!#
(generation) + #
!$
(red tape) + #
!%
(climate of authenticity) +
#
!&
(social support) + #
!'
(intersectional identity*[climate of authenticity, social support, red tape])+ ጊ
+
+ "
,
+
%
!+
+ &
,+

Table 2.3: Key Interactions in MLM Predicting Suppressed Emotions
Authentic Climate x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women 0.207 -0.0637 0.186 0.00497 0.0823 -0.0649 0.127 -0.00768 -0.116
(0.209) (0.202) (0.195) (0.185) (0.171) (0.186) (0.135) (0.127) (0.171)
men of color -0.159 0.00567 -0.224 0.0184 -0.0445 -0.201 -0.0890 -0.303* 0.0769
(0.289) (0.273) (0.264) (0.253) (0.234) (0.254) (0.183) (0.174) (0.227)
women of
color
-0.0803 -0.265 -0.475** -0.253 -0.564*** -0.542*** -0.325** -0.298** -0.431**
(0.215) (0.213) (0.206) (0.198) (0.183) (0.198) (0.143) (0.136) (0.176)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90

Social Support x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women 0.150 -0.107 0.283** -0.170 0.101 -0.245** -0.0507 -0.111 -0.258***
(0.149) (0.139) (0.139) (0.130) (0.121) (0.124) (0.0889) (0.0881) (0.0927)
men of color -0.105 -0.192 -0.0239 -0.261* -0.00715 -0.493*** -0.301*** -0.364*** -0.491***
(0.185) (0.165) (0.164) (0.153) (0.143) (0.147) (0.105) (0.104) (0.110)
women of
color
-0.0463 -0.412** -0.377** -0.530*** -0.497*** -0.915*** -0.564*** -0.501*** -0.817***
(0.183) (0.181) (0.180) (0.168) (0.157) (0.161) (0.115) (0.114) (0.120)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90

Red Tape x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.180 -0.166 -0.510** 0.110 -0.272 -0.0202 0.338*** 0.191 0.563***
(0.211) (0.207) (0.222) (0.187) (0.177) (0.212) (0.129) (0.130) (0.180)
men of color -0.779** -0.423 -0.769** -0.468 -0.525* -0.447 -0.488** -0.122 0.217
(0.368) (0.340) (0.362) (0.313) (0.295) (0.347) (0.216) (0.217) (0.292)
women of
color
-0.0805 0.588* -0.0873 0.312 0.521* 0.155 0.199 0.241 0.266
(0.342) (0.347) (0.361) (0.326) (0.307) (0.350) (0.225) (0.226) (0.285)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
A climate that welcomes emotional authenticity and where employees feel socially
supported by their colleagues should allow for more freedom of expression and less pressure to
suppress emotions. Table 2.3 reveals that women of color who perceived that the workplace
climate allowed for emotional authenticity were less likely to feel the need to suppress the
majority of the emotions measured (i.e., sadness, fear, anger, enthusiasm, happiness, and
contentment) when compared to White men. A climate of authenticity was far less influential for
71
men of color, who were only less likely to suppress happiness compared to White men. For
White women, none of the interactions were significant, suggesting a workplace climate that
allows for emotional authenticity is not influential in their decisions to suppress or not suppress
emotions. We see similar patterns for social support, where women of color who perceived high
levels of social support were less likely to express nearly all emotions (with the exception of
irritation) compared to White men. Perceptions of social support were less influential for men of
color, who tended to suppress the three positive emotions less and two negative emotions (i.e.,
concern and anger) less when social support was high compared to White men. White women
who perceived high levels of social support were only less likely to suppress contentment and, to
a lesser extent, anger. Curiously, among White women, high levels of social support contributed
to more suppression of sadness compared to White men. Taken together, these results make it
clear that women of color were by far most sensitive to perceptions of a climate of emotional
authenticity and social support, and workplaces that foster those conditions may help reduce the
amount of emotional suppression women of color take on.
An environment characterized by high levels of red tape should lead to more behavioral
constraints, meaning more emotional suppression when red tape is high. In line with this
thinking, Table 2.3 reveals that women of color who perceived high levels of red tape were more
likely to suppress negative emotions of anxiety and fear compared to White men. White women
who perceived high levels of red tape were more likely to suppress positive emotions like
enthusiasm and contentment but were actually less likely to suppress their sadness than White
men. Counterintuitively, men of color who perceived high red tape were less likely to suppress
irritation, sadness, fear, and enthusiasm than White men. These results reveal that rigid
72
procedural environments with rules that are burdensome, unnecessary, and ineffective will be
most consequential for women of color, leading them to suppress negative emotions more often.
QUALITATIVE RESULTS
The third research question asks how different groups describe their emotional
experiences. To theorize emotional labor through an intersectional lens, participants were asked
to describe their emotional experience on the job in semi-structured interviews and daily diary
prompts. In analyzing responses from 60 qualitative interviews and over 6,000 diary entries,
women and people of color, regardless of their position, described techniques to manage
emotions in interpersonal relationships in more detail than their White male counterparts.
In total 6,776 diary entries in response to 10 distinct open-ended diary prompts were
analyzed (see Appendix 2B). White women contributed the most individual entries (1,880),
accounting for roughly 28% of all responses, followed by White men (1,790) at 26.4%, women
of color (1,735) at 25.6%, and men of color (1,371) at 20% (see Figure 2.2). The entire free
response entry for each open-ended question was used as the unit of analysis for diary coding,
which allowed for incident-by-incident coding (Charmaz, 2014). It is important to note that the
descriptions below reflect the themes that were most common for each group but are not meant
to suggest they apply uniformly to all White men, White women, men of color, or women of
color. There were certainly individuals who provided exceptions to the themes below, but in
aggregate, these were the patterns that emerged with the highest frequency in the data collected.

73
Figure 2.3: Percentage of Total Diary Entries by Intersectional Identity

Using ATLAS.ti software, a coding scheme was developed based on code families after
an initial round of grounded coding produced over 100 unique codes (Campbell et al., 2013;
Charmaz, 2014; Friese, 2019). The second phase of coding involved sorting diary entries into
those explicitly mentioned emotions or emotional events. Within the primary code of emotions
and emotional events, I developed secondary codes to sort and integrate data into thematic
clusters for each group (Charmaz, 2014). The three secondary codes of interest for this analysis
were mentions of negative emotions and experiences (e.g., anger, sadness, bad day, high stress,
etc.), positive emotions and experiences (e.g., happiness, optimism, good day, low stress, etc.),
and emotional labor (e.g., suppressed emotion, amplified emotion, deep acting, etc.). For
example, an entry like “frustration with a coworker who just doesn’t listen” would have been
assigned a primary code of emotional events along with secondary code of negative emotions.
Table 2.4 shows the key terms that were inductively derived to develop the codes for negative
and positive emotions and experiences, along with the deductively established emotional labor
constructs of surface and deep acting as defined by Grandey et al. (2013), with more examples of
specific quotes available in Appendix 2H. This same coding scheme has high face validity and
74
was applied to the 60 interview transcripts to supplement the diary analysis. Descriptions of the
antecedents and consequences of emotional events are available in Appendix 2I.
Table 2.4: Key Terms for Establishing Secondary Codes
Negative Emotions & Experiences

Anger, Frustration, Irritation, Annoyance, Anxiety,
Concern, Guilt, Worry, Fear, Overwhelmed, Sadness,
High Stress, Bad Day, Distress, Depression, Empathy,
Poor Physical Health, Poor Mental Health, Poor
Relational Health, Exhaustion, Burnout

Positive Emotions & Experiences

Happiness, Joy, Contentment, Hope, Enthusiasm,
Optimism, Low Stress, Good Day, Good Physical
Health, Good Mental Health, Good Relational Health,
Job Satisfaction,
Surface Acting

Suppressing Emotions
(e.g., minimizing negativity)
Amplifying Emotions
(e.g., emphasizing positivity)
Deep Acting

Attentional Deployment
(e.g., focusing attention or emphasis on controllables),
Cognitive Change
(e.g., reappraisal or reframing the narrative),
Situation Selection
(e.g., modifying or avoiding encounters)

Emotional Events
Of the diary entries, 3,076 described emotional events. Sixty-two percent of entries
described negative emotional experiences, compared to only 38% describing positive emotional
experiences. Among those entries, an additional 29% made reference to using a form of
emotional labor.
Table 2.5: Emotional Events
White Men
(997)
Men of Color
(695)
White Women
(1,136)
Women of Color
(1,174)
62% Negative Emotions 59.5% 53.2% 62.7% 69.6%
38% Positive Emotions 40.5% 46.8% 37.3% 30.4%
29% Emotional Labor 19.5% 21.0% 21.5% 27.7%
Experiences of Negative & Positive Emotions
Table 2.5 reveals that all groups reported more negative emotions than positive, but
women of color had the highest percentage of emotional event entries dedicated to negative
75
emotions at 69.6%. As anticipated, men of color were less likely to explicitly reference feelings
of anger compared to any other group, naming “frustration” more than any other negative
sentiment. It is also worth noting that women of color and, to a lesser extent, White women were
much more likely to report feelings of anxiety than men. Men of color actually had the highest
percentage of emotional event entries dedicated to positive emotions at 46.8%, followed by
White men, White women, then women of color.
Emotional Labor
Importantly, the diary entries provide much richer information than the closed-ended
survey items about the ways participants managed their emotions in their own words. Notably,
women of color referenced engaging in emotional labor in a larger percentage of emotional event
entries than any other group at 27.7%, followed by White women and men of color at roughly
21% and White men at 19.5%. Open-ended prompts allowed participants to share some of the
ways in which they engaged in deep acting by either modifying the situation, shifting their
attention, or reframing their attitudes about events to experience more positive emotions and
avoid negative ones. The open-ended responses also provided an opportunity for respondents to
share a wider range of emotions beyond the nine discrete emotions measured in the quantitative
analysis.
Surface Acting. All groups described a pressure to present as positive, engaged, and
caring on the job while suppressing negative attitudes, irritation, and anxiety. This proved to be a
contrast from the quantitative analysis, where it was predominantly men of color, and to a lesser
extent women of color, who amplified positivity more often than White men, and women and
employees of color appeared to suppress emotions less often than White men. Across groups,
there were similar comments about “trying to remain positive” (17:583), “maintaining an
76
uplifting tone” (4:2253), “sp[eaking] calmly about what we had done and are doing” (5:492), and
“pretending to care” (4:1480) in service of “bolstering morale of others” (4:1356). For all groups,
suppressing negative emotions seemed to be tied to notions of professionalism, with participants
sharing they might want to tell their colleagues how they really feel, but they “need to keep it
professional” (4:1159) or “just don’t think there’s enough time with all the responsibilities that
my division handles” (1:606).
Amplifying Positivity. Among the women in the sample, there were frequent references
to the need for “diplomatic” styles of communication that could mask underlying frustrations.
Women described suppressing irritation to make supervisors feel more comfortable, emphasizing
their attunement to the emotions of others. As an example, one woman described growing
increasingly frustrated that she could not speak openly with her boss about her circumstances,
sharing, “I tried to diplomatically convey it, but I can’t really express it—it’s really frustrating”
(2:3). This comment in particular reveals a tension between the way she wanted to convey the
information and how she had to manipulate her tone to be more pleasant and diplomatic. Others
described fearing they would be replaced or passed over for opportunities if they did not adhere
to the emotional expectations of the work environment. One woman shared she often overstated
her confidence and comfort with the task at hand to mask her apprehension about vocalizing
dissatisfaction or tension with colleagues on the job, saying, “the minute I asked for help, maybe
I can now be laid off and that person takes my job forever” (9:7). Taken together, these
respondents reveal a theme of White women avoiding emotionally assertive behaviors, opting for
more passive emotional engagement that comes across as positive, competent, and
“professional.” But this emotional effort came at a cost to their sense of authenticity, with one
77
woman sharing, “I try to keep positive when talking with coworkers, but sometimes I feel as if
I’m putting on a mask” (14:1240).
The women interviewed described feeling like the male-dominated culture shaped
emotional expectations in the workplace. The majority of women of color described challenges
navigating sexist comments and behaviors but noted that they were subject to more stereotypes
than White women. The women of color interviewed described smiling and joking when they did
not want to in order to make those around them feel comfortable. As an example, one woman of
color shared, “sometimes I’m very comedic, and you know, we use that when we’re deflecting,
we use a lot of humor” (24:2). These quotes may reveal that women in the sample felt
constrained by the emotions evaluated in the quantitative analysis, but when given the space to
share from a broader and more general set of positive experiences, they did in fact engage in
amplification of emotion.
Suppressing Negativity. Women in the sample were explicit about the need to manage
the impression they made on others. As an example, one woman shared, “sometimes I feel I
can’t share something because there isn’t time or I have pre-existing stigmas holding me back”
(8:92). Women of color described suppressing their negative emotions for fear of being seen as
overly emotional or reactive. As an example, one woman shared she did not vocalize concerns
over gendered and racialized mistreatment because
Many of us are fearful to honestly express issues that we encounter at work due to
the stigma placed on people who speak up. The rest are just tired of speaking up
because nothing is done about their issues. (9:13)
Other women of color described their ability to suppress emotions as a personal skill they
leverage to manage the demands of the job. One woman of color in particular wrote,
I know everyone has their days, and I can do pretty well masking my feelings.
Others do not have the same ability, so their emotions can rub off and cause
stress in the work environment. (17:497).
78
In their interviews, men of color described suppressing their emotions to navigate
environments where they were underrepresented, sharing there was a heightened pressure to “be
politically correct at all times” (12:1148). One employee of color described his work
environment being dominated by White men with “a pack mentality.” To navigate the pack and
their perceived out-group status, men of color described suppressing frustration to avoid negative
stereotypes, but none described faking feelings to fit in, with one sharing, “if I’m not feeling it,
I’m not going to show it” (15:6). While the quantitative finding did not show men of color
suppressing anger more often, one respondent asked directly if his candor in the interview was
coming across like an angry Black man, saying, “so I just want to know how much it’s coming
up—like the angry Black man, how is it?” (30:9). This inconsistency between quantitative and
qualitative findings may be a function of Black men having a different set of emotional
expectations than other men of color, necessitating additional research beyond the White/non-
White binary. By contrast, White men more often wrote they tended to suppress emotions
because they perceived their colleagues to be overly sensitive. Some White men articulated an
awareness of their own power in the organization and shared concerns about being perceived as
condescending. As an example, one White man shared his challenge was “communicating
effectively with colleagues without coming off as irritated or patronizing” (4:1070). Again, these
quotes run counter to the quantitative analysis, where women and people of color did not appear
to suppress negativity more often than White men. This perhaps reveals that the emotions
measured quantitatively were too limiting, and it would be fruitful to expand the analysis into
feelings of general stress and frustration.
Deep Acting. When it comes to deep acting, diary responses from all groups revealed
three key techniques to avoid experiencing negative emotions and instead trigger positive
79
emotions: (1) shifting attention and/or focus (attentional deployment), (2) mentally reframing
events (reappraisal), and (3) changing the situation (situation selection). These techniques are
consistent with Gross’s (1998) work on emotion regulation, where (1) attentional deployment
involves directing attention in a way that the emotion-response trajectory is altered, (2)
reappraisal involves altering the meaning one makes of a situation in a way that influences the
emotions that situation will produce, and (3) situation selection involves strategically selecting
environments and interactions with the intention of increasing (or decreasing) the likelihood that
certain emotions will arise.
Attention. Women of color described attentional deployment and reappraisal strategies
more often than any other group. As examples, women of color were the group most likely to
describe directing their attention to the things within their control to manage their emotions and
cope with stress. As an example, one woman of color wrote,
I have my moments where I am a little anxious or concerned. I try to use prayer,
exercise, meditation, and/or interact with family and friends to help me when
these feelings get overwhelming and also as a means of prevention. (14:988)
A different woman of color added,
I’m overwhelmed but currently have a mantra that I am not allowing people to
stress me out. I have enough on my plate I am dealing with mentally; other people
cannot be responsible for how I should feel. (14:1650)
White women also made frequent reference to the power of focusing their attention on
the positive, writing things like “I realized that if I think positively about, for example, a
person who is bothering me, instead of focusing on the negative, it makes the situation
better” (1:338) and “we suddenly have a bunch of flowers blooming, and I took a couple
moments to enjoy them—looking for the positive, even if it’s little things” (5:1063).
What seemed to distinguish the responses of women of color was an emphasis on what
was controllable, whereas White women emphasized positivity and/or gratitude.
80
Reframing. Women were also more likely than men to minimize their work-related stress
by putting it in the context of broader problems within society. As an example, one woman of
color described her work-related anxiety as “first-world issues, so I’m not complaining too
much” (17:226). A White woman similarly wrote she was “trying to stay positive because I
know so many have it worse right now” (14:1025). As another form of reappraisal, interviews
with employees of color revealed that their family history of struggle, trauma, and oppression
gave them a perspective that made emotions evoked on the job seem trivial in light of larger
societal challenges. One man shared he rarely felt emotional in a work context in light of his
experiences coming from El Salvador:
There was civil war down there. Running every day, like gangs can kidnap you,
kill you, the government can do it to you. It’s just different. … Here I have options
to make myself safe. I feel like I have some control. (7:3)
Several other employees of color who identified as immigrants described a pressure to perform
and preserve without complaint to take full advantage of the opportunities afforded to them in
this country. One employee described suppressing work-related irritation by focusing instead on
gratitude for the opportunity to work, sharing,
My parents came from Cuba, and we’re the first-born generation [of] American
Cubans here in the States. My father used to always tell us that a man is defined
by the work that he does, so I always took that to heart, and I always told myself
I’ll always try to be the best of what I can be. (37:9)
Similarly, a woman of color interviewed shared a commitment to suppressing negative attitudes
about her job, explaining,
We are immigrants to this country, and this being the land of opportunity, you
make the most of it. We were given opportunities, and we wanted to make sure we
excelled at them because other family members were looking at us, and you don’t
want to be a failure in the eyes of others. (4:6)
White men also referenced reappraisal strategies, but to a much lesser extent than other
groups. Among White men, their reappraisal mostly took the form of “giving people the
81
benefit of the doubt” (4:796) or suggesting others needed to adopt a different mindset to
approach the same situation and achieve better interpersonal results.
Situation Selection. White men more than others referenced avoiding certain
individuals/interactions to manage their emotions on the job. Different White men described
“deciding who to engage with” (4:797) and shared that deliberately selecting the situations in
which they engaged others helped them stay even-tempered. This ability to choose avoidance as
an emotion regulation strategy reflects a degree of job autonomy and power that may not be
equally available to other groups of employees. To a lesser extent, White women also described
avoidance as a strategy for managing their emotions. As an example, one White woman shared,
“today wasn’t too bad, but then again I avoided all the phone calls I needed to make because I
just needed to give myself an easy day” (4:767). It was far less common for men or women of
color to describe avoiding interactions throughout the day, unless they were avoiding someone in
the home environment so they could decompress from the stress of the workday alone.
All groups described being mentally exhausted, but women of color described feelings of
emotional exhaustion more often than other groups, with 30.2% of the entries mentioning
consequences describing a sense of emotional fatigue. Women of color wrote, “my mental health
is wearing thin” (14:1603), “I’m drained in all aspects” (14:1841), and “I arrive home mentally
and emotionally exhausted” (12:987). Importantly, women of color noted more than any other
group that their emotional exhaustion was the cumulative effect of regulating their emotions on
the job and at home. As an example, one woman wrote, “I’m cumulatively exhausted, and this
situation isn’t helping my anxiety or depression” (1:29). Women of color often described being
worn down, unable to recover from the stress of the day, with one writing, “even the strongest
can only take this for so long” (14:10).
82
Table 2.6: Qualitative Themes by Group
Antecedents Events Consequences
ALL
Challenges managing internal
relationships with colleagues as
the biggest source of stress.
Secondary source of stress
comes from managing
competing work–life
responsibilities, particularly
caretaking
Negative experiences are
described more often than
positive ones.
All groups engage in both
surface and deep acting.
Primarily described reduced
feelings of efficacy and
accomplishment, but also an
increase in
cynicism/disengagement, along
with emotional exhaustion.
Women
of
Color
Most likely to reference office
power dynamics.

Personal and professional
responsibilities bleed into one
another.

Describe the most negative and
the fewest positive emotional
events.

Engage in more emotional labor
than any other group, especially
deep acting techniques of
attentional deployment and
reappraisal.
More emotional exhaustion than
any other group.

Incidents of inequity and
mistreatment at work as a source
of their cynicism and
disengagement.
Men of
Color
Most likely to describe internal
relationships as challenging,
especially among leaders who
feel undermined by their staff.
Work stress interferes with their
home life.
Describe the most positive and
the fewest negative emotional
events.

Emphasize the importance of
suppressing negativity to
maintain professionalism.
Incidents of inequity and
mistreatment at work as a source
of their cynicism and
disengagement.
White
Women
Conflict between others within
the office as a source of stress.

Personal and professional
responsibilities bleed into one
another.
Reference the need for
diplomatic styles of
communication, particularly for
navigating sexism.
Heightened sensitivity to
gendered performance
expectations.
White
Men
Describe colleagues as overly
sensitive and emotional.
Personal responsibilities
interfere with work life.

Rationale for suppressing
emotions is that they will be
misunderstood by others.

Leverage situation selection to
avoid engaging their emotions.
Most cynicism and
disengagement of any group.
DISCUSSION
This paper adds to public administration scholarship by making visible the gendered,
racialized, and intersectional forms of emotional labor embedded in public service work. If we
expect women and employees of color, especially women of color, to participate in
representative bureaucracy, we need to acknowledge the emotional labor required. In addition to
83
their standard job demands, this study finds that women of color, White women, and men of
color engage in emotional labor in a myriad of different ways, which carry different
consequences for their mental health, burnout, and ultimately retention of a diverse workforce.
The failure to account for the different manifestations of intersectional emotional labor creates
management blind spots with respect to supporting an increasingly diverse workforce. In
particular, this study highlights that different subgroups at the intersection of racial and gender
identities engage in different surface and deep acting techniques for different sets of discrete
emotions. Thus, theorizing at the aggregated level of surface acting techniques without exploring
norms and expectations for different emotions largely misses the mark.
The results of the quantitative modeling make clear that the experience of emotional
labor differs as a function of one’s identity, with those socially located at the intersection of
multiple marginalized identities (i.e., women of color) being more likely to engage in taxing
forms of emotional labor on the job given their relative power/status disadvantage. I also find
that women of color were more sensitive to structural elements of the workplace, with
environments characterized by high red tape increasing their emotional suppression and
environments characterized by an openness to emotional authenticity reducing their suppression.
These findings support the calls to allow more discretion, autonomy, and personal
responsibility in the emotional self-expression of women of color. It may be fruitful to foster
climates of authenticity and adopt humanistic practices designed to enhance employee well-
being through supportive emotional training, opportunities for emotional recovery, and adequate
compensation for emotional skill and effort (Grandey et al., 2015). Given that women of color
appear to be taking on the most taxing forms of emotional labor, it should be encouraging to see
that there may be tangible things leaders can do to reduce the pressure to surface act by
84
increasing a sense of psychological safety and emotional authenticity in the work group. Recent
work by Hoang et al. (2022) showed that inclusive leadership practices are necessary to facilitate
employees’ desires to balance a sense of belonging while also being able to bring their unique
identities to work. Roberson and Kulik (2007) also argued explicit discussion about stereotypes
can be useful in reducing their impact. Leadership practices that create a sense of belonging and
appreciation of individual differences have been shown to foster trust, communication, and
engagement among employees, enhancing perceptions of organizational justice (Hoang et al.,
2022).
The qualitative findings reveal there is much to be explored when it comes to differences
in deep acting between groups. It appears that perhaps we did not see as much surface acting
among women of color as anticipated in the statistical models because they have developed a
more robust set of emotion regulation techniques to manage the power asymmetries at work
without creating as much emotional dissonance. As seen in Table 2.6 and the extended analysis
in the appendix, all groups described challenges managing internal relationships with colleagues
as the biggest source of emotional stress, with the second-most prominent source of stress being
tensions managing competing personal and professional responsibilities, especially childcare.
When it came to internal relationships, women of color were most likely to reference office
power dynamics, men of color described being undermined by their colleagues, White women
referenced tensions between others as a source of their own stress, and White men often
described their peers as overly sensitive and emotional. In managing the work–life balance,
women cited pressures from home bleeding into work and work bleeding into home. Among the
men, the stress seemed to be one-directional, with men of color describing their professional life
impacting their personal life, whereas White men felt their personal responsibilities interfered
85
with work. It is clear from these findings that more work needs to be done to unpack how
positionality at work, at home, and in society at large influences the emotions and emotional
labor of different groups. Future scholars should explore how the direction of lagged effects
between emotional labor at work and at home may differ for different groups.
In terms of emotional events, all groups described negative events more often than
positive ones and engaged in both surface and deep acting to manage their emotions on the job.
Women of color described the most negative and the fewest positive experiences. They also
described engaging in a wider range of techniques than any other group, especially the deep
acting techniques of attentional deployment and reappraisal. Men of color often emphasized the
importance of suppressing negativity to maintain professionalism, which is a different rationale
than White men, who described suppressing emotions because others would misunderstand
them. This finding extends previous work by Humphrey (2021b), which suggested emotional
labor is tied to norms of professionalism, by highlighting that men of color may be more
sensitive to professionalism expectations than others. A key distinction here is that the different
rationales for suppression may reflect the fact that men of color are socialized to manage their
impressions, whereas there may be less experience reflecting on their own standing among White
men. As another expression of their power, White men also mentioned leveraging situation
selection more than other groups to avoid engaging their emotions.
With respect to the consequences of emotional events, the extended analysis in the
appendix shows all groups emphasized reduced feelings of efficacy and accomplishment but also
referenced an increase in cynicism/disengagement, along with emotional exhaustion. Women of
color in particular were most likely to explicitly name racism and sexism and, perhaps
consequently, also described experiencing more emotional exhaustion than any other group.
86
These findings make it clear that emotional labor is brought about through distinct circumstances
and has a range of consequences for different social groups, both of which have clear
implications for the retention of a diverse workforce.
Limitations & Future Directions
A limitation of this study is that it falls into the White/non-White paradigm, which limits
the ability to fully capture the interpersonal dynamics of public service work. It is important to
acknowledge that men and women of color are not homogeneous groups, and there is a diversity
of personal and professional experiences within subgroups that must be further explored. By
aggregating racial groups into a White/non-White binary, the complex reality is oversimplified,
rendering the unique experiences of different groups of color invisible (Delgado & Stefancic,
2017).
Past research on Asian and Black workers’ emotion management has shown that these
groups of employees draw from their lived experiences and understandings of the racial
dynamics present in society in order to guide their behavior on the job, which has the effect of
creating distinct emotional labor expectations and tactics for different racial groups within the
same organization (Kang, 2003; Wingfield, 2021). The stereotypes associated with Asian women
as model minorities create different sets of binds and freedoms compared to Black women or
Indigenous women (Kang, 2003). It could be the case that some emotions are socially
appropriate for some women of color, but not others. This suggests future research should
consider how emotional labor manifests for other populations, as gendered and racialized
emotions should also be understood in relation to other intersecting forms of privilege and
oppression (e.g., sexual orientation, religion, etc.) that have been shown to shape feeling rule
enforcement in organizations (Ortiz & Mandala, 2021; Wingfield, 2021; Woody, 2021).
87
Additionally, women and men in their interviews revealed a range of non-demographic
and experiential dimensions of their individual identity that they felt influenced their emotional
experience at work. Life experiences as veterans or survivors of sexual assault and battles with
substance abuse shaped the way several employees approached interpersonal interactions.
Invisible forms of neurodiversity like autism and dyslexia influenced other employees’ comfort
with emotional disclosure in meaningful ways. Navigating generational differences as norms for
emotional expression evolved at the societal level created both distance and closeness between
colleagues. All of this is to say, an intersectional approach that extends beyond race and gender
differences is sorely needed in emotional labor scholarship.
Lastly, a key difference between the quantitative and qualitative findings in this paper
may stem from two core data limitations in the way emotional labor is measured using the
DEELS Instrument. (Glomb & Tews, 2004). The first issue is that the scale does not have a way
of capturing baseline levels of emotions experienced, which means we have no way of knowing
if certain groups are experiencing anger or happiness more often than others. This is
consequential for our interpretations of suppressed emotions in particular, as those employees
who infrequently experience anger would not record much suppression of anger, but they may
suppress it more often relative to their baseline experience than others. The second issue
concerns the limited range of discrete emotions measured in the instrument. When participants
were unconstrained in their qualitative diary entries, women and employees of color showed the
anticipated patterns of suppressing negativity and amplifying positivity perhaps because they had
the ability to describe a wider set of emotions and experiences. Future scholars should take these
preliminary findings and work to develop additional measures of emotional labor that can
88
accommodate baseline levels of emotional experiences with a wider set of discrete emotions to
truly disentangle the inequitable distribution of emotional labor across groups in the workforce.

89
CHAPTER III. The Emotional Toll of COVID-19: A Longitudinal Study of Burnout in
Local Government at the Onset of the Pandemic
ABSTRACT
Despite the emotional intensity that accompanies crises, rarely is emotional labor
explicitly discussed as a required aspect of crisis response work. This paper explores workforce
burnout pre-COVID-19 and at the onset of the pandemic to understand what role the inequitable
division of emotional labor plays in contributing to burnout during times of crisis. With data
collected before and during the COVID-19 pandemic, this project has the unique ability to assess
sources of burnout during times of unanticipated organizational change, which helps to inform
our understanding of the well-being of a diverse public sector workforce under conditions of
uncertainty. Findings from multilevel modeling show that employees experienced heightened
burnout during the pandemic, and the suppression of emotion contributed to that burnout, but in
different ways for different groups. In particular, women of color who suppressed negative
emotions were more likely to experience all forms of burnout. Content analysis of open-ended
diary entries and semi-structured interviews revealed participants’ descriptions of emotional
exhaustion, depersonalization, and loss of personal accomplishment. This paper highlights the
importance of factoring emotional labor into the experience of burnout at work while
emphasizing that the relationship between the two varies for individuals of different
backgrounds.
INTRODUCTION
Our general call volume has gone up. The level of irritation and frustration from
both my citizen interactions and among my staff and my workplace has been up.
Everybody’s patience is super thin. When people’s patience is low, tempers flare,
and it’s hard to take that breath in. It’s hard for me and for my coworkers and
employees to stay professional and stay empathetic and maintain that kind of level
of professionalism. The people that are calling us don’t have that kind of baseline
90
level of courtesy, because they’re stressed out from their real lives. Myself and the
other operators that work with me, they’ve all got their own real life stresses and
impacts. So nobody’s operating from this good place where they can give it their
fullest to do their best. The call volume is higher. Everybody is in a bad mood.
Everybody’s just having this hyperinflated sense of sensitivity and offensiveness
and everybody’s super touchy about everything. They don’t have an outlet, they
don’t have somewhere else to put that [emotion]. And that goes for not just the
citizens that call us but the operators too. They’re trapped between home and
work, and their kids are home and their husbands are home, and, like, there’s no
rest from the constant nag of everybody.
On March 11, 2020, the World Health Organization (WHO, 2022) declared COVID-19 a
global pandemic crisis. Since then, there have been over 470 million confirmed cases and 6
million associated deaths worldwide. The COVID-19 pandemic has transformed aspects of
organizational life and drawn attention to issues of occupational health and safety, work–family
dynamics, telecommuting and virtual teamwork, and job insecurity as key leadership and human
resource concerns (Rudolph et al., 2021). Uncertainty can be an aversive state, and the stress
experienced by frontline workers during COVID-19 has been shown to contribute to a decrease
in work engagement, work-related self-efficacy, and sense of job security (Guy et al., 2013; Lin
et al., 2021; Liu et al., 2021; Yoon, 2021).
Prior to the pandemic, a review of public sector human capital research revealed that over
one-third of frontline workers already reported feeling burnt out, and almost half of frontline
workers resign within the first few years of service, creating costly and persistent staff shortages
(Barboza-Wilkes, Le, & Resh, 2022; Linos et al., 2021). For public servants, the pandemic
brought about abrupt shifts in job demands to meet the increased needs of communities in ways
that tested their emotional resilience (Berry et al., 2022). The social distancing requirements had
the effect of disrupting typical service delivery methods and practices while adding new
challenges to managing the combination of work and family responsibilities (Berry et al., 2022).
Self-isolation policies strained relationships that were previously sources of social support and
91
coping, exacerbating health and economic anxieties (Van Bavel et al., 2020), but the outcomes of
the pandemic might not be uniformly negative for the workforce. For some, the transition to
remote work may have created an escape hatch from toxic work environments and internal
relational pressures. Thus, this paper explores (RQ1) the extent to which the onset of the
pandemic exacerbated burnout among local government employees.
Gabriel et al. (2021b) stressed the importance of exploring the emotional complexities of
the pandemic in organizational life. For Gabriel et al. (2021b), emotions like anxiety and fear of
the unknown may be paired with hope and excitement for a different work–life balance, creating
emotional ambivalences that make it difficult to theorize employee well-being and strain. Public
administration scholars have noted that the pandemic has increased the need for more empathy in
crisis response to counterbalance the techno-rational nature and emphasis on efficiency in public
service (Zavatarro & McCandless, 2020). For public organizations in particular, crisis response
often demands managing the range of heightened emotions of the public. Additionally, crisis
response often means tasks become more interdependent across departments and organizations,
suggesting there are increased interpersonal demands as employees build and maintain
collaborative relationships (Ge et al., 2021; Steelman et al., 2021). This raises the question,
(RQ2) to what extent did emotional labor contribute to the experience of burnout during the
onset of the pandemic?
Medina and Azevedo (2021) argued representation of marginalized groups in local
government is critical in terms of sensitively addressing crisis-specific inequities
disproportionately experienced by many already marginalized groups. While the emotional
demands of the job are presumably increasing, Laster Pirtle and Wright (2021) found that
92
structural gendered racism
27
has exacerbated disadvantaged positions for women of color during
the pandemic, calling for intersectional analyses of the resources diverse employees may or may
not have to cope with job-related stress and buffer against burnout. This suggests that those who
are underrepresented in the workforce may be called upon to do more to address inequities in the
community during crisis, while simultaneously having fewer opportunities to recover from the
emotionally taxing work. Thus, the last question explores (R3) whether or not the relationship
between emotional labor and burnout during COVID-19 differs across social groups.
To explore the emotional complexity of the pandemic, I review the literature on
emotional labor, burnout, and workplace compassion to explain why times of crisis may evoke
increased emotional labor demands and when to expect burnout as a consequence of that labor.
Then, I describe the data and methods used, present the results, and discuss the implications for
human resource management.
EMOTIONAL LABOR & BURNOUT
Street-level public service jobs are often characterized as high-touch occupations with
heavy emotional labor demands, making burnout a concern (Guy et al., 2008). Drawing upon the
work of Hochschild (1983), Grandey (2000) defined emotional labor as “the process of
regulating both feelings and expressions for the organizational goals” (p. 97) and proposed
managing observable expressions through suppression or amplification of emotion (surface
acting), managing internal feelings (deep acting), and engaging in the expression of genuine
emotions in organizational settings as the working definition of emotional labor. Although
Hochschild (1983) originally argued both surface and deep acting would be alienating to one’s

27
Laster Pirtle and Wright (2021) defined structural gendered racism as “the totality of
interconnectedness of structural racism and structural sexism in shaping race and gender
inequities” (p. 8).
93
sense of self, more recent work has shown that surface and deep acting are not always beneficial
or harmful (Grandey & Melloy, 2017; Humphrey, 2021). The authenticity of emotional display is
central to whether emotional labor is beneficial for both the employee and the organization
(Grandey et al., 2013). For Rafaeli and Sutton (1987), expressed emotions contribute to either
emotional harmony when congruent with experienced emotion or emotional dissonance when
incongruent with experienced emotion. Surface acting by definition suggests employees’
emotional expressions and actual feelings are at odds (emotional dissonance), and empirical
studies have illustrated how the suppression of negative feelings leads to lower self-authenticity
and emotional exhaustion (Brotheridge & Lee, 2002; Hülsheger & Schewe, 2011; Humphrey,
2021). More recently, Sayre et al. (2021) showed that surface acting drains emotional and
cognitive resources, creating resource loss spirals that are difficult to overcome. By contrast,
when employees use deep acting strategies, their sense of authenticity is not always
compromised, and there may be some positive benefits for employee well-being (Hülsheger &
Schewe, 2011; Humphrey, 2021).
Guy et al. (2008) argued, “when the workplace does not recognize the human side of
work and the emotional labor required to effectively perform this work, then the risk of burnout
grows” (p. 119). Burnout is defined as “a psychological syndrome in response to chronic
interpersonal stressors on the job” and is characterized by three dimensions: (1) cynicism (also
depersonalization/detachment), (2) emotional exhaustion, and (3) reduced personal
accomplishment (Maslach, 1982; Maslach et al., 2008). Cynicism manifests as a detached
response reflecting lowered emotional involvement in work, exhaustion refers to overextending
and depleting emotional resources, and reduced accomplishment reflects feeling a lack of
achievement in work (Maslach et al., 2008).
94
Crisis, Compassion, and Increased Emotional Labor
As residents navigate their own emotional reactions to both the public health and
economic crises, local government employees must communicate compassionately to maintain
public trust. Recent work by Wee and Fehr (2021) showed that compassionate behavior among
team members can mitigate some of the individual suffering brought on by the hardships of the
COVID-19 pandemic. Compassion occurs when suffering is met with a caring response, making
compassion an emotion-laden interpersonal process (Cameron et al., 2003; Dutton et al., 2014).
Most empirical research on compassion has explored the healing benefits for sufferers, but
compassion can also impact those empathizing and providing the compassionate response
(Dutton et al., 2014).
For public servants, engaging in compassionate behavior can lead to a sense of
satisfaction and has been associated with a prosocial identity, but engaging in too much
compassion may lead to compassion fatigue and secondary trauma from prolonged exposure to
others’ suffering (Cameron et al., 2003). Dutton et al.’s (2014) work suggested that when
suffering is outwardly expressed, the focal actor (public employee) engages emotional skills to
(1) notice the suffering, (2) try to feel empathic concern, and (3) act to alleviate the suffering—
all while both parties continue to engage in ongoing sensemaking. During times of crisis,
sensemaking may be an especially taxing interpersonal process, with individuals collectively
trying to gain clarity amid ambiguous and unpredictable experiences. The collective
sensemaking challenge may become more difficult as some employees abruptly shift to remote
work. Mandates to telework may influence their sense of task-related self-efficacy and create
new norms of technologically mediated emotional expression (Adamovic et al., 2021; Brodsky,
2021).
95
Additionally, individuals may simultaneously experience happiness and pride in serving
the public as essential workers at the same time as fear about whether their role on the front lines
may be exposing them to the virus, suggesting emotional ambivalences and conflicting goals
(Gabriel et al., 2021a). According to Grandey et al. (2013), “multiple goals will almost always
exist and may vie for attention and implementation during a particular episode of emotion
regulation” (p. 49). Within the hierarchy of goals, there is often a tension between goals that
reflect personal desires and the needs of society, both of which contribute to a sense of identity
(Carver & Scheier, 1998). With multiple goals held simultaneously, sometimes unconsciously,
feelings may emerge that seem inappropriate or inappropriately strong, triggering the cue to
regulate outward emotional expressions, particularly when there is a sustained professional
expectation of providing compassionate public service (Carver & Scheier, 1998).
The Dissonance–Burnout Relationship Through Conservation of Resources (COR) Theory
Hobfoll’s (1989) conservation of resources (COR) theory is a stress model that suggests
emotion regulation strategies are driven by attempts to gain, conserve, or protect resources.
According to COR, individuals have finite psychological, emotional, and physical resources, and
the resources used for one task limit the resources available for other undertakings (Kammeyer-
Mueller et al., 2013). When resources are threatened or lost, anxiety and distress begin to
increase (Hobfoll, 2002; Jeung et al., 2018). Applied to emotional labor, the effort of regulating
one’s emotions is believed to contribute to the depletion of regulatory resources associated with
self-control (Grandey, 2000; Grandey et al., 2013). We should anticipate burnout when the
emotional demands of the work exceed an employee’s resources or ability to recover from that
work (Grandey, 2000; Xanthopoulou et al., 2018). During times of crisis, organizational goals
may be in flux, and recent research has found that goal uncertainty negatively affects well-being
96
during episodes of emotional labor (Davis & Stazyk, 2021). Taken together, this suggests that
times of crisis will demand norms of compassion while employees are experiencing
unanticipated events, necessitating surface acting to reconcile conflicting emotional experiences
and subsequently contributing to burnout (H1, see Figure 3.1).
Figure 3.1: Hypothesized Relationship Between Crisis, Emotions, and Burnout

COR theory needs to be tested empirically through an intersectional lens. Moments of
crisis involve fewer opportunities to replenish resources, potentially making it harder to recover
from surface acting in a pandemic (Xanthopoulou et al., 2018). Importantly, COR theory asserts
that cultural groups share sets of resources and attribute different values to specific resources
(Hobfoll, 2002). Particularly in the context of organizations, research has shown there are
implicit gendered (Acker 2006, 2012; Ridgeway, 2011) and racialized (Dumas et al., 2013; Ray,
2019; Wingfield, 2010; Wingfield & Alston, 2014) norms that limit the resources and behavioral
choices available to different groups, creating disproportionate strain on women and employees
of color. Previous work has shown that it is often uncomfortable for members of
underrepresented groups in the workforce to form social connections with others in the
workplace when they perceive cultural differences from the majority (Dumas et al., 2013;
97
Phillips et al., 2018). More specifically, building trust across racial lines demands a sense of
psychological safety and common cultural understandings, leading many minority group
members to hold back, limiting their ability to be emotionally authentic in work relationships
(Dumas et al., 2013; Philips et al., 2018). This suggests we should expect to see differences
across social groups when it comes to the relationship between emotional labor and burnout,
with those social groups experiencing more privilege (e.g., White men) having more agency to
express authentic emotions, less pressure to surface act, and more resources to recover from
emotional labor than their lower-status counterparts (e.g., women of color; H2, see Figure 3.2).
Figure 3.2: Hypothesized Relationships Between Cultural Resources, Surface Acting, and
Burnout

DATA, METHODS, AND MEASURES
This analysis explores the extent to which (RQ1) the onset of pandemic exacerbated
burnout among local government employees, (RQ2) emotional labor contributed to the
98
experience of burnout, and (R3) the relationship between emotional labor and burnout during
COVID-19 differed across social groups. Gabriel et al. (2021b) suggested the need for within-
person research methods to unpack the emotional toll of COVID-19, and longitudinal designs are
key to understanding the role that time plays in the dynamics of recovery from resource loss
(Davis et al., 2022; Hobfoll et al., 2018). Experience sampling approaches are especially well
suited for studying self-regulation at work over time, so this study uses a mixed-methods
approach blending qualitative interviews and daily diaries for data collection (Johnson et al.,
2018).
In April 2020, a recruiting email was distributed electronically to local government
employees in a large metropolitan area in the state of California, with 181 employees opting in
and completing at least one daily diary entry. The diary approach is a method of data collection
in which participants respond in real time to repeated assessments while functioning within their
natural settings. As such, diary studies combine high ecological validity with internal validity,
since they allow researchers to study emotional episodes in real-life settings while using
individuals as their own controls. Diary studies that capitalize on major events and transitions
help to model the processes underlying psychological change, making the method particularly
attractive for assessing emotional labor during the COVID-19 pandemic (Bolger et al., 2003).
Data was collected successively in the period of April 13–30, 2020. Respondents were
asked to fill in a daily questionnaire for a period of 3 weeks. The online questionnaire was sent to
the respondents towards the end of the working day, at 4 p.m. Respondents were instructed to fill
in the diary on the day it was sent to them, and daily entries were closed after a period of 24
hours to reduce retrospective biases.
99
Participants were recruited from a previous research project of local government
employees across Los Angeles County, with existing data on measures of burnout before the
onset of the COVID-19 pandemic. General population data from the U.S. Census Bureau (2020)
for Los Angeles Country shows the region is 51% women and 49% men. In terms of
race/ethnicity, Los Angeles County is 48% Hispanic, 26% White, 15% Asian, 8% Black, and 3%
two or more races (U.S. Census Bureau, 2020). The sample included 51.9% women (94) and
48.1% men (87). The racial makeup of the sample was 51.9% White, 21.5% Mixed/Other, 11%
Hispanic, 7.2% Black, 5% Asian, 1.7% Pacific Islander/Native Hawaiian, and 1.7% Native
American/Alaskan Native. Looking at race and gender combined, the sample was 28.2% women
of color, 28.2% White men, 23.8% White women, and 19.8% men of color. Workforce
population data was only available for the City of Los Angeles, where the workforce is 71%
men, 29% women, 38% Hispanic, 29% White, 16% Black, 11% Asian, and 4% Filipino. As a
point of caution, the workforce population data did not allow for respondents to identify as more
than one race/ethnicity and only accounts for one of the 10 municipalities covered, making it
difficult to accurately compare the sample to the workforce population.
An onboarding survey was used primarily to introduce participants to the research
project, providing instructions and guidance to set expectations for the diaries, while also
collecting additional pandemic-specific data on work-from-home status, attitudes and
preferences surrounding the integration of personal and professional boundaries, workload,
workplace climate, and updated measures of burnout.
Quantitative Analysis
Multilevel models leveraged validated scales and measures in the diary data to estimate
the impact of emotional labor on different dimensions of burnout while controlling for a series of
100
other predictors (see Appendix 3A). Importantly, given the complexity of individual differences
like personality that contribute to well-being and resilience (Yi-Feng Chen et al., 2021), this
study analyzes both between- and within-person differences in emotional labor and burnout.
Dependent Variable—Burnout
The dependent variable of interest was burnout, which was modeled in four distinct ways
to account for (1) overall burnout, (2) cynicism/depersonalization, (3) emotional exhaustion, and
(4) loss of personal accomplishment. The overall burnout measure combined all three
dimensions of burnout using 16 items (eigenvalue = 5.79) from the Maslach Burnout
Inventory—General Survey (MBI-GS; Maslach & Jackson, 1986). Burnout was then
disaggregated into its multiple dimensions using confirmatory factor analysis.
28
The cynicism
factor was composed of four items (eigenvalue = 2.53), emotional exhaustion was composed of
five items (eigenvalue = 2.96), and loss of personal accomplishment was composed of six
reverse-coded items (eigenvalue = 2.33). Pre- and post-COVID burnout was measured using two
items from the MBI-GS available from the previous research study that asked respondents the
extent to which they agreed or disagreed with the following statements: (1) “I feel burned out
from my work” and (2) “I have accomplished many worthwhile things in this job.” Because
burnout was not a focal point of the original survey, these two items are useful but imperfect
measures of the multidimensional construct. Factor scores were created for all four types of
burnout at the point of the onboarding survey, the midway diary, and the final diary using the
same item combinations as above. Eigenvalues for the factors corresponding to the onboarding,
midway, and final timepoints fall between 1.97 and 7.79 and are listed in Appendix 3B.

28
When factor loadings were plotted, one item, “I just want to do my job and not be bothered,”
did not cluster among any of the three factors.
101
Predictors of Interest—Emotional Labor & Intersectional Identity
The primary predictor of interest was emotional labor. Emotional labor was measured
using a modified version of Glomb and Tews’ (2004) Discrete Emotions Emotional Labor Scale
(DEELS), which measures the extent to which 14 discrete emotions are (1) genuinely expressed,
(2) faked/amplified, and (3) suppressed. To minimize participant burden, nine discrete emotions
were measured (i.e., contentment, concern, happiness, enthusiasm, fear, anxiety, sadness,
irritation, and anger). These items were measured on consecutive days, with the exception of the
midway and final-day surveys (see Figure 3.3).
When conducting confirmatory factor analysis for the expression of genuine emotions,
discrete emotions clustered into what appear to be socially desirable versus undesirable
groupings, with a factor for the genuine expression of “desirable/positive” emotions (i.e.,
contentment, happiness, and enthusiasm; eigenvalue = 1.77) and a separate factor for the genuine
expression of “undesirable/negative” emotions (i.e., irritation, anxiety, sadness, concern, fear,
and anger; eigenvalue = 2.95). The same clustering of three positive emotions and six negative
emotions also held for amplified and suppressed emotions, creating four additional emotional
labor factors for faking positive emotions (eigenvalue = 1.82), faking negative emotions
(eigenvalue = 3.69), suppressing positive emotions (eigenvalue = 1.80), and suppressing negative
emotions (eigenvalue = 3.49).

102
Figure 3.3: Sequencing of Data Collection

Additionally, the models sought to explore the extent to which intersectional identity was
associated with both emotional labor and burnout. Recent meta-analytic work by Humphrey
(2021) found gender to be insignificant with respect to the relationship between emotional labor
and emotional exhaustion, but much of the existing research has looked at gender without the
intersection of race, lumping together the experiences of White women and women of color. The
models presented here included an interaction term combining gender and race to acknowledge
the unique experiences at the intersection of multiple social identities. Four distinct groups were
created for women of color, men of color, White women, and White men as the reference group,
given that they were the largest group in the population sampled.
Additional Control Variables
At the individual level, the models controlled for age, highest level of education, and
political ideology. Lower levels of education and more conservative political ideology have been
associated with lower pandemic-related anxiety (Freiling et al., 2021; Rothgerber et al., 2020),
which may influence feelings of burnout. The models also accounted for several perceptual
103
variables associated with motivations and constraints on the job. The more an individual
identifies with their role, the more likely they are to spontaneously feel the socially appropriate
emotions they are expected to express, or to use deep acting to bring their feelings into alignment
with role expectations to preserve their identity (Humphrey et al., 2015). Thus, a prosocial
motivation factor composed of four items (eigenvalue = 2.02) and an additional item asking
respondents the extent to which they found their work meaningful were used to measure
motivation and role identification. Wharton (1993) argued that rather than having uniformly
negative effects, the consequences of emotional labor vary by levels of job autonomy. To
measure perceived autonomy versus constraints, Borry’s (2016) Three-Item Red Tape (TIRT)
Scale was used to create a red tape factor (eigenvalue = 1.22), and four additional items from
Leiter and Maslach’s (2011) Areas of Worklife Survey (AWS) were used to create a workload
factor (eigenvalue = 1.80). Two additional items were used to control for the frequency with
which respondents interacted with (1) coworkers or professional colleagues and (2) residents on
a given day.
Grandey et al. (2012) argued a climate of authenticity among coworkers can offset the
resource loss spiral caused by interpersonal mistreatment by replenishing resources and buffering
against the strain of emotional labor. The climate of authenticity is a unit-level construct defined
by the extent to which coworkers value, accept, and respect authentic emotional expressions
among coworkers (Grandey et al., 2012). In inauthentic environments, coworkers have to
regulate their emotions around both customers and coworkers, but in a climate of authenticity,
there is an opportunity to restore depleted emotion-regulation resources through authentic
interactions with coworkers, a form of social support (Grandey et al., 2012). At the work group
level, Grandey et al.’s (2012) seven-item measure was used to create a perceived climate of
104
authenticity factor (eigenvalue = 4.17). The models included two additional items from the AWS
(Leiter & Maslach, 2011) assessing perceived social support and closeness to one’s colleagues.
To accommodate the unique dynamics of the COVID-19 pandemic, several controls were
implemented to understand work–home boundaries.
29
The models controlled for the number of
individuals living in the household, the telecommuting status of the employee, and whether the
respondent was living with a dependent child or children. Additionally, three items from
Kreiner’s (2006) measure of work-family segmentation preferences, three items from Desrochers
et al.’s (2005) work–family role integration-blurring scale, four items from Netemeyer et al.’s
(1996) measure of work–family and family–work conflict, and four items from Kossek et al.’s
(2006) measure of integration/segmentation enactment behaviors were combined into a
segmentation/integration factor (eigenvalue = 2.17) to account for preferences for separating the
personal and professional domains.
RESULTS
Empirical Tests
A series of chi-square tests were run comparing responses to two burnout indicators
before and after the onset of the pandemic. The test revealed a statistically significant difference
in pre- and post-pandemic onset responses to the statement “I feel burned out from my work” (χ
2

= 96.96, p ≤ 0.001), with more burnout reported during COVID-19. The second indicator asked
the extent to which the respondent agreed or disagreed with the statement, “I have accomplished

29
Mirroring COR logic, the scarcity hypothesis treats work and family obligations as competing
over the course of daily activities for a fixed amount of a person’s energy (Wharton, 1993).
Premised on a zero-sum view of emotional energy, expending energy at home limits the
resources an individual has to devote at work and vice versa, and Wharton (1993) showed that
emotion work at home has negative consequences for women’s job-related well-being.
105
many worthwhile things in my job,” and a chi-square test again revealed a statistically significant
difference in pre- and post-pandemic onset responses (χ
2
= 98.63, p ≤ 0.001), with more
disagreement with the statement during COVID-19, reflecting reduced achievement. While these
two items are limited in their ability to confirm increases in all dimensions of burnout, they
signal that onset of the pandemic was a period of increased burnout.
To better understand the mechanisms underpinning burnout at the onset of COVID-19,
the more comprehensive battery of 16 burnout questions from the MBI-GS was used on
participants at the beginning, middle, and end of the diary. Four multilevel models
30
were run to
predict burnout in its aggregate and disaggregated forms. A positive coefficient for covariates of
interest suggests (1) more overall burnout, (2) more cynicism, (3) more emotional exhaustion,
and (4) reduced feelings of accomplishment. Importantly, these models interacted intersectional
identity with the amplification of positive emotions and the suppression of negative emotions.
These interactions let us look at whether or not specific forms of emotional labor had a stronger
relationship with burnout for some groups, with White men omitted as the reference group. The
models also included department fixed effects (ጊ
!
) to account for professional norms and agency-
specific characteristics and day fixed effects ("
"
) to address job tasks that might be specific to
days of the week.
2%&1.%/
,+
= #
!!
+ #
"!
(4)1%01) (.*) + #
#!
(4)1%01) 1)4) + #
$!
(amplify pos) + #
%!
(6-(7089 1)4) +
#
&!
(suppress pos) + #
'!
(suppress neg) + #
!"
(intersectional identity) + #
!#
(generation) + #
!$
(education) +
#
!%
(ideology) + #
!&
(prosocial motivation) + ##
!'
(avg interaction colleagues) + #
!(
(avg interaction residents) +
#
!)
(children) + #
!*
(telecommute) + #
!"!
(work-home) + #
!""
(workload) + #
!"#
(red tape) + #
!"$
(climate of
authenticity) + #
!"%
(social support) + +#
$"
(amplification pos * intersectional identity) + #
'"
(suppression neg *
intersectional identity) + ጊ
+
+ "
,
+ %
!+
+ &
,+

30
A series of unconditional models assessed the extent to which variation in burnout was
attributed to between- versus within-person differences and is available in Appendix 3C.
106
Table 3.1: Coefficients for Intersectional Identity & Emotional Labor
(1)
Burnout All
(2)
Cynicism
(3)
Exhaustion
(4)
Accomplishment
White women -2.213***
(0.258)
-1.899***
(0.389)
-2.196***
(0.296)
-0.944***
(0.331)
men of color 0.377
(0.296)
-0.414
(0.404)
-0.603*
(0.349)
0.985***
(0.351)
women of color 0.432**
(0.215)
0.149
(0.330)
0.253
(0.252)
0.370
(0.279)
genuinely expressed
positive emotions
-1.093***
(0.120)
-0.611***
(0.169)
-0.844***
(0.141)
-0.775***
(0.144)
genuinely expressed
negative emotions
-0.0744
(0.188)
-0.149
(0.253)
-0.655***
(0.212)
0.699***
(0.216)
amplified
positive emotions
-0.628*
(0.363)
-0.268
(0.432)
1.188***
(0.431)
-1.539***
(0.357)
amplified
negative emotions
-0.438***
(0.137)
-0.294
(0.213)
-0.458***
(0.162)
-0.0625
(0.177)
suppressed
positive emotions
0.279
(0.230)
-0.225
(0.276)
1.150***
(0.260)
-0.770***
(0.232)
suppressed
negative emotions
-0.393*
(0.230)
0.0277
(0.263)
-1.473***
(0.262)
0.214
(0.221)
Interactions: Identity * Amplification of Positive Emotions
White women -0.0277
(0.498)
-0.298
(0.431)
-1.937***
(0.583)
1.280***
(0.354)
men of color 0.778**
(0.333)
0.333
(0.426)
-0.754*
(0.394)
1.778***
(0.352)
women of color -2.576***
(0.665)
-0.908
(0.745)
-4.780***
(0.737)
-0.773
(0.638)
Interactions: Identity * Suppressed of Negative Emotions
White women 0.452*
(0.263)
0.360
(0.338)
1.469***
(0.308)
-0.387
(0.281)
men of color 0.344
(0.302)
0.590*
(0.319)
0.969***
(0.362)
0.716***
(0.274)
women of color 2.769***
(0.583)
1.823***
(0.608)
3.427***
(0.645)
1.238**
(0.521)
Obs
# Groups
118
61
136
63
127
61
128
63
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Table 3.1 shows the coefficients for intersectional identity, emotional labor, and
interactions between the two, with full model results for all controls available in Appendix 3D.
When employees genuinely express the positive emotions they are feeling, the results show a
107
negative relationship with all forms of burnout. When employees are able to genuinely express
negative emotions, the results suggest there is less emotional exhaustion but a reduced sense of
personal accomplishment. This makes intuitive sense, as challenges in the workplace likely
evoke negative emotions and make it harder for employees to do their job, but having the ability
to vent or vocalize frustrations would not necessarily increase their sense of emotional
exhaustion because it is an authentic expression.
Counter to the hypothesis that emotional labor, particularly surface acting, would
contribute to an increase in burnout during COVID-19, the main effects for emotional labor
suggest there were only statistically significant increases in burnout in specific scenarios: (1)
when employees faked or amplified positive emotions and (2) when employees suppressed
positive emotions. A more nuanced look at the interaction terms reveals that amplifying
positivity is actually associated with reduced emotional exhaustion for women of color, whereas
men of color experience less emotional exhaustion but more overall burnout and reduced
personal accomplishment, and White women similarly experience less emotional exhaustion but
a reduced sense of personal accomplishment compared to White men. The suppression of
negative emotions is associated with increased burnout of all types for women of color, three
types for men of color (i.e., cynicism, emotional exhaustion, and reduced accomplishment), and
only two types for White women (i.e., overall burnout and emotional exhaustion) compared to
White men. Taken together, these results show that women of color are most affected by
pressures to suppress negative emotions on the job but may find some benefit from amplifying
positive emotions. Men of color and White women also experience increased burnout when
suppressing negative emotions, but to a lesser extent than women of color. An important
difference between groups is that men of color and White women are more prone to experiencing
108
a reduced sense of accomplishment when they fake positivity, which is not the case for women
of color.
Qualitative Content Analysis
To understand how different groups experienced burnout, responses from 60 qualitative
interviews and 793 diary entries mentioning burnout were analyzed (for the wording of
qualitative prompts, see Appendix 3A). Within the diaries, the entire free-response entry for each
open-ended question was used as the unit of analysis (Charmaz, 2014). ATLAS.ti software was
used to develop a coding scheme based on code families (Campbell et al., 2013; Charmaz, 2014;
Friese, 2019). After an initial round of grounded coding produced over 100 unique codes, the
second phase of coding involved sorting existing codes into three primary categories that
incorporated descriptions of (1) cynicism/disengagement, (2) emotional exhaustion, or (3) loss of
personal accomplishment. Consistent with the multidimensional definitions of burnout by
Maslach and Jackson (1986), diary entries were deductively coded with
“cynicism/depersonalization” when there was mention of (a) loss of interest and enthusiasm for
work, (b) distancing themselves from peers and the work environment, or (c) doubting the
significance of and becoming cynical about their role and/or contributions. Diary entries were
given the “emotional exhaustion” code when there was mention of physical, mental, or
emotional fatigue or numbness. Entries that mentioned a lack of confidence in their ability to
execute the job or an inability to complete work were given the code “reduced accomplishment.”
With these primary categories established, I inductively developed secondary codes that
seemed to be the drivers of the descriptions of burnout. Secondary codes for (1) challenges in
managing COVID in their personal lives (personal life), (2) changes to the workload or job tasks
(job design), and (3) interpersonal interactions or events that triggered negative emotions
109
(emotional events) were applied across the primary categories to identify the sources of burnout.
The personal life code encompassed mentions of participants’ social lives or family lives being
impacted by COVID-19, including mentions of social isolation, challenges with children, and
difficulties with spouses or partners. The job design code encompassed mentions of mixed
signals or changing priorities, new protocols and inconsistencies in processes, and increased
workloads. Lastly, the emotional events code included mentions of high-stress experiences,
accounting for the emotions of others and negative emotions like fear, concern, anger, and
anxiety. As examples, a diary entry that read “Toddler was having a day of big emotions, which
is exhausting” would have been assigned a primary code of emotional exhaustion and a
secondary code of personal life, whereas an entry like “I was anxious about doing a task I wasn’t
comfortable with, which can have negative repercussions if I do it incorrectly” would have been
assigned a primary code of reduced accomplishment along with secondary code of emotional
experience for its specific mention of heightened anxiety (for more examples, see Appendix 3E).
This coding scheme was also applied to all 60 interview transcripts in which participants were
explicitly asked about feelings of burnout.
Figure 3.4: Type of Burnout

Interestingly, reduced feelings of efficacy or loss of personal accomplishment were
mentioned most frequently (43%), followed by mentions of increased cynicism or
disengagement (32%) and lastly, emotional exhaustion (25%) (see Figure 3.4). Using ATLAS.ti
110
co-occurrence tables, there were clear indications that burnout was associated with (1) challenges
managing personal life stressors (personal life), (2) fluctuations in job responsibilities and
workload (job design), and (3) stressful interpersonal interactions that elicited negative emotions
(emotional events). The co-occurrence of primary and secondary codes is illustrated in Figure
3.5, which shows that descriptions of reduced personal accomplishment most frequently co-
occurred with negative emotional events and changes to job goals and workload, whereas
cynicism predominantly co-occurred with negative emotional events, and emotional exhaustion
predominantly co-occurred with personal life stressors and negative emotional experiences.
Figure 3.5: Percent of Secondary Codes Applied to Burnout Dimensions

Reduced Personal Accomplishment & Job-Related Confidence
Across men of color (49%), White men (47%), White women (40%), and women of color
(35%), descriptions of reduced personal accomplishment and a lack of job-related efficacy were
mentioned more frequently than any other dimension of burnout. In all groups, participants
wrote, “not feeling I did enough” (17:938), “anxiety about whether or not I am ‘useful’”
(17:671), and “feeling discouraged by how much some other folks are accomplishing” (4:462) as
111
examples of a reduced sense of personal accomplishment. For others, entries like “I’m spinning
my wheels on a project” (4:55), “I am unequipped” (4:1393), and “I worry a great deal about
someone complaining about my work performance to a supervisor” (17:848) reflected a lack of
confidence in their abilities to do their job.
Among descriptions of reduced accomplishment/confidence, 62% of diary entries
mentioned negative emotional events, 47% mentioned challenges in the workload or job tasks,
and only 8% referenced personal life challenges interfering with work performance and
confidence. Negative emotional events often took the form of feeling helpless to help others. As
examples, participants wrote that their stress stemmed from things like “communicating with
constituents we could not help” (4:808), and “talking to the public makes me sad because we
can’t fix every problem” (5:1526). This negative emotional state paired with the organizational
norms for suppressing negative emotions may create feelings of isolation among employees,
which is of particular concern for women of color given the negative outcomes of emotional
suppression from the quantitative analysis.
With respect to job design, many employees saw their roles shifting in ways that
significantly increased their workloads and put them in positions where they were engaging in
emergency management functions that were unfamiliar to them. One interview respondent
shared that the increased pace of work felt like “running a marathon at a sprinter’s pace” (9:7).
Others shared that it was the combination of workload and uncertainty that evoked the strongest
feelings of fear, self-doubt, and stress. As an example, one employee shared,
Yes, my workload has dramatically increased, given my role in my workplace, but
the uncertainty of “Oh my god, am I going to get this? Oh, my God, [are] any of
my friends or family going to get this? Are we all gonna die? Is the world ending?
Am I gonna have a job?” You have all of that stress on top of the work stress.
(9:1)
112
Cynicism & Disengagement
Descriptions of increased cynicism or disengagement from work were mentioned in 36%
of diary entries from White men, 32% from White women, 28% from women of color, and 25%
from men of color. Participants shared entries describing a lack of motivation, inability to focus,
and outright skepticism about the intentions of leadership. As examples, one participant shared it
was “hard to get motivated to do work today” (4:1057), while others described “feeling like a
pawn in someone else’s chess game” (4:35) and concerns that “management is pretty much ready
to sacrifice us whenever it suits them and enhances their position” (20:292). These descriptions
most often co-occurred with mentions of negative emotional experiences (40%) and, to a lesser
extent, personal life challenges (20%) and job design changes (19%). The negative emotions
characterized by cynicism or disengagement entries tended to be somewhat passive, using terms
like “withdrawn,” “unmotivated,” “bored,” or general descriptions of apathy.
An employee who was approached by an upset resident using expletives shared, “I
thought I should have reacted with greater emotion, but I didn’t—I didn’t really feel any strong
emotions, and I don’t know why” (5:1842). Similarly, an employee taking on a new role as a
disaster service worker assigned to manage a call center added the calls from the public “have
been a little more difficult, and I feel almost, like, a degree of compassion fatigue” (31:1). While
these reactions to stressful encounters show employees were able to maintain a neutral tone, it
stemmed from a maladaptive coping mechanism of becoming numb to the work and disengaged
from the emotions of residents.
With respect to personal life challenges, entries mentioning cynicism and disengagement
often co-occurred with accounts of news or media images that were distressing. As an example,
one participant speculated residents were especially challenging to work with because of the
113
misinformation that makes their job “impossible,” writing, “the tv and print media is complicit in
causing divisiveness in this country and they get away with it” (1:405). Another participant
wrote that when she watched the news, she felt she had to
will myself to not speculate about how dystopian this may all become. Will people
be stealing food soon? Who could blame them? Will my circle of people be
victimized by unrest somehow? I can’t afford to descend into that thinking, too
much to keep track of on the job. (1:497)
With respect to job design and cynicism, entries primarily described the mixed signals
and uncertainty surrounding their jobs. One interviewee shared, “I think there’s a lot more
emotionally, it’s just the unknown, it’s very difficult to predict what will happen for my job, for
the folks I work with, the programming we put on … it’s pretty overwhelming” (22:1). Many
employees shared they feared losing their jobs given the economic fallout, and that fear led them
to distance themselves from their work and their peers. As an example, one interviewee shared,
“I worry every day that I could be let go; I felt completely secure in my employment before the
stay home order” (48:67). For others, anxiety and fear contributed to disengagement from work,
with one respondent sharing,
I woke up this morning with an overwhelming feeling that I did not want to return
to work in any capacity at any time in the future. That really concerned me … I
considered sharing my feelings with my former boss but decided not to because I
didn’t want him to worry. (5:1639)
Emotional Exhaustion
Women of color described emotional exhaustion more than any other group, and it
accounted for roughly 30% of their diary entries, whereas emotional exhaustion only emerged in
23% of entries for White women, 22% for men of color, and 15% for White men. Diary entries
of emotional exhaustion included comments like “I arrive home mentally and emotionally
exhausted” (12:987) and “I’m cumulatively exhausted, and this situation isn’t helping my
anxiety or depression” (1:29). Another employee described being exhausted from dealing with
114
the public and the gravity of the pandemic combined, sharing, “I am more emotional driving
home, sometimes crying. I am tired, concerned” (48:81). Emotional exhaustion most frequently
co-occurred with descriptions of personal life challenges (55%), followed closely by negative
emotional experiences (52%) and, to a lesser extent, job design elements (22%).
Challenges managing emotions across the work–home boundary emerged more
frequently among women of color compared to other groups, necessitating further exploratory
research. Exhausting emotional experiences often stemmed from managing the emotions of
others personally and professionally. In the personal life domain, this involved dealing with
relationships becoming strained while socially distant, managing grief over lost loved ones, and
household dynamics with children, spouses, and roommates. As an example, one respondent
wrote,
My daughter who is homeschooling for third grade has lots of anxiety and
sadness from school being closed that surfaces occasionally in emotional bursts
and leaves me feeling drained and anxious about her needs not being met.
(14:175)
A respondent who lived alone shared,
My work required me to be focused on COVID all day, so it’s draining mentally.
I’m trying to maintain communication with family members, but it’s also
sometimes draining, since conversations always go back to feelings about how it’s
affecting everyone. I have trouble escaping it, which only adds to the anxiety.
(14:796)
Managing grief was another challenge that depleted emotional resources. Several women of
color mentioned the pain they experienced not being able to engage in meaningful cultural rituals
surrounding the loss of life, with one sharing, “there’s never a good time to die but, if it
would’ve happened in January, we could’ve been there—we could’ve grieved it the right way”
(15:6).
115
In the work domain, employees described feeling responsible for managing the emotions
of both colleagues and the public during the crisis. Those in leadership positions described
feeling concern for the emotional well-being of their employees. One supervisor shared,
There’s just a lot of extra anxiety and stress that seeps into everything. We talk a
lot about whether that person seemed a little down or that person seems short-
tempered, and we’re just kind of acknowledging everyone’s got this added burden
of stress that everyone’s feeling. (38:1)
Another supervisor added they felt the emotions of their colleagues were difficult to manage,
sharing, “everyone is operating on such a short fuse, myself included. My job is already
frustrating, but I have less patience/capacity for it than I used to because of this” (48:66). While
the sense of responsibility for others seemed heightened among supervisors, the general
sentiment was not exclusive to leaders. Many employees described concern for their colleagues.
One employee shared of their work group, “everyone is affected, and you kind of just can get
this sense of people being a little more testy or stressed” (38:1).
Importantly, the qualitative insights from diary entries and interviews suggest women of
color were emotionally exhausted, but the root of their exhaustion extended beyond the
boundaries of the work context and varied among different racial/ethnic groups. Several women
of color mentioned being exhausted by the emotional demands of the job coupled with the
emotional demands at home, particularly with respect to the health and emotional well-being of
loved ones. Within response and recovery efforts, disasters typically highlight social
vulnerabilities that require culturally competent practices, with marginalized and socially
vulnerable populations facing more economic, political, physical, and sociocultural obstacles
(Knox & Haupt, 2020). Although the samples for distinct racial/ethnic groups of women were
small in this study, some important distinctions emerged warranting further exploration. Given
that the sample was based in Southern California, where there is a large Hispanic population that
116
often lives in multigenerational households, many of the Hispanic women in the sample (n=19)
described juggling multiple meaningful relationships at home that required emotional attention.
As an example, a Hispanic woman shared, “it has been extremely stressful for my family,
particularly the ones who don’t know how to speak English—I’m an immigrant, so my family
members are relying on me a lot to just navigate things for them and bring clarity” (14:6).
By contrast, Asian women in the sample (n=17) described the challenge of navigating the
anti-Asian sentiment and rhetoric that emerged around COVID-19, given its presumed origins in
China. The sometimes asymptomatic nature of the virus has led people to turn to non-health-
related traits to identify potential carriers (Roberto et al., 2020). Misplaced blame for carrying
the virus has contributed to an increase in discrimination targeting Asians during the pandemic,
impacting their sense of safety (Roberto et al., 2020). Multiple Asian women described feeling
on edge and emotionally guarded in public, as they were concerned about verbal and physical
abuse for themselves and their families from strangers who held hostile attitudes towards the
Asian community.
Black women in the sample (n=15) found themselves extending their roles as
representatives of the government into their communities. Several Black women described a
historical skepticism of government intervention in public health matters given the decades-long
history of neglect and mistreatment of Black patients in the health care system. One Black
woman interviewed said the public viewed her as a “community institution.” She shared that her
community had so little faith in local government that they took every concern directly to her,
meaning she carried the weight of the entire institution on her shoulders as the only point of
contact her community could trust. While this study can clearly demonstrate a link between
117
emotional labor and burnout, particularly among women of color, the qualitative findings point
to more cultural nuances among women of color that need to be explored.
It is important to note that not all interviewees described a heightened emotional state.
Particularly among those whose roles shifted to require less public engagement, working during
the pandemic was described as a “relief” or “respite.” As an example, one interviewee shared,
“part of what has been a little bit easier is we can’t do the public engagement necessary for
[projects], so the work has been a lot more planning for the future, which can be a little more
positive” (29:1). Other employees shared that teleworking had lightened interpersonal burdens
for them. One respondent shared that her teleworking status was a marked improvement from her
job prior to COVID-19, saying, “I find it to be much less stressful, and I feel very removed and
remote from the real activity of what’s going on” (17:1). Another teleworker added, “I feel like
it’s a little less, I don’t want to say less emotional, but it’s a little lighter in that we’re helping
people in a more abstract way” (23:2). These findings are consistent with work by Grandey et al.
(2021), who found service employees with reduced interpersonal workloads during the pandemic
had more personal recovery from the taxing aspects of their jobs.
DISCUSSION/CONCLUSION
This paper offers lessons that public managers and human resource professionals might
draw from to better cope in future crises. Increasing gender and racial representation is one
mechanism for increasing the in-house knowledge public organizations have to better serve
diverse communities during times of crisis, but that may place outsized emotional burdens on
already underrepresented employees. Quantifying the disparities in emotional labor across
groups, and the subsequent impacts on burnout, is a first step towards standardizing functions to
ensure all employees are treated fairly and emotional burdens are equitably distributed. Taking
118
stock of the emotional labor demanded of public servants during the pandemic is grounded in a
desire for more compassion and empathy for the people in public service, helping to humanize
employees. While existing literature has shown that the pandemic exposed and exacerbated
existing inequities for women and people of color (Berry et al., 2022), this research provides
much-needed context for the ways in which the challenges of those groups might be manifesting
in the work environment. This paper makes it clear that emotional labor performed during a
crisis response is clearly linked to more burnout, particularly among women of color, which has
consequences for public sector human resource management.
The results from statistical modeling show that in general, women of color were more
likely to experience overall burnout, and when they suppressed negative emotions, they were
more likely to experience all forms of burnout compared to White men. Interestingly, amplifying
positive emotions was not associated with increased burnout among women of color and was
actually associated with less emotional exhaustion and overall burnout. Perhaps like with any
hard task, practice at suppression may make emotional work easier to manage over time,
requiring less effort and reducing the likelihood of becoming exhausted. It may be that women of
color have developed a set of emotional labor skills throughout their lives that make
amplification of emotion less effortful than might be anticipated. Alternatively, perhaps this
distinction between amplifying positivity and suppressing negativity is because faking positivity
to match interpersonal and organizational expectations is a proactive strategy towards achieving
a desired outcome, whereas suppressing negativity is a reactive strategy to minimize potential
negative outcomes if negative emotions are not tolerated in the workplace.
119
Recommendations
While the COVID-19 pandemic is unique in many respects, we can draw some lessons
about how to better prepare the public workforce for emotionally demanding crisis response
work. Rather than relying on the prosocial motivation of public servants to weather the storm,
public managers might consider institutional changes to better support their workforce. In
particular, I recommend public institutions extend autonomy and institutionalize a climate of
emotional authenticity, psychological safety, and social support to help buffer against burnout
during times of crisis.
Institutionalizing practices that provide more autonomy to employees is a way of
communicating trust and respect and allows individual employees more agency to innovate and
think creatively about finding solutions to tough problems in difficult circumstances. As a form
of autonomy, institutionalizing forms of reflection, sensemaking, and critical thinking during
times of crisis will allow public organizations to capture the learnings of this challenge to better
understand how to prepare for the next. Recent work has shown that organizations that foster a
learning culture when adapting to crises and unanticipated external circumstances can increase
work engagement, employee resilience, and psychological empowerment (Blaique et al., 2022).
This is largely believed to be because employees with more autonomy feel safe to innovate. A
culture that focuses on creating avenues for all employees to participate in sharing information
and providing constructive feedback alongside continuous training opportunities has been shown
to increase employees’ confidence in their ability to deal with new work challenges (Blaique et
al., 2022). Autonomy in this sense helps sustain commitment, allows for innovation, and helps
promote equity (Berry et al., 2022).
120
When exploring the additional control variables across the models, the perception of a
workplace climate where emotions could be authentically expressed was associated with less
burnout in all of its forms. Humphrey et al. (2015) argued that organizations need to create a safe
and positive emotional climate to reduce the need to leverage surface acting, but the mechanisms
for developing such a climate are still unknown. Nembhard and Edmondson (2011) defined
psychological safety as “the general belief that one is comfortable being oneself—being open,
authentic, and direct—in a particular setting or role” (p. 491). Psychological safety has been
shown to vary significantly across groups within the same organization (Edmondson, 1999,
2004), implying it is a group-level climate variable subject to local influences such supervisory
behavior, coworker relationships, goal clarity, task interdependence, and work group
demographics (Nembhard & Edmondson, 2011).
Relatedly, prior research has suggested that increased social support is a workplace
resource that can help employees cope with adversity and buffer against burnout (Linos et al.,
2021). Linos et al. (2021) provided experimental evidence that workplace-based interventions
bolstering perceived social support among peers can reduce burnout and turnover. More
specifically, they found that nudging employees to share advice and read about their peers’
experiences increased perceived social support and functioned as a job resource that significantly
reduced employee burnout and resignations (Linos et al., 2021). This is a particularly promising
finding given that it is low-cost intervention that can be implemented without direct interaction
in person.
Additional research is needed to identify the antecedents of psychological safety, a
climate of authenticity, and social support to see if such conditions can be replicated
systematically (Grandey et al., 2012). The challenge is that diverse groups may struggle to
121
achieve psychological safety and feel socially supported when differences translate into status
and power differences
31
(Nembhard & Edmondson, 2011). Amis et al. (2019) made the case that
identifying who has power and to what end in organizations is critical to understanding how
inequality is persistently reproduced through everyday processes and interactions. This
necessitates future research that expands beyond race and gender in intersectional analyses to
include additional markers of status and difference (e.g., sexual orientation, neurodiversity, etc.)
(Ortiz & Mandala, 2021).
Limitations & Future Directions
Future work should account for idiosyncrasies across professions and seek to measure
genuine emotional expression and deep acting more rigorously to address some of the data
limitations in this paper. To the first point, the relationship between emotional labor and burnout
is likely different for a city attorney compared to a firefighter or a public school teacher given the
differences in the emotional demands of those jobs. While departments were controlled in this
study, future research should focus on comparisons between different occupations to provide a
more comprehensive understanding of how to best support employees across job types. To the
second point, this paper, along with the bulk of the emotional labor literature, focused on the
harmful consequences of response-focused regulation of emotion that either fakes or suppresses
emotional expression. Beyond surface acting, the genuine expression of emotion was
consequential for burnout and warrants further exploration. Prior research has suggested that the
nature of the relationship between emotional labor and burnout is contingent upon whether the
employee is surface or deep acting (Humphrey et al., 2015; Kammeyer-Mueller et al., 2013), and

31
For Nembhard and Edmondson (2011), status is the amount of prominence, respect, and
influence associated with a characteristic (e.g., age, gender, ethnicity, etc.), while power refers to
the ability to influence people and get them to do things they would not otherwise do (from
Pfeffer, 1994).
122
recent meta-analytic work has found that deep acting improves depersonalization and the sense
of personal accomplishment (Humphrey, 2021).
It is generally unhelpful to categorize specific forms of emotional labor like deep acting
as “always good” or surface acting as “always bad” because of variations in people’s immediate
and longer-term goals (Grandey et al., 2013). Additionally, a major challenge in studying
emotional labor is the emerging consensus in the social cognitive literature that many regulatory
processes can be engaged automatically so they are operating the background of our conscious
awareness (Grandey et al., 2013). Grandey et al. (2013) suggested emotion regulation strategies
may slowly become automatic through repetition, eventually requiring little conscious reflection
and feeling genuine. This may partially explain why amplifying positivity is not as emotionally
taxing for women of color compared to White men. Women of color may have developed the
skill over time as a coping mechanism for managing their impression and countering negative
group stereotypes in the workplace. Pointing to differences between collectivist and individualist
societies, Hochschild (1983) suggested, “ironically, cultures which require the most emotional
labor—and may be home to its most highly-trained practitioners—may also be those that inhibit
the very recognition of it” (p. xi). It may be the case that certain groups have grown so
accustomed to managing their emotions that it no longer feels effortful and thus doesn’t get
reported as emotional labor and/or has little impact on burnout. This could also be the case
among women of color of different generational cohorts, who were socialized under different
emotional expectations. A more in-depth analysis of the emotional skills and resources of
different women of color would be valuable as public institutions seek to retain a diverse
workforce during turbulent times.

123
Conclusion
This dissertation set out with three main goals: (1) to bring much-needed attention to
emotional labor in the public sector by rigorously measuring the ways external and internal
interactions demand emotional effort, (2) to leverage an intersectional lens to showcase how
emotional labor is unevenly distributed across employees, and (3) to detail how the inequitable
distribution of emotional labor contributes to burnout, especially for women of color. Increasing
gender and racial representation in the workforce is one mechanism for increasing the in-house
knowledge public organizations have to better serve diverse communities, but this research
shows there is an outsized emotional burden placed on women and employees of color, who
subsequently experience more burnout.
Whether it’s interacting with an angry resident or a hostile colleague, this research
reveals that public service work is emotion-laden. In alignment with past research, in the first
chapter I show that external service encounters are emotionally taxing, and when employees
engage residents, they often suppress negative emotions to manage the interaction. I also
demonstrate that internal team dynamics demand different forms of emotional labor, and
relationships with colleagues are especially memorable sources of stress. The techniques
employees use to navigate one-off interactions with the public are different from the ones needed
to manage ongoing relationships, with a tendency towards surface acting in the former and deep
acting in the latter. This important distinction between external and internal interactions sheds
light on the range of skills needed to effectively serve residents and navigate public institutions.
In the second chapter, I also make visible the gendered, racialized, and intersectional
forms of emotional labor embedded in public service work. I find that White women, men of
color, and women of color engage in emotional labor in different ways, which carry different
124
consequences for their mental health, burnout, and ultimately retention of a diverse workforce.
Measuring those between-group differences makes the case that theorizing about gender or race
at the aggregated level without exploring norms and expectations for different emotions at the
intersections largely misses the mark. Those socially located at the intersection of multiple
marginalized identities (i.e., women of color) are more likely to engage in taxing forms of
emotional labor on the job given their relative power/status disadvantage. The qualitative
components of the research revealed key differences in the rationales different groups have for
engaging in emotional labor. Women of color were most likely to reference office power
dynamics and were also more sensitive to structural elements of the workplace, with
environments characterized by high red tape increasing their emotional suppression and
environments characterized by an openness to emotional authenticity reducing their suppression.
Women of color in particular were most likely to explicitly name racism and sexism and, perhaps
consequently, also described experiencing more emotional exhaustion than any other group.
These findings make it clear that emotional labor is brought about through distinct circumstances
and has a range of consequences for different social groups.
Lastly, Chapter 3 shows how times of crisis evoke emotional labor, with harmful
consequences for employee burnout. While the COVID-19 pandemic is unique in many respects,
scholars interested in crisis management can draw some lessons about how to better prepare the
public workforce for emotionally demanding disaster service work. I show that women of color
in particular experienced more overall burnout, and when they suppressed negative emotions,
they were more likely to experience emotional exhaustion, cynicism, and a reduced sense of
accomplishment. Rather than relying on the prosocial motivation of public servants to weather
the proverbial storm, scholars should consider what institutional changes might help better
125
support this workforce. Preliminary findings from my work suggest that extending autonomy and
institutionalizing a climate of emotional authenticity, psychological safety, and social support
may help buffer against burnout during times of crisis, particularly for the most vulnerable
employees.
Future Directions
An important finding from this work is the emotional sophistication women of color
displayed when describing how they think about managing their emotions. While the inequitable
distribution of emotional labor is certainly something that needs to be addressed, the fact that
women of color are forced to engage in more emotional labor and a wider range of emotional
techniques has contributed to a degree of emotional fluency across contexts. This dissertation
emphasizes the costs associated with the uneven emotional demands on the job, but it should be
noted that those lopsided demands have accelerated the development of a sophisticated set of
emotional skills among women of color. In one sense, this research reveals a hidden emotional
tax on women of color, but in another sense, I also reveal a set of underappreciated assets they
leverage on the job.
Future work should explore how enhancing competencies for emotional labor might
positively impact public administrators, public organizations, and the communities they serve.
Building upon Gabriel et al.’s (2021b) original call to study the emotional complexity of crisis
response, future research might consider integrating literature from positive organizational
scholarship (POS) that emphasizes the role of positive emotions, positive processes, and positive
institutions to enrich our understanding of individual, group, and organizational assets and
balance the emotional labor literature’s emphasis on dissonance, exhaustion, and burnout. The
dominant perspective is that emotional labor is harmful to personal well-being (Hochschild,
126
1983), but for those providing compassion during times of crisis, emotional labor can be
beneficial (Ashforth & Humphrey, 1993; Côté, 2005; Grandey & Gabriel, 2015). Shuler and
Sypher (2000) departed from most treatments of emotional labor by featuring workers who seek
out emotional labor as a fun, exciting, and rewarding part of their work. For Maslach et al.
(2008), engagement represents the opposite pole of the burnout continuum, and Chapter 3
showed that several emotional labor techniques actually had negative relationships with burnout.
White women and employees of color who faked/amplified positive emotions actually
experienced less emotional exhaustion, as did those who engaged in the genuine expression of
positive emotions.
Fredrickson’s broaden-and-build theory suggests positive emotions broaden thought-
action repertoires in the short term and in the long term build personal and social resources in a
self-reinforcing and contagious process of upward spirals that can transform groups and
organizations by increasing engagement (Cameron et al., 2003; Fredrickson, 2003;
Vacharkulksemsuk & Fredrickson, 2013). Although positive affect is transient, the building of
durable personal and social resources in moments of positivity accumulates and equips
individuals to better meet future opportunities and challenges (Fredrickson & Losada, 2005;
Fredrickson et al., 2011). Resources accrued from positive emotions can take a variety of forms,
including “cognitive (e.g., mindfulness skills or intellectual complexity); social (e.g., dense
social networks and high-quality friendship bonds); psychological (e.g., resilience or optimism in
the face of adversity); or physical (e.g., the ability to rebound from stress-induced cardiovascular
activity or ward off the common cold)” (Vacharkulksemsuk & Fredrickson, 2013, p. 48).
Importantly, because people can control which emotions they feel and when through emotion
regulation techniques, they can proactively choose techniques that steer them toward upward
127
spirals (Fredrickson et al., 2011). This reinforces the notion of emotional labor as a skill, which
means emotional competencies should be integrated into hiring, training, and compensation
processes to facilitate more positive spirals and resilience.
To balance the majority of critical diversity research focused on examining
organizational inequalities, the skill with which women of color engage in emotional labor can
also be seen as an asset. A positive approach to diversity management shifts from deficit inquiry
(what is the problem?) to appreciative inquiry (what works well?) and complements existing
studies of gendered and racialized emotional labor by expanding the view of how organizations
can create sustained competitive advantage through elements such as meaning creation, positive
emotion cultivation, and high-quality connections (Cameron et al., 2003). Scholars interested in
the positive outcomes of emotional labor should consider exploring questions related to coping
behaviors to differentiate those who are able to replenish their emotional resources, as well as
mechanisms for enhancing emotion management skills and self-efficacy.
Specific to the idea of coping, women of color described the root of their emotional
exhaustion extending beyond the boundaries of the work context. Scholars exploring employee
well-being should extend this work beyond the professional domain to explore how far-reaching
the effects of emotional labor may be on the employee’s personal life. It is clear from my
findings that more work needs to be done to unpack how positionality at work, at home, and in
society influences the emotions and emotional labor of different groups. Participants described
challenges balancing competing personal and professional demands, and there was often
emotional spillover from one domain to the other. Future research should aim to explore lagged
effects of emotional labor both on the job and at home. Emotional effort in the home
environment may have a similar lagged effect on emotional labor in the work environment. An
128
intersectional approach to analyzing emotional effort across the work–home boundary might
include exploring disparities in mental and physical health outcomes, quality of non-work
relationships, and changes to self-concept.
Lastly, I invite scholars interested in an intersectional approach to consider the
antecedents, experiences, and consequences of emotional labor at the intersection of gender and
specific racial groups while also taking into account other markers of difference like age,
neurodiversity, and sexual orientation. Aggregating racial groups into a White/non-White binary
oversimplifies a complex reality, obscuring the unique experiences of different groups of color.
Gendered and racialized emotions need to be understood in relation to other intersecting forms of
privilege and oppression. Men and women of color are not homogeneous groups, and there is a
diversity of personal and professional experiences within subgroups that must be further
explored.

129
Bibliography
Acker, J. (2006). Inequality regimes: Gender, class, and race in organizations. Gender & Society,
20(4), 441–464. https://doi.org/10.1177/0891243206289499
Acker, J. (2012). Gendered organizations and intersectionality: Problems and possibilities. Equality,
Diversity and Inclusion: An International Journal, 31(3), 214–224.
https://doi.org/10.1108/02610151211209072
Adamovic, M., Gahan, P., Olsen, J., Gulyas, A., Shallcross, D., & Mendoza, A. (2021). Exploring the
adoption of virtual work: The role of virtual work self-efficacy and virtual work climate. The
International Journal of Human Resource Management. Advance online publication.
https://doi.org/10.1080/09585192.2021.1913623
Agho, O. A. (2022). In public administration and public policy, can theory capture caste?
Administrative Theory & Praxis. Advance online publication.
https://doi.org/10.1080/10841806.2022.2025736
Amis, J. M., Mair, J., & Munir, K. A. (2019). The organizational reproduction of inequality. Academy
of Management Annals, 14(1), 195–230. https://doi.org/10.5465/annals.2017.0033
Ashforth, B. E., & Humphrey, R. H. (1993). Emotional labor in service roles: The influence of
identity. The Academy of Management Review, 18(1), 88–115. https://doi.org/10.2307/258824
Ashkanasy, N. M. (2003). Emotions in organizations: A multi-level perspective. In F. Dansereau & F.
J. Yammarino (Eds.), Multi-level issues in organizational behavior and strategy (Vol. 2, pp. 9–
54). Emerald Group Publishing Limited. https://doi.org/10.1016/S1475-9144(03)02002-2
Ashkanasy, N. M., Härtel, C. E. J., & Daus, C. S. (2002). Diversity and emotion: The new frontiers in
organizational behavior research. Journal of Management, 28(3), 307–338.
https://doi.org/10.1016/S0149-2063(02)00130-7
Awasthi, P., & Mastracci, S. H. (2021). Integrating emotive competencies in public affairs education.
Journal of Public Affairs Education, 27(4), 472–489.
https://doi.org/10.1080/15236803.2021.1947072
Bagozzi, R. P. (2003). Positive and negative emotions in organizations. In K. S. Cameron, J. E.
Dutton, & R. E. Quinn (Eds.), Positive organizational scholarship: Foundations of a new
discipline (pp. 176–193). Berrett-Koehler Publishers.
Bakker, A. B., & Oerlemans, W. G. M. (2011). Subjective well-being in organizations. In K. S.
Cameron (Ed.), The Oxford handbook of positive organizational scholarship.
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199734610.001.0001/oxfordhb-
9780199734610-e-014
Barboza-Wilkes, C. J., Le, T. V., & Resh, W. (2022). Deconstructing Burnout at the Intersections of
Race, Gender, and Generation in Local Government. Journal of Public Administration Research
and Theory. Advance online publication. https://doi.org/10.1093/jopart/muac018
Berry, S., Trochmann, M. B., & Millesen, J. L. (2022). Putting the humanity back into public human
resources management: A narrative inquiry analysis of public service in the time of COVID-19.
Review of Public Personnel Administration, 0734371X211069656.
https://doi.org/10.1177/0734371X211069656
130
Blaique, L., Ismail, H. N., & Aldabbas, H. (2022). Organizational learning, resilience and
psychological empowerment as antecedents of work engagement during COVID-19.
International Journal of Productivity and Performance Management. Advance online
publication. https://doi.org/10.1108/IJPPM-04-2021-0197
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived. Annual
Review of Psychology, 54(1), 579–616. https://doi.org/10.1146/annurev.psych.54.101601.145030
Bono, J. E., & Vey, M. A. (2005). Toward understanding emotional management at work: A
quantitative review of emotional labor research. In C. E. Hartel, W. J. Zerbe, & N. M. Ashkanasy
(Eds.), Emotions in organizational behavior (pp. 213–233). Erlbaum Publishers.
Borry, E. L. (2016). A new measure of red tape: Introducing the Three-Item Red Tape (TIRT) Scale.
International Public Management Journal, 19(4), 573–593.
https://doi.org/10.1080/10967494.2016.1143421
Boyd, N. M., & Martin, E. C. (2020). Sense of community responsibility at the forefront of crisis
management. Administrative Theory & Praxis. Advance online publication.
https://doi.org/10.1080/10841806.2020.1765288
Brodsky, A. (2021). Virtual surface acting in workplace interactions: Choosing the best technology to
fit the task. Journal of Applied Psychology, 106(5), 714–733.
https://doi.org/10.1037%2Fapl0000805
Brotheridge, C. M., & Lee, R. T. (2003). Development and validation of the Emotional Labour Scale.
Journal of Occupational and Organizational Psychology, 76(3), 365–379.
https://doi.org/10.1348/096317903769647229
Browne, I., & Misra, J. (2003). The intersection of gender and race in the labor market. Annual
Review of Sociology, 29, 487–513.
Cameron, K. S., Dutton, J. E., & Quinn, R. E. (2003). Positive organizational scholarship:
Foundations of a new discipline (1st ed.). Berrett-Koehler.
Campbell, J. L., Quincy, C., Osserman, J., & Pedersen, O. K. (2013). Coding in-depth semistructured
interviews: Problems of unitization and intercoder reliability and agreement. Sociological
Methods & Research, 42(3), 294–320. https://doi.org/10.1177/0049124113500475
Cartwright, A. (2022). A theory of racialized cultural capital. Sociological Inquiry, 92(2), 317–340.
https://doi.org/10.1111/soin.12479
Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge University
Press.
Charmaz, K. (2014). Constructing grounded theory (2nd ed.). SAGE.
Chawla, N., Gabriel, A. S., Rosen, C. C., Evans, J. B., Koopman, J., Hochwarter, W. A., Palmer, J. C.,
& Jordan, S. L. (2020). A person‐centered view of impression management, inauthenticity, and
employee behavior. Personnel Psychology, 74(4), 657–691. https://doi.org/10.1111/peps.12437
Cole, E. R. (2009). Intersectionality and research in psychology. American Psychologist, 64(3), 170–
180. https://doi.org/10.1037/a0014564
Collins, P. H. (2008). Black feminist thought: Knowledge, consciousness, and the politics of
empowerment (1st ed.). Routledge.
131
Colquitt, J. A., Scott, B. A., Rodell, J. B., Long, D. M., Zapata, C. P., Conlon, D. E., & Wesson, M. J.
(2013). Justice at the millennium, a decade later: A meta-analytic test of social exchange and
affect-based perspectives. Journal of Applied Psychology, 98(2), 199–236.
http://dx.doi.org.libproxy2.usc.edu/10.1037/a0031757
Corbin, J., & Strauss, A. (2014). Basics of qualitative research: Techniques and procedures for
developing grounded theory (4th ed.). SAGE.
Côté, S. (2005). A social interaction model of the effects of emotion regulation on work strain.
Academy of Management Review, 30(3), 509–530. https://doi.org/10.5465/amr.2005.17293692
Côté, S., Miners, C. T. H., & Moon, S. (2006). Chapter 1: Emotional intelligence and wise emotion
regulation in the workplace. In W. J. Zerbe, N. M. Ashkanasy, & C. E.J. Härtel (Eds.), Individual
and organizational perspectives on emotion management and display (Vol. 2, pp. 1–24).
Emerald Group Publishing Limited. https://doi.org/10.1016/S1746-9791(06)02001-3
Côté, S., van Kleef, G. A., & Sy, T. (2013). The social effects of emotion regulation in organizations.
In Emotional labor in the 21st century: Diverse perspectives on emotion regulation at work (pp.
79–100). Routledge.
Cottingham, M. D., Johnson, A. H., & Erickson, R. J. (2018). “I can never be too comfortable”: Race,
gender, and emotion at the hospital bedside. Qualitative Health Research, 28(1), 145–158.
https://doi.org/10.1177/1049732317737980
Crenshaw, K. (1989). Demarginalizing the intersection of race and sex: A Black feminist critique of
antidiscrimination doctrine, feminist theory and antiracist politics. University of Chicago Law
Forum, 1(1), 31.
Crenshaw, K. W. (1991). Demarginalizing the intersection of race and sex: A Black feminist critique
of antidiscrimination doctrine, feminist theory and antiracist politics. In K. Bartlett & R.
Kennedy (Eds.), Feminist Legal Theory: Readings in Law and Gender (pp. 57–80). Westview
Press.
Davis, R. S., & Stazyk, E. C. (2021). Ambiguity, appraisal, and affect: Examining the connections
between goal perceptions, emotional labour, and exhaustion. Public Management Review.
Advance online publication. https://doi.org/10.1080/14719037.2021.1909348
Davis, R. S., Stazyk, E. C., Kline, E. D., & Green, A. C. (2022). Experience, emotion, and exhaustion:
how unionization influences emotional labor. Review of Public Personnel Administration,
0734371X211068171. https://doi.org/10.1177/0734371X211068171
Delgado, R. (1989). Storytelling for oppositionists and others: A plea for narrative. Michigan Law
Review, 87(8), 2411–2441. https://doi.org/10.2307/1289308
Delgado, R., & Stefancic, J. (2017). Critical race theory: An introduction (3rd ed.). New York
University Press.
Desrochers, S., Hilton, J. M., & Larwood, L. (2005). Preliminary validation of the work-family role
integration-blurring scale. Journal of Family Issues, 26, 442–466.
Diefendorff, J., Croyle, M., & Gosserand, R. (2005). The dimensionality and antecedents of emotional
labor strategies. Journal of Vocational Behavior, 66, 339–357.
https://doi.org/10.1016/j.jvb.2004.02.001
132
Diefendorff, J. M., & Gosserand, R. H. (2003). Understanding the emotional labor process: A control
theory perspective. Journal of Organizational Behavior, 24(8), 945–959.
https://doi.org/10.1002/job.230
Diefendorff, J. M., Lee, F., & Hynes, D. (2021, June 28). Longitudinal designs for organizational
research. Oxford Research Encyclopedia of Business and Management.
https://doi.org/10.1093/acrefore/9780190224851.013.211
Ding, F., Lu, J., & Riccucci, N. M. (2021). How bureaucratic representation affects public
organizational performance: A meta‐analysis. Public Administration Review, 81(6), 1003–1018.
https://doi.org/10.1111/puar.13361
Dumas, T. L., Phillips, K. W., & Rothbard, N. P. (2013). Getting closer at the company party:
Integration experiences, racial dissimilarity, and workplace relationships. Organization Science,
24(5), 1377–1401. https://doi.org/10.1287/orsc.1120.0808
Duncan, G. A. (2005). Critical race ethnography in education: Narrative, inequality and the problem
of epistemology. Race Ethnicity and Education, 8(1), 93–114.
https://doi.org/10.1080/1361332052000341015
Durr, M., & Harvey Wingfield, A. M. (2011). Keep your ‘N’ in check: African American women and
the interactive effects of etiquette and emotional labor. Critical Sociology, 37(5), 557–571.
https://doi.org/10.1177/0896920510380074
Dutton, J. E., Workman, K. M., & Hardin, A. E. (2014). Compassion at work. Annual Review of
Organizational Psychology and Organizational Behavior, 1(1), 277–304.
https://doi.org/10.1146/annurev-orgpsych-031413-091221
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative
Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999
Edmondson, A. C. (2004). Psychological safety, trust, and learning in organizations: A group-level
lens. In R. Kramer & K. Cook (Eds.), Trust and distrust in organizations: Dilemmas and
approaches (pp. 239–272). Russell Sage Foundation.
Fredrickson, B. L. (2003). Positive emotions and upward spirals in organizations. In K. S. Cameron, J.
E. Dutton, & R. E. Quinn (Eds.), Positive organizational scholarship: Foundations of a new
discipline (pp. 163–175). Berrett-Koehler Publishers.
Fredrickson, B. L., & Kurtz, L. E. (2011). Cultivating positive emotions to enhance human
flourishing. In S. I. Donaldson, M. Csikszentmihalyi, & J. Nakamura (Eds.), Applied positive
psychology: Improving everyday life, health, schools, work, and society (pp. 35–48). Routledge.
Fredrickson, B. L., & Losada, M. F. (2005). Positive affect and the complex dynamics of human
flourishing. American Psychologist, 60(7), 678–686.
http://dx.doi.org.libproxy2.usc.edu/10.1037/0003-066X.60.7.678
Freiling, I., Krause, N. M., Scheufele, D. A., & Brossard, D. (2021). Believing and sharing
misinformation, fact-checks, and accurate information on social media: The role of anxiety
during COVID-19. New Media & Society, 14614448211011452.
https://doi.org/10.1177/14614448211011451
133
Gabriel, A. S., Butts, M. M., Chawla, N., da Motta Veiga, S. P., Turban, D. B., & Green, J. D. (2021).
Feeling positive, negative, or both? Examining the self-regulatory benefits of emotional
ambivalence. Organization Science. https://doi.org/10.1287/orsc.2021.1553
Gabriel, A. S., Ganster, M. L., Slaughter, J. E., & MacGowan, R. L. (2021). The emotional
complexities of the COVID-19 pandemic and organizational life. Industrial and Organizational
Psychology, 14(1–2), 85–89. https://doi.org/10.1017/iop.2021.7
Glomb, T. M., & Tews, M. J. (2004). Emotional labor: A conceptualization and scale development.
Journal of Vocational Behavior, 64(1), 1–23. https://doi.org/10.1016/S0001-8791(03)00038-1
Goleman, D. (1995). Emotional intelligence. Bantam Books.
Grandey, A. A. (2000). Emotional regulation in the workplace: A new way to conceptualize
emotional labor. Journal of Occupational Health Psychology, 5(1), 95–110.
https://doi.org/10.1037//1076-8998.5.1.95
Grandey, A. A., & Gabriel, A. S. (2015). Emotional labor at a crossroads: Where do we go from here?
Annual Review of Organizational Psychology and Organizational Behavior, 21(1), 323–349.
https://doi.org/10.1146/annurev-orgpsych-032414-111400
Grandey, A. A., & Melloy, R. C. (2017). The state of the heart: Emotional labor as emotion regulation
reviewed and revised. Journal of Occupational Health Psychology, 22(3), 407–422.
https://doi.org/10.1037/ocp0000067
Grandey, A. A., Rupp, D., & Brice, W. N. (2015). Emotional labor threatens decent work: A proposal
to eradicate emotional display rules. Journal of Organizational Behavior, 36(6), 770–785.
https://doi.org/10.1002/job.2020
Grandey, A., Diefendorff, J., & Rupp, D. E. (2013). Emotional labor in the 21st century: Diverse
perspectives on the psychology of emotion regulation at work. Routledge.
Grandey, A., Foo, S. C., Groth, M., & Goodwin, R. E. (2012). Free to be you and me: A climate of
authenticity alleviates burnout from emotional labor. Journal of Occupational Health
Psychology, 17(1), 1–14. https://doi.org/10.1037/a0025102
Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of
General Psychology, 2(3), 271–299. https://doi.org/10.1037%2F1089-2680.2.3.271
Grund, S., Lüdtke, O., & Robitzsch, A. (2019). Missing data in multilevel research. In S. E.
Humphrey & J. M. LeBreton (Eds.), The handbook of multilevel theory, measurement, and
analysis (pp. 365–386). American Psychological Association.
Guy, M. E., & Newman, M. A. (2013). Emotional labor, job satisfaction and burnout: How each
affects the other. In R. J. Burke, A. J. Noblet, & C. L. Cooper (Eds.), Human management in the
public sector (pp. 132–150). Elgar. https://doi.org/10.4337/9780857937322.00016
Guy, M. E., Newman, M. A., & Emel Ganapati, N. (2013). Managing emotions while managing
crises. International Journal of Emergency Services, 2(1), 6–20. https://doi.org/10.1108/IJES-07-
2012-0033
Guy, M. E., Newman, M. A., & Mastracci, S. H. (2008). Emotional labor: Putting the service in
public service. M.E. Sharpe, Inc.
134
Hall, E. V., Hall, A. V., Galinsky, A. D., & Phillips, K. W. (2018). MOSAIC: A model of
stereotyping through associated and intersectional categories. Academy of Management Review,
44(3), 643–672. https://doi.org/10.5465/amr.2017.0109
Hoang, T., Suh, J., & Sabharwal, M. (2022). Beyond a numbers game? Impact of diversity and
inclusion on the perception of organizational justice. Public Administration Review. Advance
online publication. https://doi.org/10.1111/puar.13463
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American
Psychologist, 44(Mar 89), 513–524.
Hobfoll, S. E. (2002). Social and psychological resources and adaptation. Review of General
Psychology, 6(4), 307–324.
Hobfoll, S. E., Halbesleben, J., Neveu, J.-P., & Westman, M. (2018). Conservation of resources in the
organizational context: The reality of resources and their consequences. Annual Review of
Organizational Psychology and Organizational Behavior, 5(1), 103–128.
https://doi.org/10.1146/annurev-orgpsych-032117-104640
Hochschild, A. R. (1983). The managed heart commercialization of human feeling. University of
California Press.
Holman, D., Martinez‐Iñigo, D., & Totterdell, P. (2008). Emotional labor, well‐being, and
performance. In The Oxford handbook of organizational well being.
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199211913.001.0001/oxfordhb-
9780199211913-e-014
Hsieh, C.-W. (2014). Burnout among public service workers: The role of emotional labor
requirements and job resources. Review of Public Personnel Administration, 34(4), 379–402.
https://doi.org/10.1177/0734371X12460554
Hülsheger, U. R., & Schewe, A. F. (2011). On the costs and benefits of emotional labor: A meta-
analysis of three decades of research. Journal of Occupational Health Psychology, 16(3), 361–
389. http://dx.doi.org.libproxy2.usc.edu/10.1037/a0022876
Humphrey, N. M. (2021a). Emotional labor and employee outcomes: A meta-analysis. Public
Administration. Advance online publication. https://doi.org/10.1111/padm.12818
Humphrey, N. M. (2021b). Racialized emotional labor: An unseen burden in the public sector.
Administration & Society, 54(4), 741–758. https://doi.org/10.1177/00953997211037583
Humphrey, N. M. (2021c). Emotional labor and professionalism: Finding balance at the local level.
State and Local Government Review, 53(3), 260–270.
https://doi.org/10.1177/0160323X211048847
Humphrey, R., Ashforth, B., & Diefendorff, J. (2015). The bright side of emotional labor. Journal of
Organizational Behavior, 36(6), 749–769. https://doi.org/10.1002/job.2019
Humphrey, R., Pollack, J., & Hawver, T. (2008). Leading with emotional labor. Journal of
Managerial Psychology, 23(2), 151–168. https://doi.org/10.1108/02683940810850790
Imose, R. A., & Finkelstein, L. M. (2018). A multilevel theoretical framework integrating diversity
and emotional labor. Group & Organization Management, 43(5), 718–751.
https://doi.org/10.1177/1059601118793549
135
Jeung, D.-Y., Kim, C., & Chang, S.-J. (2018). Emotional labor and burnout: A review of the
literature. Yonsei Medical Journal, 59(2), 187–193. https://doi.org/10.3349/ymj.2018.59.2.187
Johnson, R. E., Muraven, M., Donaldson, T. L., & Lin, S.-H. (Joanna). (2018). Self-control in work
organizations. In The self at work: Fundamental theory and research (pp. 119–144). Routledge.
Kahneman, D. (2013). Thinking, fast and slow (1st ed.). Farrar, Straus and Giroux.
Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., & Stone, A. A. (2004). A survey
method for characterizing daily life experience: The day reconstruction method. Science (New
York, N.Y.), 306(5702), 1776–1780. https://doi.org/10.1126/science.1103572
Kammeyer‐Mueller, J. D., Rubenstein, A. L., Long, D. M., Odio, M. A., Buckman, B. R., Zhang, Y.,
& Halvorsen‐Ganepola, M. D. K. (2013). A meta-analytic structural model of dispositional
affectivity and emotional labor. Personnel Psychology, 66(1), 47–90.
https://doi.org/10.1111/peps.12009
Kang, M. (2003). The Managed Hand: The commercialization of bodies and emotions in Korean
immigrant–owned nail salons. Gender & Society, 17(6), 820–839.
https://doi.org/10.1177/0891243203257632
Kanter, R. M. (1977). Men and women of the corporation. Basic Books.
Kim, E., Bhave, D. P., & Glomb, T. M. (2013). Emotion regulation in work groups: The roles of
demographic diversity and relational work context. Personnel Psychology, 66(3), 613–644.
https://doi.org/10.1111/peps.12028
Knox, C. C., & Haupt, B. “Brie” (Eds.). (2020). Cultural competency for emergency and crisis
management: Concepts, theories and case studies (1st ed.). Routledge.
Koopman, J., Lin, S.-H. (Joanna), Lennard, A. C., Matta, F. K., & Johnson, R. E. (2020). My
coworkers are treated more fairly than me! A self-regulatory perspective on justice social
comparisons. Academy of Management Journal, 63(3), 857–880.
https://doi.org/10.5465/amj.2016.0586
Kossek, E. E., Lautsch, B. A., & Eaton, S. C. (2006). Telecommuting, control, and boundary
management: Correlates of policy use and practice, job control, and work–family effectiveness.
Journal of Vocational Behavior, 68(2), 347–367. https://doi.org/10.1016/j.jvb.2005.07.002
Kreiner, G. E. (2006). Consequences of work–home segmentation or integration: A person–
environment fit perspective. Journal of Organizational Behavior, 27(4), 485–507.
https://doi.org/10.1002/job.386
Laster Pirtle, W. N., & Wright, T. (2021). Structural gendered racism revealed in pandemic times:
Intersectional approaches to understanding race and gender health inequities in COVID-19.
Gender & Society, 35(2), 168–179.
http://journals.sagepub.com/stoken/default+domain/DDTXBJB4Z53GG8QR6WRX/full
Leiter, M. P., & Maslach, C. (2000). Preventing burnout and building engagement: A complete
program for organizational renewal. Jossey-Bass.
Leiter, M. P., & Maslach, C. (2011). Areas of worklife manual (5th ed). Mind Garden, Inc.
136
Lin, W., Shao, Y., Li, G., Guo, Y., & Zhan, X. (2021). The psychological implications of COVID-19
on employee job insecurity and its consequences: The mitigating role of organization adaptive
practices. Journal of Applied Psychology, 106(3), 317. https://doi.org/10.1037/apl0000896
Linos, E., Ruffini, K., & Wilcoxen, S. (2021). Reducing burnout and resignations among frontline
workers: A field experiment. Journal of Public Administration Research and Theory, muab042.
https://doi.org/10.1093/jopart/muab042
Liu, D., Chen, Y., & Li, N. (2021). Tackling the negative impact of COVID-19 on work engagement
and taking charge: A multi-study investigation of frontline health workers. Journal of Applied
Psychology, 106(2), 185. https://doi.org/10.1037/apl0000866
Locke, K., Golden-Biddle, K., & Feldman, M. S. (2008). Making doubt generative: Rethinking the
role of doubt in the research process. Organization Science, 19(6), 907–918.
Lu, X., & Guy, M. E. (2019). Emotional labor, performance goal orientation, and burnout from the
perspective of conservation of resources: A United States/China comparison. Public
Performance & Management Review, 42(3), 685–706.
https://doi.org/10.1080/15309576.2018.1507916
Martin, J. (2000). Hidden gendered assumptions in mainstream organizational theory and research.
Journal of Management Inquiry, 9(2), 207–216. https://doi.org/10.1177/105649260092017
Martin, J., Knopoff, K., & Beckman, C. (1998). An alternative to bureaucratic impersonality and
emotional labor: Bounded emotionality at the Body Shop. Administrative Science Quarterly,
43(2), 429. https://doi.org/10.2307/2393858
Maslach, C., & Jackson, S. E. (1986). Maslach burnout inventory: Manual. Consulting Psychologists
Press.
Maslach, C., Leiter, M. P., & Jackson, S. E. (2012). Making a significant difference with burnout
interventions: Researcher and practitioner collaboration. Journal of Organizational Behavior,
33(2), 296–300. https://doi.org/10.1002/job.784
Maslach, C., Leiter, M. P., & Schaufeli, W. (2008). Measuring burnout. In The Oxford handbook of
organizational well being.
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199211913.001.0001/oxfordhb-
9780199211913-e-005
Mastracci, S., & Arreola, V. I. (2016). Gendered organizations: How human resource management
practices produce and reproduce administrative man. Administrative Theory & Praxis, 38(2),
137–149. https://doi.org/10.1080/10841806.2015.1130523
Mastracci, S., & Mourtgos, S. (2021). The ethics of emotional labor in public service: The case of
children’s services forensic interviewers. Human Service Organizations: Management,
Leadership & Governance, 45(3), 260–272. https://doi.org/10.1080/23303131.2021.1883787
Mastracci, S., & Sawbridge, Y. (2019). Showing you care: Emotional labor and public service work.
In H. Dickinson, C. Needham, C. Mangan, & H. Sullivan (Eds.), Reimagining the future public
service workforce (pp. 39–53). Springer. https://doi.org/10.1007/978-981-13-1480-3_3
May, V. M. (2015). Pursuing intersectionality, unsettling dominant imaginaries. Routledge.
http://ebookcentral.proquest.com/lib/socal/detail.action?docID=1924472
137
McCall, L. (2005). The complexity of intersectionality. Signs: Journal of Women in Culture and
Society, 30(3), 1771–1800. https://doi.org/10.1086/426800
Medina, P. S., & Azevedo, L. (2021). Latinx COVID-19 outcomes: Expanding the role of
representative bureaucracy. Administrative Theory & Praxis. Advance online publication.
https://doi.org/10.1080/10841806.2021.1910411
Mele, V., & Belardinelli, P. (2019). Mixed methods in public administration research: Selecting,
sequencing, and connecting. Journal of Public Administration Research and Theory, 29(2), 334–
347. https://doi.org/10.1093/jopart/muy046
Myin-Germeys, I., Kuppens, P., & Kuppens, M.-G.-. (2021). The open handbook of experience
sampling methodology: A step-by-step guide to designing, conducting, and analyzing ESM
studies. Independently published.
Nembhard, I. M., & Edmondson, A. C. (2011). Psychological safety: A foundation for speaking up,
collaboration, and experimentation in organizations. In The Oxford handbook of positive
organizational scholarship.
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199734610.001.0001/oxfordhb-
9780199734610-e-037
Netemeyer, R. G., Boles, J. S., & McMurrian, R. (1996). Development and validation of work–family
conflict and family–work conflict scales. Journal of Applied Psychology, 81(4), 400–410.
https://doi.org/10.1037/0021-9010.81.4.400
Ohly, S., Sonnentag, S., Niessen, C., & Zapf, D. (2010). Diary studies in organizational research: An
introduction and some practical recommendations. Journal of Personnel Psychology, 9(2), 79–
93. https://doi.org/10.1027/1866-5888/a000009
Ortiz, S. M., & Mandala, C. R. (2021). “There is queer inequity, but I pick to be happy”: Racialized
feeling rules and diversity regimes in university LGBTQ resource centers. Du Bois Review:
Social Science Research on Race, 18(2), 347–364. https://doi.org/10.1017/S1742058X21000096
Ozcelik, H. (2013). An empirical analysis of surface acting in intra-organizational relationships.
Journal of Organizational Behavior, 34(3), 291–309. https://doi.org/10.1002/job.1798
Pacheco-Vega, R., & Parizeau, K. (2018). Doubly engaged ethnography: Opportunities and
challenges when working with vulnerable communities. International Journal of Qualitative
Methods, 17(1), 1609406918790653. https://doi.org/10.1177/1609406918790653
Phillips, K. W., Dumas, T. L., & Rothbard, N. P. (2018). Diversity and authenticity. Harvard
Business Review, March-April 2018, 8.
Portillo, S., Humphrey, N., & Bearfield, D. A. (2022). Representative bureaucracy theory and the
implicit embrace of Whiteness & masculinity. Public Administration Review. Advance online
publication. https://doi.org/10.1111/puar.13477
Purdie-Vaughns, V., & Eibach, R. P. (2008). Intersectional invisibility: The distinctive advantages
and disadvantages of multiple subordinate-group identities. Sex Roles, 59(5–6), 377–391.
https://doi.org/10.1007/s11199-008-9424-4
Rafaeli, A., & Sutton, R. I. (1987). Expression of emotion as part of the work role. The Academy of
Management Review, 12(1), 23–37. https://doi.org/10.2307/257991
138
Rafaeli, A., & Sutton, R. I. (1989). The expression of emotion in organizational life. In Research in
organizational behavior (Vol. 11, pp. 1–42). JAI Press Inc.
Ray, V. (2019). A theory of racialized organizations. American Sociological Review, 84(1), 26–53.
https://doi.org/10.1177/0003122418822335
Ray, V., Herd, P., & Moynihan, D. (2022). Racialized burdens: Applying racialized organization
theory to the administrative state. Journal of Public Administration Research and Theory,
muac001. https://doi.org/10.1093/jopart/muac001
Razzante, R. J., Boylorn, R. M., & Orbe, M. P. (2021). Embracing intersectionality in co-cultural and
dominant group theorizing: Implications for theory, research, and pedagogy. Communication
Theory, 31(2), 228–249. https://doi.org/10.1093/ct/qtab002
Ridgeway, C. L. (2011). Framed by gender: How gender inequality persists in the modern world (1st
ed.). Oxford University Press.
Roberson, L., & Kulik, C. T. (2007). Stereotype threat at work. Academy of Management
Perspectives, 21(2), 24–40. https://doi.org/10.5465/amp.2007.25356510
Roberto, K. J., Johnson, A. F., & Rauhaus, B. M. (2020). Stigmatization and prejudice during the
COVID-19 pandemic. Administrative Theory & Praxis, 42(3), 364–378.
https://doi.org/10.1080/10841806.2020.1782128
Rodriguez, J. K., Holvino, E., Fletcher, J. K., & Nkomo, S. M. (2016). The theory and praxis of
intersectionality in work and organisations: Where do we go from here? Gender, Work &
Organization, 23(3), 201–222. https://doi.org/10.1111/gwao.12131
Rothgerber, H., Wilson, T., Whaley, D., Rosenfeld, D. L., Humphrey, M., Moore, A., & Bihl, A.
(2020). Politicizing the COVID-19 pandemic: Ideological differences in adherence to social
distancing. PsyArXiv. https://doi.org/10.31234/osf.io/k23cv
Rudolph, C. W., Allan, B., Clark, M., Hertel, G., Hirschi, A., Kunze, F., Shockley, K., Shoss, M.,
Sonnentag, S., & Zacher, H. (2021). Pandemics: Implications for research and practice in
industrial and organizational psychology. Industrial and Organizational Psychology, 14(1–2), 1–
35. https://doi.org/10.1017/iop.2020.48
Sayre, G. M., Chi, N.-W., & Grandey, A. A. (2021). Investment in emotional labor: Daily resource
spirals and recovery. Academy of Management Proceedings, 2021(1), 14316.
https://doi.org/10.5465/AMBPP.2021.14316abstract
Scollon, C. N., Kim-Prieto, C., & Diener, E. (2003). Experience Sampling: Promises and Pitfalls,
Strengths and Weaknesses. Journal of Happiness Studies, 4(1), 5–34.
https://doi.org/10.1023/A:1023605205115
Shuler, S., & Sypher, B. D. (2000). Seeking emotional labor. Management Communication Quarterly,
14(1), 50–89.
Smith, A. N., Watkins, M. B., Ladge, J. J., & Carlton, P. (2019). Making the invisible visible:
paradoxical effects of intersectional invisibility on the career experiences of executive Black
women. Academy of Management Journal, 62(6), 1705–1734.
https://doi.org/10.5465/amj.2017.1513
139
Snijders, T. A. B. (2011). Multilevel analysis: An introduction to basic and advanced multilevel
modeling (2nd ed.). SAGE.
Steelman, T., Nowell, B., Velez, A.-L., & Scott, R. (2021). Pathways of Representation in Network
Governance: Evidence from Multi-Jurisdictional Disasters. Journal of Public Administration
Research and Theory, muab004. https://doi.org/10.1093/jopart/muab004
Thiele, C., Laireiter, A.-R., & Baumann, U. (2002). Diaries in clinical psychology and psychotherapy:
A selective review. Clinical Psychology & Psychotherapy, 9(1), 1–37.
https://doi.org/10.1002/cpp.302
U.S. Census Bureau. (2020). American Community Survey 5-year estimates. Census Reporter profile
page for Los Angeles County, CA. http://censusreporter.org/profiles/05000US06037-los-angeles-
county-ca/
Vacharkulksemsuk, T., & Fredrickson, B. L. (2013). Looking back and glimpsing forward: The
broaden-and-build theory of positive emotions as applied to organizations. In A. B. Bakker (Ed.),
Advances in Positive Organizational Psychology (Vol. 1, pp. 45–60). Emerald Group Publishing
Limited. https://doi.org/10.1108/S2046-410X(2013)0000001005
Van Bavel. J. J., Baicker, K., Boggio, P. S., Valerio, C., Aleksandra, C., Mina, C., Crockett, M. J.,
Crum, A. J., Douglas, K. M., Druckman, J. N., Drury, J. (2020). Using social and behavioural
science to support COVID-19 pandemic response. Nature Human Behaviour, 4(5), 460–471.
http://dx.doi.org.libproxy1.usc.edu/10.1038/s41562-020-0884-z
A. (2009). How emotions regulate social life: The Emotions as Social Information (EASI) Model.
Current Directions in Psychological Science, 18(3), 184–188. https://doi.org/10.1111/j.1467-
8721.2009.01633.x
Van Kleef, G. A., & Fischer, A. H. (2016). Emotional collectives: How groups shape emotions and
emotions shape groups. Cognition and Emotion, 30(1), 3–19.
https://doi.org/10.1080/02699931.2015.1081349
Walter, F., & Bruch, H. (2008). The positive group affect spiral: A dynamic model of the emergence
of positive affective similarity in work groups. Journal of Organizational Behavior, 29(2), 239–
261. https://doi.org/10.1002/job.505
Wang, G., Seibert, S. E., & Boles, T. L. (2011). Chapter 1: Synthesizing what we know and looking
ahead: A meta-analytical review of 30 years of emotional labor research. In C. E. J. Härtel, N. M.
Ashkanasy, & W. J. Zerbe (Eds.), What have we learned? Ten years on (Vol. 7, pp. 15–43).
Emerald Group Publishing Limited. https://doi.org/10.1108/S1746-9791(2011)0000007006
Wee, E. X. M. (2021). Compassion during difficult times: Team compassion behavior, suffering,
supervisory dependence, and employee voice during COVID-19. Journal of Applied Psychology,
106(12), 1805. https://doi.org/10.1037/apl0001001
Wharton, A. S. (1993). The affective consequences of service work: Managing emotions on the job.
Work and Occupations, 20(2), 205–232. https://doi.org/10.1177/0730888493020002004
Wharton, A. S. (1999). The psychosocial consequences of emotional labor. Annals of the American
Academy of Political and Social Sciences, 561, 19.
Wharton, A. S., & Erickson, R. J. (1995). The consequences of caring: Exploring the links between
women’s job and family emotion work. The Sociological Quarterly, 36(2), 273–296.
140
Wilkins, A. C., & Pace, J. A. (2014). Class, race, and emotions. In J. E. Stets & J. H. Turner (Eds.),
Handbook of the sociology of emotions: Volume II (pp. 385–409). Springer Netherlands.
https://doi.org/10.1007/978-94-017-9130-4_18
Wingfield, A. H. (2007). The modern Mammy and the angry Black man: African American
professionals’ experiences with gendered racism in the workplace. Race, Gender & Class,
14(1/2), 196–212.
Wingfield, A. H. (2009). Racializing the glass escalator: Reconsidering men’s experiences with
women’s work. Gender & Society, 23(1), 5–26. https://doi.org/10.1177/0891243208323054
Wingfield, A. H. (2010). Are some emotions marked “Whites only”? Racialized feeling rules in
professional workplaces. Social Problems, 57(2), 251–268.
https://doi.org/10.1525/sp.2010.57.2.251
Wingfield, A. H. (2020). Where work has been, where it is going: Considering race, gender, and class
in the neoliberal economy. Sociology of Race and Ethnicity, 6(2), 137–145.
https://doi.org/10.1177/2332649220903715
Wingfield, A. H. (2021). The (un)managed heart: Racial contours of emotion work in gendered
occupations. Annual Review of Sociology, 47(1), 197–212. https://doi.org/10.1146/annurev-soc-
081320-114850
Wingfield, A. H., & Alston, R. S. (2014). Maintaining hierarchies in predominantly White
organizations: A theory of racial tasks. American Behavioral Scientist, 58(2), 274–287.
https://doi.org/10.1177/0002764213503329
Woody, A. (2021). Emotions and ambient racism in America’s Whitest big city. Social Problems,
spab074. https://doi.org/10.1093/socpro/spab074
World Health Organization. (2022). WHO coronavirus (COVID-19) dashboard.
https://covid19.who.int
Xanthopoulou, D., Bakker, A. B., Oerlemans, W. G. M., & Koszucka, M. (2018). Need for recovery
after emotional labor: Differential effects of daily deep and surface acting. Journal of
Organizational Behavior, 39(4), 481–494. https://doi.org/10.1002/job.2245
Yi-Feng Chen, N., Crant, J. M., Wang, N., Kou, Y., Qin, Y., Yu, J., & Sun, R. (2021). When there is a
will there is a way: The role of proactive personality in combating COVID-19. Journal of
Applied Psychology, 106(2), 199. https://doi.org/10.1037/apl0000865
Yoon, S. (2021). Working through an “infodemic”: The impact of COVID-19 news consumption on
employee uncertainty and work behaviors. Journal of Applied Psychology, 106(4), 501.
https://doi.org/10.1037/apl0000913
Zavattaro, S. M., & McCandless, S. A. (2020). Editor’s introduction: Our public service manifesto
during pandemic. Administrative Theory & Praxis, 42(2), 233–239.
https://doi.org/10.1080/10841806.2020.1752593
141
Appendices
Appendix 1A: Full Diary Protocol

Survey Items and Factor Loading
Dependent Variable: Emotional Labor
Genuine Emotion

How often did you express feelings of ___ on the job today? (i.e.,
contentment, happiness, enthusiasm, irritation, anxiety, sadness,
concern, fear, anger)
Glomb & Tews,
2004
Amplified Emotion

How often did you express feelings of ___ on the job when you
really did not feel that way today? (i.e., contentment, happiness,
enthusiasm, irritation, anxiety, sadness, concern, fear, anger)
Suppressed Emotion

How often did you keep feelings of ___ to yourself when you
really felt that way today? (i.e., contentment, happiness,
enthusiasm, irritation, anxiety, sadness, concern, fear, anger)
Event-Level Controls
External Interactions How often did you interact with the public today?
Internal Interactions How often did you interact with professional colleagues today?
Person-Level Controls
Gender With which gender do you most identify?
Race Which of these reflects how you would describe yourself? Check all that apply.
Age What year were you born?
Person-Level Perceptual Measures of the Organization
Prosocial Motivation
Factor
(Eigenvalue: 2.02)
Why are you motivated to do your work?
1. Because I care about benefiting others through my work
2. Because I want to help others through my work
3. Because it is important to me to do good work for others through my work
4. Because I want to have a positive impact on others
Climate of
Authenticity Factor
(Eigenvalue: 4.17)
1. If you show anxiety or distress with this team, it is held
against you (R)
2. Members of this team are able to discuss how they feel
about problems and issues
3. People on this team reject others for showing irritation or
frustration in the team (R)
4. It is safe to show how you really feel with this team
Grandey et al.,
2012
142
5. It is uncomfortable for team members to show sadness or
disappointment with each other (R)
6. No one on this team would deliberately act in a way that
disrespects another member’s feelings
7. Working with members of this team, expressions of
feelings are respected
Social Support 1. I am a member of a supportive work group
2. I don’t feel close to my colleagues (R)
Leiter & Maslach,
2011
Job-Level Measures
% Public
Engagement
1. Please indicate the average percentage of the week that
you spend directly engaging with the public

Cadence & Entries for Open-Ended Diary Prompts
Construct Prompt Cadence Entries
Emotional
Experience
Please describe the events that led to some of your emotional
experiences today.
Daily 961
What were some of the main sources of stress in your work
environment today?
Daily 959
What were some of the hardest things about your job today? Daily 1,190
Work–Home
Boundary
Did your job affect your home life (or your home life affect your job)
today? If so, how?
Daily 1,194
Please describe how your childcare responsibilities/arrangements
have changed since the outbreak of COVID-19.
32

Onboarding 46
COVID-19 How are you doing personally with the “stay at home” order?
(physical health, mental health, relationships, etc.)
Daily 965
Job Context Is there anything you would like to share about your workload or
work group? Please describe.
Midpoint &
Final
164
Are there any thoughts you would like to share about the
management/leadership environment? Please describe.
Midpoint &
Final
163
Is there anything you would like to share about your overall job
satisfaction? Please describe.
Midpoint &
Final
161
General Is there anything else you would like to share about your day that we
did not ask about directly?

Daily 1,094
Total Entries 6,897

32
Note: This question was only shown to those respondents who had previously indicated they
had children.
143

144
Appendix 1B: OLS Models Predicting Attrition
To assess whether certain groups systematically dropped out of the study, two logistic
regression models predicting missed days were run with gender, race, and workload variables as
predictors. None of the coefficients were statistically significant at the 5% level in either model,
helping to rule out systematic attrition between groups in the sample. This step is important in
determining whether or not a diary design is appropriate for all groups by accounting for a
degree of selection bias. Cases with missing data at the day level were not deleted. It is possible
to work with partially missing data in multilevel modeling, since this method does not hold the
assumption that there is an equal number of observations per respondent or that there are fixed
time points. Moreover, it is important not to delete cases with missing data, as the deletion of
these cases could result in biased parameter estimates (Grund et al., 2019; Snijders & Bosker,
2012).
MODEL 1 Missed Days =!
#
+ !
$
(gender) + !
%
(race) + !
&
(workload) + ɛ
MODEL 2 Missed Days =!
#
+ !
$
(gender*race) + !
%
(workload) + ɛ

Model 1 Model 2
White women
1.265*
(0.690)
men of color
0.682
(0.750)
women of color
-0.211
(0.685)
workload_factor
0.0776
(0.296)
0.0879
(0.295)
woman
0.268
(0.510)

employee of color
-0.506
(0.505)

Obs
R-squared
367
0.004
367
0.016

Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

145
Appendix 1C: Pandemic-Specific Considerations for Compassion, Positionality, and
Reflexivity
As frontline service providers during the pandemic, I considered all the participants to be
members of a vulnerable community, which necessitated extra care in considering my own
positionality, the tension between engaging versus exploiting subjects, as well as the ethics of
how I represented their emotional experiences (Pacheco-Vega & Parizeau, 2018). Pacheco-Vega
and Parizeau (2018) argued, “under the no-harm principle, we owe it to our communities to
undertake responsible scholarship that minimizes the possibilities of harm” (p. 10). Specifically,
as an outsider to the community of frontline public servants, in the research design, I reflected on
how to meaningfully incentivize participation without being coercive/extractive or
overburdening participants. While the diary design was used primarily for its research utility, a
diary method has also been shown to have therapeutic benefits (Thiele et al., 2002). Given the
heightened emotions at the onset of the COVID-19 pandemic, this method allowed for high-
quality research while also being compassionately designed with the well-being of participants in
mind.
I leveraged my outsider status to encourage respondents to articulate social dynamics that
might be assumed to be understood by a researcher who worked in the same context, but my
identity as a woman of color likely influenced the comfort level different respondents had
disclosing their emotional experiences. By engaging participants across the levels of the
organizational hierarchy, from entry-level to senior management, there was not a consistent
power asymmetry between myself and the participants. Maintaining a clear understanding of my
own positionality and systematically reflecting on its impact on the data collection and analysis
is a necessity for the rigor and integrity of the research (Pacheco-Vega & Parizeau, 2018).
Critical reflexivity in the data collection process involved paraphrasing the participants’
comments to check my own understanding during interviews and probing explicitly into how
their social identity informed their emotional experience. Additionally, I circulated a preliminary
report of findings to all participants to solicit their feedback on any areas where I may have
mischaracterized their experiences. Future scholars should engage a diverse research team to
mitigate some of the concerns of researcher positionality.

146
Appendix 1D: Primary and Secondary Coding Scheme Examples
Primary
Code
Secondary
Code
Examples
Internal
Interactions
Negative
Emotions
&
Appraisals
“Having to deal with colleagues taking out their frustrations on me.”

“Feeling disrespected, ignored, not listened to, not valued as important by my
supervisor.”
Positive
Emotions
&
Appraisals
“Had some pleasant collaborative and brief emails today between colleagues.”

“Led a virtual staff meeting, which went very well. It was really nice to see some of
the faces and was great at relieving some of the stress everyone is feeling.”
Emotional
Labor
“Allowing feelings of frustration to pass and being patient in responding to obtuse
questions from coworkers.”

“Presenting a positive face for subordinates while stressed out.”

“Keeping negative emotions private so as not to affect those around me, coworkers
and subordinates.”

“Trying to keep a group of new hires engaged and excited when I’m not engaged
and excited.”
External
Interactions
Negative
Emotions
&
Appraisals
“Dealing with people who have no respect for public servants or the law in
general.”

“Responding to upset constituents.”
Positive
Emotions
&
Appraisals
“He was in a good mood and pleasant to speak with … I was able to close the
case”

“Conference call with faith-based organizations and leaders was encouraging”

“It’s wonderful working with the community and changing lives”
Emotional
Labor
“Continuing to be polite and nice to someone, even though they are yelling at you
for no reason.”

“I listened to the frustration and the fear and spoke calmly about what we had
done and are doing.”

“Giving people the benefit of the doubt while educating them on the rule.”

147
Appendix 1E: Full OLS Models for Disaggregated Forms of Emotional Labor
1E.1 Modeled Without % Public Engagement
VARIABLES (1)
Genuine Pos
(2)
Genuine Neg
(3)
Amplified Pos
(4)
Amplified Neg
(5)
Suppressed Pos
(6)
Suppressed Neg
Avg Genuine Pos

0.270***
(0.0243)
0.356***
(0.0401)
0.222***
(0.0342)
0.139***
(0.0502)
-0.316***
(0.0287)
Avg Genuine Neg

0.338***
(0.0386)

-0.231***
(0.0404)
0.243***
(0.0453)
-0.0537
(0.0475)
0.379***
(0.0407)
Avg Amplified Pos

0.365***
(0.0341)
-0.189***
(0.0264)

-0.104***
(0.0357)
-0.0593
(0.0389)
0.469***
(0.0308)
Avg Amplified Neg 0.225***
(0.0275)
0.197***
(0.0245)
-0.103***
(0.0341)

0.285***
(0.0269)
0.142***
(0.0255)
Avg Suppressed Pos 0.185***
(0.0442)
-0.0570
(0.0458)
-0.0768*
(0.0430)
0.373***
(0.0674)

0.265***
(0.0364)
Avg Suppressed Neg -0.492***
(0.0442)
0.471***
(0.0334)
0.711***
(0.0385)
0.218***
(0.0392)
0.310***
(0.0498)

Avg Freq coworkers 0.0671
(0.100)
-0.0122
(0.0824)
-0.332***
(0.108)
0.170
(0.104)
-0.458***
(0.0989)
0.491***
(0.0835)
Avg Freq residents 0.244***
(0.0644)
-0.520***
(0.0753)
0.229***
(0.0713)
0.153*
(0.0844)
0.0788
(0.0677)
0.141***
(0.0515)
White Women

-0.250***
(0.0578)
-0.0325
(0.0441)
0.476***
(0.0640)
0.239***
(0.0703)
-0.142**
(0.0616)
-0.300***
(0.0551)
Men of Color

0.0355
(0.0928)
-0.332***
(0.0697)
0.770***
(0.0974)
-0.127
(0.0937)
0.319***
(0.0829)
-0.213***
(0.0689)
Women of Color

-0.331***
(0.0628)
0.150**
(0.0748)
0.736***
(0.0740)
0.370***
(0.0881)
0.188*
(0.104)
-0.501***
(0.0705)
Baby Boomers

0.158***
(0.0387)
-0.232***
(0.0390)
-0.211***
(0.0389)
-0.266***
(0.0444)
0.0592*
(0.0340)
0.218***
(0.0322)
Millennials

0.314***
(0.0746)
-0.220***
(0.0543)
0.464***
(0.0866)
-0.482***
(0.103)
0.342**
(0.142)
-0.124**
(0.0574)
Avg Authenticity

0.129***
(0.0445)
-0.114***
(0.0327)
-0.0363
(0.0414)
-0.0271
(0.0409)
-0.309***
(0.0451)
0.0568*
(0.0297)
Avg Support

0.0493*
(0.0255)
0.122***
(0.0268)
-0.0404
(0.0371)
0.00216
(0.0253)
0.0267
(0.0291)
-0.0257
(0.0249)
Avg Closeness

-0.124***
(0.0284)
0.0751***
(0.0173)
0.0703***
(0.0216)
0.178***
(0.0330)
-0.122***
(0.0225)
-0.0357*
(0.0188)
Observations
R-squared
1,026
0.657
1,026
0.697
1,026
0.693
1,026
0.664
1,026
0.674
1,026
0.781

148
1E.2 Modeled With % Public Engagement
VARIABLES
(1)
Genuine Pos
(2)
Genuine Neg
(3)
Amplified Pos
(4)
Amplified Neg
(5)
Suppressed Pos
(6)
Suppressed Neg
% Public Engagement -0.00176
(0.0119)
-0.00647
(0.00757)
0.0393***
(0.00885)
0.105***
(0.0120)
0.0140
(0.00950)
-0.0414***
(0.00806)
Avg Genuine Pos 0.270***
(0.0243)
0.350***
(0.0390)
0.190***
(0.0323)
0.139***
(0.0504)
-0.307***
(0.0283)
Avg Genuine Neg 0.338***
(0.0388)
-0.223***
(0.0398)
0.216***
(0.0393)
-0.0522
(0.0481)
0.363***
(0.0390)
Avg Amplified Pos 0.366***
(0.0341)
-0.186***
(0.0268)
-0.138***
(0.0355)
-0.0656*
(0.0384)
0.474***
(0.0310)
Avg Amplified Neg 0.228***
(0.0337)
0.206***
(0.0260)
-0.158***
(0.0394)
0.264***
(0.0332)
0.197***
(0.0295)
Avg Suppressed Pos 0.185***
(0.0444)
-0.0556
(0.0465)
-0.0837**
(0.0413)
0.295***
(0.0531)
0.266***
(0.0381)
Avg Suppressed Neg -0.493***
(0.0463)
0.466***
(0.0319)
0.727***
(0.0376)
0.264***
(0.0391)
0.320***
(0.0506)

Avg Freq coworkers 0.0652
(0.102)
-0.0191
(0.0843)
-0.284***
(0.105)
0.258***
(0.0964)
-0.442***
(0.0964)
0.431***
(0.0798)
Avg Freq residents 0.244***
(0.0642)
-0.518***
(0.0759)
0.211***
(0.0704)
0.0922
(0.0789)
0.0736
(0.0677)
0.152***
(0.0511)
White Women -0.251***
(0.0590)
-0.0362
(0.0437)
0.489***
(0.0620)
0.263***
(0.0682)
-0.134**
(0.0617)
-0.314***
(0.0578)
Men of Color 0.0371
(0.0914)
-0.325***
(0.0702)
0.720***
(0.0936)
-0.206**
(0.0947)
0.305***
(0.0848)
-0.168**
(0.0679)
Women of Color -0.330***
(0.0637)
0.155**
(0.0737)
0.687***
(0.0714)
0.219***
(0.0817)
0.175
(0.107)
-0.448***
(0.0729)
Baby Boomers 0.157***
(0.0383)
-0.234***
(0.0387)
-0.197***
(0.0380)
-0.197***
(0.0367)
0.0629*
(0.0326)
0.200***
(0.0299)
Millennials 0.315***
(0.0744)
-0.216***
(0.0549)
0.436***
(0.0827)
-0.463***
(0.0923)
0.334**
(0.145)
-0.0989*
(0.0545)
Avg Authenticity 0.129***
(0.0443)
-0.111***
(0.0326)
-0.0508
(0.0411)
-0.0636
(0.0394)
-0.313***
(0.0443)
0.0709**
(0.0301)
Avg Support 0.0475
(0.0294)
0.116***
(0.0256)
-0.000266
(0.0361)
0.108***
(0.0254)
0.0407
(0.0303)
-0.0664***
(0.0254)
Avg Closeness -0.125***
(0.0295)
0.0715***
(0.0176)
0.0906***
(0.0213)
0.209***
(0.0316)
-0.114***
(0.0239)
-0.0573***
(0.0190)
Observations
R-squared
1,026
0.657
1,026
0.697
1,026
0.699
1,026
0.715
1,026
0.675
1,026
0.788
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

149
Appendix 1F: Full ML Models for Within-Person Variation in Emotional Labor

(1)
Genuine Pos
(2)
Genuine Neg
(3)
Amplified Pos
(4)
Amplified Neg
(5)
Suppressed Pos
(6)
Suppressed Neg
Daily Genuine Pos

0.166***
(0.0602)
0.161***
(0.0595)
0.104**
(0.0484)
0.187***
(0.0639)
-0.137**
(0.0552)
Daily Genuine Neg 0.124***
(0.0482)

-0.159***
(0.0532)
0.0709
(0.0432)
-0.00456
(0.0581)
0.291***
(0.0474)
Daily Amplified Pos 0.133***
(0.0489)
-0.163***
(0.0545)

0.178***
(0.0433)
-0.118**
(0.0587)
0.305***
(0.0478)
Daily Amplified Neg 0.0757
(0.0573)
0.0965
(0.0647)
0.283***
(0.0620)

0.250***
(0.0700)
-0.0148
(0.0595)
Daily Suppressed Pos 0.105**
(0.0441)
-0.0238
(0.0500)
-0.0849*
(0.0491)
0.156***
(0.0408)

0.184***
(0.0447)
Daily Suppressed Neg -0.125**
(0.0527)
0.346***
(0.0565)
0.347***
(0.0556)
0.0133
(0.0478)
0.256***
(0.0621)

Daily Freq coworkers 0.224*
(0.117)
-0.0656
(0.133)
-0.0128
(0.131)
-0.0500
(0.110)
-0.223
(0.144)
0.128
(0.121)
Daily Freq residents 0.0973
(0.0873)
0.0691
(0.0980)
0.256***
(0.0958)
-0.0686
(0.0805)
-0.128
(0.106)
0.0260
(0.0899)
White Women -0.0229
(0.208)
-0.479**
(0.214)
0.266
(0.217)
-0.0889
(0.143)
-0.180
(0.200)
-0.180
(0.197)
Men of Color 0.252
(0.235)
-0.585**
(0.245)
0.665***
(0.245)
-0.489***
(0.169)
0.415*
(0.234)
0.0222
(0.225)
Women of Color -0.128
(0.224)
-0.0643
(0.232)
0.432*
(0.232)
0.165
(0.147)
0.0343
(0.209)
-0.393*
(0.210)
Baby Boomers -0.186
(0.142)
-0.159
(0.146)
-0.0120
(0.147)
-0.146
(0.0899)
0.0189
(0.130)
0.0746
(0.133)
Millennials 0.350
(0.242)
-0.372
(0.251)
0.535**
(0.252)
-0.438***
(0.164)
0.453*
(0.231)
-0.296
(0.230)
Prosocial Motivation 0.0816
(0.0894)
0.0397
(0.0921)
-0.0772
(0.0928)
-0.0985*
(0.0581)
-0.0116
(0.0828)
0.179**
(0.0834)
Avg Authenticity 0.134
(0.133)
-0.296**
(0.136)
0.0857
(0.138)
-0.0305
(0.0839)
-0.364***
(0.121)
-0.0807
(0.125)
Avg Support 0.128
(0.0894)
0.225**
(0.0922)
-0.0787
(0.0935)
-0.0612
(0.0599)
0.00164
(0.0849)
-0.0443
(0.0849)
Avg Closeness 0.0265
(0.0859)
0.0714
(0.0880)
0.0487
(0.0889)
0.0759
(0.0522)
-0.206***
(0.0763)
-0.0724
(0.0803)
Observations
# of groups
336
95
336
95
336
95
336
95
336
95
336
95
Standard errors in parentheses
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

150

Appendix 2A: OLS Models Predicting Attrition
MODEL 1 Missed Days =!
!
+ !
"
(gender) + !
#
(race) + !
$
(workload) + ɛ
MODEL 2 Missed Days =!
!
+ !
"
(gender*race) + !
#
(workload) + ɛ

Model 1 Model 2
White women
1.265*
(0.690)
men of color
0.682
(0.750)
women of color
-0.211
(0.685)
workload_factor
0.0776
(0.296)
0.0879
(0.295)
woman
0.268
(0.510)

employee of color
-0.506
(0.505)

Obs
R-squared
367
0.004
367
0.016

Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

151
Appendix 2B: Full Diary Protocol
Survey Items and Factor Loadings
Dependent Variables: Emotional Labor
Genuine Emotion

How often did you express feelings of ___ on the job today? (i.e.,
contentment, happiness, enthusiasm, irritation, anxiety, sadness,
concern, fear, anger)
Glomb & Tews,
2004
Amplified Emotion

How often did you express feelings of ___ on the job when you
really did not feel that way today? (i.e., contentment, happiness,
enthusiasm, irritation, anxiety, sadness, concern, fear, anger)
Suppressed Emotion

How often did you keep feelings of ___ to yourself when you
really felt that way today? (i.e., contentment, happiness,
enthusiasm, irritation, anxiety, sadness, concern, fear, anger)
Independent Variables of Interest
Intersectional
Identity
Gender x Race (i.e., White men, White women, men of color,
women of color)
Interaction Term
Individual Controls
Demographics 1. Gender
2. Race
3. Age (generational cohorts)

Organizational & Job-Related Controls
Interpersonal
Interaction
Did you interact with ____ today? How often? (i.e., the public,
professional colleagues)

Red Tape Factor
(Eigenvalue: 1.22)
According to the following characteristics, how would you
describe policies and procedures in your department?
1. Burden (0 “not burdensome”–7 “burdensome”)
2. Necessity (0 “necessary”–7 “not necessary”)
3. Effectiveness (0 “effective”–7 “ineffective”)
Borry, 2016
Climate of
Authenticity Factor
(Eigenvalue: 4.17)
1. If you show anxiety or distress with this team, it is held
against you (R)
2. Members of this team are able to discuss how they feel
about problems and issues
3. People on this team reject others for showing irritation or
frustration in the team (R)
4. It is safe to show how you really feel with this team
5. It is uncomfortable for team members to show sadness or
disappointment with each other (R)
6. No one on this team would deliberately act in a way that
Grandey et al.,
2012
152
disrespects another member’s feelings
7. Working with members of this team, expressions of
feelings are respected
Social Support I am a member of a supportive work group Leiter & Maslach,
2011
Representation 1. My ethnic background is adequately represented in my
department
2. My gender is adequately represented in my department

Cadence for Open-Ended Diary Prompts
Construct Prompt Cadence
Emotional
Experience
Please describe the events that led to some of your emotional
experiences today.
Daily
What were some of the main sources of stress in your work
environment today?
Daily
What were some of the hardest things about your job today? Daily
Work–Home
Boundary
Did your job affect your home life (or your home life affect your job)
today? If so, how?
Daily
Please describe how your childcare responsibilities/arrangements have
changed since the outbreak of COVID-19.
33

Onboarding
COVID-19 How are you doing personally with the “stay at home” order?
(physical health, mental health, relationships, etc.)
Daily
Job Context Is there anything you would like to share about your workload or work
group? Please describe.
Midpoint & Final
Are there any thoughts you would like to share about the
management/leadership environment? Please describe.
Midpoint & Final
Is there anything you would like to share about your overall job
satisfaction? Please describe.
Midpoint & Final
General Is there anything else you would like to share about your day that we
did not ask about directly?

Daily

33
Note: This question was only shown to those respondents who had previously indicated they
had children.
153
Appendix 2C: Overall Frequency of Emotional Labor & Differences by Group
Frequency of Emotional Labor Across All Participants
Amplified Emotions Suppressed Emotions Genuinely Expressed Emotions
happiness (mean = 0.224)
enthusiasm (mean = 0.218)
concern (mean = 0.185)
contentment (mean = 0.175)
irritation (mean = 0.127)
anxiety (mean = 0.065)
sadness (mean = 0.055)
fear (mean = 0.055)
anger (mean = 0.052)
irritation (mean = 1.134)
anxiety (mean = 0.627)
concern (mean = 0.539)
sadness (mean = 0.435)
anger (mean = 0.380)
fear (mean = 0.331)
contentment (mean = 0.260)
happiness (mean = 0.214)
enthusiasm (mean = 0.175)
happiness (mean = 1.172)
irritation (mean = 1.036)
concern (mean = 1.000)
contentment (mean = 0.841)
enthusiasm (mean = 0.779)
anxiety (mean = 0.503)
anger (mean = 0.474)
sadness (mean = 0.425)
fear (mean = 0.253)
Chi-Square Tests for Differences in Emotional Labor Across Groups
happiness χ
2
= 69.2***
enthusiasm χ
2
= 69.0***
concern χ
2
= 29.2***
contentment χ
2
= 80.8***
irritation χ
2
= 46.5***
anxiety χ
2
= 26.6**
sadness χ
2
= 16.7*
fear χ
2
= 34.8***
anger χ
2
= 16.7*
irritation χ
2
= 41.4***
anxiety χ
2
= 91.8***
concern χ
2
= 51.1***
sadness χ
2
= 83.2***
anger χ
2
= 38.2***
fear χ
2
= 74.2***
contentment χ
2
= 32.8***
happiness χ
2
= 26.8***
enthusiasm χ
2
= 35.1***
happiness χ
2
= 62.5***
irritation χ
2
= 73.0***
concern χ
2
= 105.6***
contentment χ
2
= 43.3***
enthusiasm χ
2
= 39.8***
anxiety χ
2
= 124.8***
anger χ
2
= 60.3***
sadness χ
2
= 77.5***
fear χ
2
= 103.5***

154
Appendix 2D: Group Means for Genuinely Expressed Emotions
Among the negative emotions, all groups seemed most comfortable expressing irritation
and concern. Interestingly, women of color seemed to genuinely express anxiety, sadness, fear,
and anger more often than White women and men of color, potentially signaling more emotional
freedom as non-prototypical members of their gender or racial groups. It is also noteworthy that
men of color genuinely expressed positive emotions more than any other group and had the
biggest gap between their mean values for positive versus negative emotions.

APPENDIX 2E: Full Results of Single-Axis vs. Double-Axis Models of Emotional Labor
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
VARIABLES Amplified
Pos Single
Amplified
Neg Single
Amplified
Pos Double
Amplified
Neg Double
Suppressed
Pos Single
Suppressed
Neg Single
Suppressed
Pos Double
Suppressed
Neg Double
Genuine
Pos Single
Genuine
Neg Single
Genuine
Pos Double
Genuine
Neg Double
women -0.172 0.152 -0.151 -0.353** -0.120 -0.128
(0.125) (0.227) (0.159) (0.172) (0.194) (0.200)
employees of color 0.508** -0.220 0.334 0.112 0.203 0.216
(0.220) (0.269) (0.257) (0.242) (0.197) (0.298)
White women -0.00590 -0.128 -0.201 -0.473*** -0.0641 -0.522***
(0.151) (0.172) (0.186) (0.154) (0.210) (0.176)
men of color 0.851** -0.799* 0.231 -0.134 0.317 -0.596*
(0.409) (0.473) (0.347) (0.338) (0.404) (0.349)
women of color 0.353 -0.0962 0.178 -0.253 0.0889 0.0480
(0.241) (0.264) (0.317) (0.286) (0.271) (0.347)
freq. colleagues 0.435 -0.0845 0.309 0.127 -0.359 0.561 -0.321 0.651 -0.218 0.360 -0.260 0.657
(0.287) (0.343) (0.296) (0.347) (0.261) (0.415) (0.272) (0.416) (0.337) (0.424) (0.349) (0.414)
freq. residents 0.592** -0.238 0.551** -0.169 0.127 0.265 0.140 0.294 0.402** -0.323 0.389** -0.226
(0.275) (0.279) (0.250) (0.251) (0.267) (0.250) (0.268) (0.245) (0.190) (0.347) (0.184) (0.306)
baby boomers -0.273 -0.249 -0.193 -0.383 0.0536 -0.0111 0.0294 -0.0684 -0.110 -0.0862 -0.0837 -0.275
(0.173) (0.237) (0.150) (0.280) (0.169) (0.172) (0.181) (0.184) (0.181) (0.181) (0.165) (0.198)
millennials 0.219 -0.659 0.284 -0.769 0.462 0.490 0.443 0.444 0.329 0.403 0.350 0.249
(0.279) (0.464) (0.283) (0.496) (0.427) (0.389) (0.434) (0.394) (0.250) (0.419) (0.252) (0.415)
red tape 0.0271 -0.109 0.0590 -0.163 -0.218** -0.0670 -0.227** -0.0899 -0.0398 -0.153* -0.0291 -0.228**
(0.0810) (0.0960) (0.0885) (0.108) (0.0828) (0.0879) (0.0886) (0.0960) (0.117) (0.0886) (0.124) (0.106)
authentic climate 0.188 -0.273* 0.129 -0.174 -0.245* -0.0714 -0.227* -0.0292 0.0268 -0.0441 0.00716 0.0952
(0.160) (0.162) (0.145) (0.120) (0.131) (0.158) (0.124) (0.158) (0.157) (0.151) (0.149) (0.147)
social support -0.140 -0.0537 -0.0960 -0.129 -0.0213 -0.112 -0.0347 -0.144 0.158 -0.0662 0.173 -0.172
(0.144) (0.113) (0.132) (0.119) (0.103) (0.114) (0.106) (0.119) (0.108) (0.115) (0.108) (0.124)
represent: ethnic 0.0325 -0.190 0.0244 -0.176 -0.0714 -0.0414 -0.0690 -0.0355 -0.0110 -0.0443 -0.0137 -0.0250
(0.0532) (0.117) (0.0527) (0.108) (0.0676) (0.0752) (0.0664) (0.0726) (0.0499) (0.0824) (0.0477) (0.0756)
represent: gender -0.0555 0.0608 -0.0425 0.0390 -0.0109 -0.123* -0.0148 -0.132* 0.00987 -0.117* 0.0142 -0.148**
(0.0556) (0.0680) (0.0566) (0.0629) (0.0393) (0.0690) (0.0415) (0.0704) (0.0363) (0.0675) (0.0369) (0.0641)
Obs 937 937 937 937 937 937 937 937 937 937 937 937
R-squared 0.613 0.462 0.622 0.486 0.546 0.597 0.547 0.602 0.611 0.495 0.612 0.548
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

156
APPENDIX 2F: Full Results of Double-Axis OLS Models of Discrete Emotions

Discrete Amplified Emotions
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.115 -0.0266 -0.0141 -0.107 -0.0271 -0.0263 -0.00505 -0.0736 0.102
(0.0763) (0.0576) (0.0468) (0.0746) (0.0541) (0.0447) (0.114) (0.113) (0.101)
men of color -0.403** -0.262 -0.202 -0.140 -0.244 -0.193 0.565* 0.582* 0.597**
(0.178) (0.162) (0.131) (0.175) (0.154) (0.120) (0.302) (0.307) (0.253)
women of color -0.118 -0.0269 -0.000847 0.104 -0.0302 -0.0387 0.402** 0.167 0.210
(0.121) (0.0874) (0.0746) (0.145) (0.0821) (0.0652) (0.200) (0.178) (0.147)
freq. colleagues 0.0713 0.0288 -0.00295 0.0911 0.0114 0.0351 0.269 0.214 0.148
(0.145) (0.115) (0.0973) (0.166) (0.108) (0.0902) (0.218) (0.239) (0.200)
freq. residents -0.172 -0.0579 -0.0364 0.203 -0.0456 -0.0516 0.359* 0.343* 0.432***
(0.106) (0.0834) (0.0675) (0.136) (0.0790) (0.0630) (0.186) (0.185) (0.154)
baby boomers -0.150 -0.132 -0.0954 -0.128 -0.107 -0.0968 -0.111 -0.0926 -0.231***
(0.111) (0.0940) (0.0772) (0.107) (0.0898) (0.0707) (0.115) (0.116) (0.0837)
millennials -0.344* -0.259 -0.218 -0.0364 -0.259 -0.180 0.384* 0.199 0.0174
(0.178) (0.166) (0.136) (0.241) (0.161) (0.123) (0.219) (0.193) (0.211)
red tape -0.0801 -0.0486 -0.0256 -0.0549 -0.0460 -0.0528* 0.0564 0.0460 -0.0328
(0.0514) (0.0365) (0.0305) (0.0501) (0.0345) (0.0285) (0.0646) (0.0682) (0.0603)
authentic climate -0.0726 -0.0472 -0.0502 -0.102* -0.0498 -0.0435 0.130 0.0828 0.130
(0.0560) (0.0410) (0.0331) (0.0552) (0.0391) (0.0316) (0.106) (0.110) (0.0877)
social support -0.0359 -0.0481 -0.0332 -0.00468 -0.0435 -0.0352 -0.106 -0.0587 -0.110
(0.0531) (0.0405) (0.0325) (0.0475) (0.0385) (0.0303) (0.0982) (0.0986) (0.0715)
represent: ethnic -0.0588 -0.0631* -0.0510* -0.0242 -0.0593* -0.0481* 0.0288 0.0169 -0.0105
(0.0377) (0.0365) (0.0300) (0.0394) (0.0349) (0.0270) (0.0410) (0.0389) (0.0317)
represent: gender 0.0175 0.0115 0.0175 0.00430 0.0148 0.00647 -0.0453 -0.0295 0.00329
(0.0243) (0.0213) (0.0184) (0.0237) (0.0201) (0.0157) (0.0457) (0.0400) (0.0294)
Obs 937 937 937 937 937 937 937 937 937
R-squared 0.491 0.481 0.529 0.559 0.488 0.495 0.630 0.599 0.684
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

157

Discrete Suppressed Emotions
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.302* -0.497*** -0.439*** -0.315** -0.378*** -0.416** -0.0767 -0.137 -0.189
(0.156) (0.131) (0.159) (0.126) (0.142) (0.162) (0.112) (0.113) (0.131)
men of color -0.547* -0.229 -0.189 -0.0517 -0.00455 -0.0343 0.241 0.129 -0.0525
(0.301) (0.317) (0.304) (0.297) (0.315) (0.305) (0.199) (0.206) (0.263)
women of color -0.418* -0.214 -0.208 -0.00253 -0.0681 -0.502* 0.145 0.0545 0.157
(0.246) (0.268) (0.293) (0.228) (0.259) (0.269) (0.186) (0.187) (0.218)
freq. colleagues 0.726* 0.542 0.318 0.625* 0.258 0.785** -0.280* -0.0451 -0.255
(0.388) (0.372) (0.421) (0.335) (0.360) (0.363) (0.150) (0.188) (0.191)
residents 0.0828 0.410* 0.184 0.219 0.357 0.136 0.0851 -0.0116 0.288
(0.239) (0.225) (0.239) (0.216) (0.223) (0.203) (0.154) (0.158) (0.187)
baby boomers -0.193 -0.135 -0.0205 -0.00237 -0.0748 0.00318 0.0782 0.0168 -0.0706
(0.164) (0.164) (0.181) (0.147) (0.169) (0.183) (0.104) (0.112) (0.122)
millennials 0.526 0.526 0.326 0.399 0.379 0.270 0.212 0.175 0.433
(0.340) (0.319) (0.379) (0.337) (0.366) (0.344) (0.246) (0.250) (0.290)
red tape -0.0530 -0.0388 -0.218** -0.0317 -0.0387 -0.112 -0.0984* -0.178*** -0.135**
(0.0886) (0.0878) (0.0992) (0.0805) (0.0898) (0.0970) (0.0582) (0.0472) (0.0622)
authentic climate -0.0556 0.0324 0.00873 -0.0455 0.0199 -0.107 -0.168** -0.179** -0.0461
(0.128) (0.136) (0.163) (0.114) (0.156) (0.131) (0.0705) (0.0704) (0.100)
social support -0.114 -0.178 -0.164 -0.0624 -0.123 -0.0888 -0.00370 -0.0197 -0.0323
(0.0984) (0.107) (0.122) (0.0929) (0.114) (0.0999) (0.0616) (0.0582) (0.0845)
represent: ethnic -0.0236 -0.0389 -0.0698 -0.0210 0.0149 -0.0268 -0.0433 -0.0340 -0.0445
(0.0526) (0.0576) (0.0744) (0.0529) (0.0741) (0.0678) (0.0393) (0.0394) (0.0429)
represent: gender -0.0415 -0.101 -0.106 -0.135** -0.161** -0.0413 -0.0107 -0.00548 -0.0161
(0.0490) (0.0624) (0.0745) (0.0580) (0.0686) (0.0564) (0.0254) (0.0243) (0.0291)
Obs 937 937 937 937 937 937 937 937 937
R-squared 0.628 0.673 0.559 0.632 0.574 0.554 0.571 0.556 0.542
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Appendix 2G: MLM Models Predicting Suppressed Emotions With Key Interactions

Authentic Climate x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
VARIABLES Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment
White women -0.326 -0.542*** -0.486*** -0.296 -0.204 -0.419** 0.0261 -0.0562 -0.336**
(0.199) (0.194) (0.187) (0.182) (0.168) (0.181) (0.130) (0.125) (0.159)
men of color -0.180 -0.119 -0.0830 0.252 0.332 0.145 0.379** 0.330** -0.294
(0.269) (0.254) (0.246) (0.235) (0.217) (0.235) (0.170) (0.161) (0.212)
women of
color
-0.242 -0.170 -0.210 0.0182 0.110 -0.489** 0.0119 -0.0945 -0.284
(0.219) (0.227) (0.219) (0.208) (0.192) (0.209) (0.151) (0.143) (0.190)
authentic
climate
-0.0294 0.0575 0.0737 0.0257 0.0797 0.0889 -0.0822 0.0263 0.191
(0.201) (0.193) (0.187) (0.180) (0.166) (0.180) (0.129) (0.124) (0.161)
Interactions: Intersectional Identity # Authentic Climate
White women 0.207 -0.0637 0.186 0.00497 0.0823 -0.0649 0.127 -0.00768 -0.116
(0.209) (0.202) (0.195) (0.185) (0.171) (0.186) (0.135) (0.127) (0.171)
men of color -0.159 0.00567 -0.224 0.0184 -0.0445 -0.201 -0.0890 -0.303* 0.0769
(0.289) (0.273) (0.264) (0.253) (0.234) (0.254) (0.183) (0.174) (0.227)
women of
color
-0.0803 -0.265 -0.475** -0.253 -0.564*** -0.542*** -0.325** -0.298** -0.431**
(0.215) (0.213) (0.206) (0.198) (0.183) (0.198) (0.143) (0.136) (0.176)
freq.
colleagues
0.0100 -0.0526 -0.170 -0.0577 -0.0592 0.186 -0.153 -0.135 -0.0879
(0.163) (0.142) (0.137) (0.141) (0.130) (0.138) (0.0973) (0.0971) (0.100)
residents 0.208 0.209* 0.0747 0.144 0.140 0.0778 -0.0339 -0.0620 0.0651
(0.127) (0.115) (0.110) (0.113) (0.104) (0.111) (0.0783) (0.0777) (0.0820)
baby boomers -0.205 0.207 0.125 0.0977 0.233* 0.251* 0.0454 0.0127 -0.0935
(0.158) (0.154) (0.149) (0.142) (0.131) (0.143) (0.103) (0.0974) (0.130)
millennials 0.522* 0.663** 0.215 0.553** 0.587** 0.400 0.209 0.155 0.608***
(0.267) (0.263) (0.254) (0.249) (0.230) (0.247) (0.177) (0.171) (0.213)
red tape 0.189* 0.0428 -0.161 -0.0265 -0.0105 -0.0844 -0.144** -0.225*** -0.251***
(0.106) (0.102) (0.0989) (0.0926) (0.0856) (0.0936) (0.0679) (0.0637) (0.0880)
social support -0.111 -0.171* -0.217** -0.0593 -0.131* -0.143* -0.0717 -0.0595 -0.142*
(0.0976) (0.0903) (0.0874) (0.0828) (0.0765) (0.0833) (0.0602) (0.0569) (0.0764)
ethnic
representation
0.0423 -0.00869 0.0386 0.0240 0.112*** 0.0200 0.0410 0.0173 -0.0299
(0.0460) (0.0450) (0.0436) (0.0409) (0.0378) (0.0413) (0.0299) (0.0281) (0.0387)
gender
representation
-0.123*** -0.129*** -0.150*** -0.173*** -0.183*** -0.0792** -0.00217 -0.00589 -0.00842
(0.0378) (0.0395) (0.0383) (0.0355) (0.0328) (0.0360) (0.0262) (0.0244) (0.0347)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

159
Social Support x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
VARIABLES Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment

White women -0.900* -0.158 -1.520*** 0.282 -0.582 0.415 0.170 0.299 0.655**
(0.536) (0.498) (0.498) (0.460) (0.430) (0.442) (0.315) (0.313) (0.329)
men of color 0.126 0.588 -0.0157 1.163** 0.361 1.885*** 1.406*** 1.581*** 1.542***
(0.685) (0.607) (0.606) (0.563) (0.526) (0.540) (0.386) (0.382) (0.402)
women of
color
-0.0332 1.531** 1.401* 2.134*** 2.195*** 3.202*** 2.299*** 1.904*** 3.031***
(0.758) (0.751) (0.750) (0.696) (0.650) (0.668) (0.477) (0.473) (0.497)
social support -0.127 -0.0270 -0.244* 0.144 -0.0827 0.206* 0.100 0.147* 0.161*
(0.141) (0.127) (0.125) (0.120) (0.112) (0.115) (0.0824) (0.0817) (0.0860)
Interactions: Intersectional Identity # Social Support
White women 0.150 -0.107 0.283** -0.170 0.101 -0.245** -0.0507 -0.111 -0.258***
(0.149) (0.139) (0.139) (0.130) (0.121) (0.124) (0.0889) (0.0881) (0.0927)
men of color -0.105 -0.192 -0.0239 -0.261* -0.00715 -0.493*** -0.301*** -0.364*** -0.491***
(0.185) (0.165) (0.164) (0.153) (0.143) (0.147) (0.105) (0.104) (0.110)
women of
color
-0.0463 -0.412** -0.377** -0.530*** -0.497*** -0.915*** -0.564*** -0.501*** -0.817***
(0.183) (0.181) (0.180) (0.168) (0.157) (0.161) (0.115) (0.114) (0.120)
authentic
climate
-0.00504 -0.0127 -0.0123 0.000853 -0.0122 -0.0678 -0.0866 -0.0859 0.150**
(0.122) (0.114) (0.113) (0.106) (0.0990) (0.102) (0.0726) (0.0720) (0.0757)
freq.
colleagues
0.0100 -0.0600 -0.212 -0.0511 -0.0884 0.182 -0.156 -0.141 -0.121
(0.163) (0.142) (0.136) (0.140) (0.131) (0.134) (0.0957) (0.0949) (0.0998)
residents 0.215* 0.233** 0.100 0.158 0.166 0.101 -0.0157 -0.0557 0.110
(0.126) (0.115) (0.110) (0.112) (0.105) (0.108) (0.0768) (0.0761) (0.0801)
baby boomers -0.190 0.259* 0.135 0.152 0.230* 0.344*** 0.116 0.0575 0.0430
(0.154) (0.148) (0.147) (0.138) (0.129) (0.132) (0.0944) (0.0936) (0.0985)
millennials 0.522** 0.671*** 0.229 0.529** 0.549** 0.389* 0.196 0.153 0.612***
(0.266) (0.257) (0.253) (0.245) (0.229) (0.235) (0.168) (0.166) (0.175)
red tape 0.178 0.0728 -0.215** 0.0133 -0.0406 -0.0196 -0.112* -0.193*** -0.140**
(0.108) (0.103) (0.103) (0.0954) (0.0892) (0.0916) (0.0654) (0.0648) (0.0682)
ethnic
representation
0.0265 -0.00550 0.0181 0.0190 0.0938*** 0.0170 0.0251 0.0102 -0.0298
(0.0437) (0.0412) (0.0410) (0.0384) (0.0359) (0.0368) (0.0263) (0.0261) (0.0274)
gender
representation
-0.118*** -0.125*** -0.139*** -0.169*** -0.177*** -0.0698** 0.00934 -0.000865 0.00578
(0.0384) (0.0383) (0.0384) (0.0354) (0.0331) (0.0340) (0.0243) (0.0241) (0.0253)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

160
Red Tape x Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
VARIABLES Irritation Anxiety Sadness Concern Fear Anger Enthusiasm Happiness Contentment

White women -0.447** -0.496*** -0.557*** -0.337* -0.214 -0.440** -0.117 -0.137 -0.322**
(0.192) (0.186) (0.193) (0.174) (0.163) (0.187) (0.120) (0.120) (0.153)
men of color -0.317 -0.110 -0.231 0.203 0.279 0.00985 0.252 0.271* -0.387*
(0.270) (0.251) (0.264) (0.233) (0.219) (0.254) (0.161) (0.162) (0.210)
women of
color
-0.281 -0.0243 -0.269 0.0801 0.242 -0.473** 0.0215 -0.0729 -0.240
(0.225) (0.230) (0.243) (0.212) (0.200) (0.234) (0.146) (0.147) (0.193)
red tape 0.326** 0.0974 0.119 -0.0891 0.119 -0.106 -0.310*** -0.351*** -0.571***
(0.165) (0.158) (0.165) (0.147) (0.138) (0.159) (0.102) (0.102) (0.131)
Interactions: Intersectional Identity # Red Tape
White women -0.180 -0.166 -0.510** 0.110 -0.272 -0.0202 0.338*** 0.191 0.563***
(0.211) (0.207) (0.222) (0.187) (0.177) (0.212) (0.129) (0.130) (0.180)
men of color -0.779** -0.423 -0.769** -0.468 -0.525* -0.447 -0.488** -0.122 0.217
(0.368) (0.340) (0.362) (0.313) (0.295) (0.347) (0.216) (0.217) (0.292)
women of
color
-0.0805 0.588* -0.0873 0.312 0.521* 0.155 0.199 0.241 0.266
(0.342) (0.347) (0.361) (0.326) (0.307) (0.350) (0.225) (0.226) (0.285)
authentic
climate
0.0154 -0.0586 -0.00476 -0.0107 -0.0549 -0.0672 -0.0414 -0.0736 0.138
(0.122) (0.117) (0.124) (0.108) (0.102) (0.119) (0.0747) (0.0750) (0.0986)
freq.
colleagues
0.0249 -0.0374 -0.168 -0.0409 -0.0153 0.176 -0.113 -0.108 -0.0688
(0.162) (0.141) (0.136) (0.140) (0.131) (0.137) (0.0964) (0.0968) (0.100)
residents 0.192 0.222** 0.0513 0.161 0.142 0.0764 -0.0432 -0.0907 0.0692
(0.124) (0.112) (0.109) (0.111) (0.104) (0.110) (0.0763) (0.0767) (0.0810)
baby boomers -0.195 0.278* 0.0886 0.0746 0.265* 0.145 -0.0492 -0.0582 -0.208
(0.160) (0.157) (0.163) (0.146) (0.137) (0.158) (0.101) (0.101) (0.129)
millennials 0.478* 0.520* -0.00777 0.502** 0.325 0.275 0.212 0.131 0.633***
(0.272) (0.266) (0.273) (0.252) (0.237) (0.266) (0.174) (0.175) (0.214)
social support -0.108 -0.148 -0.165* -0.0811 -0.0758 -0.151 -0.136** -0.0805 -0.216***
(0.100) (0.0941) (0.0990) (0.0870) (0.0819) (0.0954) (0.0601) (0.0604) (0.0788)
ethnic
representation
-0.00375 -0.0204 -0.0314 -0.0179 0.0640* -0.0397 -0.0356 -0.0160 -0.0778**
(0.0458) (0.0443) (0.0467) (0.0409) (0.0385) (0.0449) (0.0282) (0.0284) (0.0373)
gender
representation
-0.137*** -0.114*** -0.147*** -0.180*** -0.171*** -0.0889** -0.0245 -0.0171 -0.0157
(0.0405) (0.0410) (0.0437) (0.0377) (0.0355) (0.0418) (0.0260) (0.0261) (0.0352)
Obs 315 367 367 367 367 367 367 367 367
# of groups 90 90 90 90 90 90 90 90 90
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

161
Appendix 2H: Primary and Secondary Coding Scheme Examples
Primary
Code
Secondary
Code
Examples
Events Negative
Emotions
“It was a very stressful and frustrating day”
“Today I feel exhausted, irritated, tired, and cranky”
Positive
Emotions
“Today was a good day”
“I am so happy to do my job”
Emotional
Labor
“Presenting a positive face for subordinates while stressed out”
“Made an effort to pass along info in a calm, objective way so they don’t worry
or jump to conclusions”

162
Appendix 2I: Antecedents and Consequences of Emotional Events
Primary
Code
Secondary
Code
Examples
Antecedents Internal
Relationships
“Mainly it was resolving some relationship issues between coworkers”
“Dealing with negative attitudes from negative colleagues”
Client
Relationships
“Responding to upset constituents”
“Dealing with the public, who feel that they are above the rules and are
offended when you point out what they need to do”
Work–Life
Balance
Challenges
“I feel like I cannot escape work at home anymore”
“The job continues being a source of stress even when home, and home
remains a source of stress even at work”
Children “Stress regarding the kids doing their work for school.”
“Rough day with kids. Many interruptions, had to help them a lot.”
Consequences Emotional
Exhaustion
“Finding the constructive language to do so was challenging and emotionally
draining”
Cynicism “Pretending to care”
“The feeling that nothing will come of the effort”
Loss of Self-
Efficacy
“Working on projects without a sense of completion or accomplishment”
“I am unequipped”

Antecedents of Emotional Events
Among the diary entries, 2,175 described contextual details about some of the
antecedents of their emotional events. Forty-four percent of those entries described internal
relationships among colleagues, 29% mentioned challenges managing work–home boundaries,
16% referenced childcare challenges, and 11% described external relationships with members of
the public (see Table 2E.1). The columns in Table 8 reveal the percentage of responses within
each group that were attributed to the secondary theme identified within the primary code of
antecedents to emotional events.
Table 2E.1 Antecedents of Emotional Events
Men Women
White Men
(485)
Men of Color
(334)
White Women
(715)
Women of Color
(641)
44% Internal Relationships 41.4% 49.1% 42.2% 45.2%

Coworkers: 68.7%
Staff: 13.4%
Supervisors: 17.9%
Coworkers: 60.4%
Staff: 25.6%
Supervisors: 14.0%
Coworkers: 68.9%
Staff: 14.2%
Supervisors: 16.9%
Coworkers: 67.2%
Staff: 17.2%
Supervisors: 15.5%
29% Work–Life Challenges 28.9% 28.7% 29.4% 30.4%
16% Children 15.9% 14.7% 17.6% 15.0%
163
11% External Relationships 13.8% 7.4% 10.8% 9.4%
Internal Relationships. Men of color described internal relationships more often than
any other group as a percentage of their total entries relating to antecedents of emotional events,
at 49.1%, followed by women of color, White women, then White men. When it came to internal
relationships, 60–69% of entries referred to coworkers, with staff and supervisory relationships
being referenced far less often. Curiously, men of color were much more likely than other groups
in the sample to describe internal interactions with employees under their supervision as a source
of negative emotions, with 25.6% of responses making explicit references to staff or
subordinates. Men of color described feeling undermined when in leadership positions. As an
example, one man described a “constant backlash from staff when trying to move forward with
an issue” (5:1908). In interactions upwards in the hierarchy, with supervisors and leadership,
men of color described feeling misunderstood by management and unable to share their feelings.
Interestingly, men of color also referenced the inability of others to suppress their negative
emotions as a source of frustration. When asked about their challenges throughout the day, men
of color described other “employees showing signs of stress and anger” (5:1439) and “employees
unable to cope with the stress of work” (5:1549). One man in particular was frustrated by an
“irate employee because he was not prepared to deal with the stress of being a frontline provider
to the public” (5:1969). These comments suggest that men of color may expect others to manage
their emotions and suppress negativity in the way they do.
Interestingly, women of color more than any other group referenced office power
dynamics. They described empathizing with other colleagues who “brought up legitimate [race-
based] concerns about the workplace” (5:73), but when women of color broached the concern
with leadership, there was a pattern of their complaints “not being taken seriously” and “nothing
being done about it,” which made them feel angry and frustrated (5:1699). They also described
feeling excluded from social relationships in the workplace, with one woman writing, “It feels
like being back in elementary/middle/high school again with cliques” (5:797). More than any
other group, women of color described incidents of overt disrespect. Women of color described
navigating a range of sexist and racist “microaggressions” in the internal work environment,
even among women of color in leadership positions. As an example, one woman shared she was
tired of dealing with mistreatment her male colleagues in leadership did not have to endure,
writing she “felt one of my staff was rude in email when I inquired on [the] status of a project”
(5:1373). Another woman of color in leadership wrote, “I was on a conference call with other
City Department members, and it was a little frustrating being interrupted while speaking”
(5:984), something she said rarely, if ever, happened to White men or men of color. For other
women of color, it was “staff’s pettiness” (5:1089), “rudeness” (17:1162), and generally “being
harassed by colleagues that I know don’t care for me” (4:1139).
Regardless of their seniority in the organization, women of color described a constant
awareness of the impression they were making on others. One woman shared, “I’m Latin and a
female who served the U.S. Navy … does it change things … no, we are still treated in a
demeaning way” (48:254). Being a woman of color in a society structured to support patriarchy
and White supremacy was described as constantly demanding emotion management beyond the
work domain. Several women of color interviewed said they felt they were well equipped to
“code switch” because they were “forced to do it” on and off the clock. In reference to
navigating interpersonal challenges at work, one woman described being “raised with a lot of
164
childhood trauma, which actually makes it easier (I think) for me to manage some of what’s
going on now than [for] people who were raised more healthily than I was” (48:249).
White women more than any other group referenced the emotions of colleagues they
were not directly interacting with, sharing that observing “sniping between coworkers” was a
source of their own negative emotions (5:1876). White women described being the person their
colleagues turned to for social and emotional support, but providing that support could feel
taxing. As an example, one White woman wrote,
One coworker is very stressed about the city budget and potential layoffs, and I’m
not enjoying being the person they go to express that. I’m doing ok myself but not
interested in managing someone else’s stress right now. I guess my patience is
shorter than it is in better times. (5:1142)
White women often described feeling slighted by coworkers and unsupported by
supervisors. One White woman described “frustration with a coworker who just doesn’t listen”
(4:2097), while another shared, “I was feeling anxiety today over a group email interaction
which I thought was a subtle way of trying to make me feel bad” (5:1058). A woman in a male-
dominated team shared,
This group is not supportive. The group, being male (except me), creates
situations all the time where I am aware that I am not in the know about things.
(9:138)
White women often described feeling like they could not turn to leadership for emotional
support. They described “feeling disrespected, ignored, not listened to, not valued as important
by my supervisor” (4:207) and that supervisors would either “constantly put me down” (4:372)
or neglect what they perceived to be the responsibility to “call people out for rudeness” (17:615).
These sentiments seemed to culminate in feeling a lack of psychological safety and support in
the workplace among White women.
By contrast, White women in leadership positions often described the challenge of
managing the emotions of their staff. Whether it was “supervising worried employees” (4:13) or
“presenting a positive face for subordinates while stressed out” (4:15), White women often
referenced the emotional climate of the workplace. One leader in particular shared, “I think the
hardest [thing] is feeling helpless that my staff is having such a hard time” (4:458).
White men in the sample often described the need to work with difficult people and their
emotions. Whether it was described as “managing expectations” (4:1066) or “addressing other
people’s fears and concerns” (4:329), there was a sense among White men that they needed to try
to “keep others positive and steer away from the negative” (4:527). Many of the White men in
the sample interviewed suggested that emotional expression in general was more common
among women in the workplace. One White man interviewed who worked on a team of all men
said he believed women “display their emotions better, [and] know how to do it better,” so they
express their emotions more often (39:16). Interview responses from several White men in the
sample suggested their coworkers were overly sensitive, sharing they felt they needed to
constrain their behavior to manage the emotions of their colleagues, but that their colleagues’
emotions were not justified. As an example, one White man shared, “I can be kind of sharp-
witted sometimes, so I have to pay more attention to not making any of those snide remarks
because they quite often were misinterpreted or taken defensively” (26:2). Interestingly, White
165
men more than any other group had a tendency to describe their colleagues in disparaging terms,
calling them things like “morons” (4:230), “assholes” (4:232), and “lazy” (17:1483). In
interviews with White men, it often took multiple probes before they shared any emotional
experiences. This may be a function of less emotional attunement but could also be a response to
my own identity as a woman of color conducting the interviews.
Work–Life Challenges. All groups described challenges managing their work and home
boundaries at similar rates (29–30% of their entries describing antecedents), but different aspects
of the work–life negotiation emerged for each group. For women of color, the challenge of
keeping emotions from spilling over from one domain to another seemed to flow in both
directions. When describing negative experiences at work, women of color shared, “my husband
might be tired of hearing my stories” (12:19) and “my job ALWAYS affects my home life,
because I’m often too exhausted to be productive and engaged at home (housework, playing with
the kids, talking with the elders, etc.)” (12:35). For other women of color, the challenge had
more to do with keeping challenges in the personal life from affecting their work. As examples,
one woman of color shared it was challenging how “my home life is affecting how much work I
get done, and work is not getting 100% of my effort” (12:756). Another woman added that her
stress mostly came from “dealing with personal issues [and] trying to function to the best of my
ability at work [while] avoiding discussing these issues” (4:689). Women of color more than any
other group described personal life challenges stemming from familial responsibilities for both
children and elders. One woman described taking care of her dad in between video calls
(12:755), while another shared, “I’m having a problem managing care for my aging aunt, uncle,
and mother” (4:975). A consistent sentiment among this group was wishing they had the
flexibility to balance their own needs with those of the job, but they often coped by releasing
their anxiety, stress, and frustration privately at home, where they felt safer to do so. Other
women of color expressed they felt guilty venting about work and losing their patience around
their families, so they “compartmentalized” (12:991) by becoming distant (12:991) and in some
cases becoming anxious (12:1144) and depressed (12:1140).
White women similarly described increasingly blurred boundaries between work and
home responsibilities, with one woman sharing she felt “wiped out” because
They feel inseparable at this point. My job is constantly there; although I try not
to check email or think too much about work during non-hours, it’s ever-present.
(12:548)
Other White women described work and home responsibilities as “constantly bleeding
into one another” (12:965) and “so intermeshed at this point it’s hard to say which is
which” (12:969), with one woman adding, “it’s getting increasingly difficult to separate
the two in the sense that I cannot seem to stop the stress” (12:2139). This was especially
true of White women with caretaking responsibilities while working, which was
described as “overwhelming” (14:159).
By contrast, men predominantly wrote about work–life boundary challenges only flowing
in one direction. Men of color most often described the challenge of separating stress from work
so it would not affect their mood at home, but not necessarily the stress from home bleeding into
work. They shared that they “tend to be short at home when stressed from work” (12:649) and
“anxiety from work created a negative mood” (12:824) that made them “more easily irritated at
home.” The general sentiment from men of color was that it would be unprofessional to reveal
166
their stress or negative emotions on the job, and as a consequence, they “come home very, very
stressed out [and] get angry easy” (12:938). Interestingly, for White men, the stress seemed to
come from home and bleed into work, but not the other way around. White men described their
relationships with spouses becoming more emotionally effortful in the transition to COVID-19
“stay-at-home” orders than before. White men often described it as being “progressively harder
to focus on work-related tasks while working from home” (1:1077), often blaming family
members for being distracting. As an example, one man shared, “my wife does think since I’m
home I can stop here and there more or less for whatever” (12:62). While a full analysis of each
group’s changing dynamics across the work–home boundary is beyond the scope of this paper,
these findings suggest expectations and attitudes about balancing one’s home responsibilities
vary across groups.
Children. All groups mentioned challenges with childcare at similar rates (14.7–17.6%),
but again, the nature of the challenges varied by group. In heterosexual partnerships, most White
men described deferring homeschooling duties to their wives, who they noticed were becoming
burnt out and unable to take on the household responsibilities they had previously managed. As
an example, one White father shared,
Our child is needing to be homeschooled, taking much time and energy, which is
fine, but my wife does it, and it’s time she used to spend making me breakfast and
lunch to take to work, so now I’m doing that myself. (2:1)
Others added, “wife is getting burnt out dealing with kids all day while I have to work” (12:270)
and “since children are being homeschooled and my wife works, we are sharing responsibilities
much more in my time at home” (2:4). While some expressed empathy for their wives, others
described “extreme frustration in my family dynamic” that “led to some emotional outbursts”
(5:1695). Many White men shared that they felt guilty when their children expressed confusion
and frustration that they were home but could not play with them (12:267; 12:316), but others
described their children as increasingly “whiny,” “irritating,” and “a source of constant
interruptions.”
By contrast, White women were much more likely than White men to describe guilt and
concern about the well-being of children, including adult children outside the home. Many White
mothers shared that the anxiety and stress of their children weighed on them personally and
stayed with them throughout the day. One White mother expressed remorse that her job “doesn’t
allow me to give my daughter the attention she deserves” (12:1489), and another shared,
Toddler was having a day of big emotions, which is exhausting. Exhaustion leads
to the inability to self-regulate, so my husband and I had some tense interactions.
This of course bleeds into all other interactions with work.(4:614)
This concern extended to adult children no longer living at home. As an example, one White
woman shared,
My son suffers from serious anxiety, and he’s really struggling with being
confined at home, away from friends, with no sports. Being unable to see his
friends, play music with his band, play baseball, attend Dodger games, is
overwhelming him. He called me, and the conversation led to him getting very
angry, which I can’t respond to. This is the hardest part for me to handle. (5:813)
167
Many of the women with school-age children described taking on the remote learning
responsibilities, sharing,
One of my children was unable to get into the online learning, so I sat through 2
hours of schooling immediately after getting home, and my other child had been
given assignments to do with their parent as well. (1:12)
The abrupt transitions from work to homeschooling made many women feel exhausted, making
it difficult to be patient with their children (12:15). Several White women described feeling
guilty stepping away from their children out of concern that their development would suffer.
Another White woman added,
My kids need me and keep asking why I care about work so much. I’m only able
to work a max of 4 hours a day, and they come running in with legitimate needs
throughout. I’m supposed to work a 40-hour week. (12:599)
White women more than any other group referenced raising these childcare concerns with
leadership to try to advocate for change. One in particular shared how comforting it was to have
children normalized in meetings:
A colleague allowed her daughter to sit and listen to a minor coordination call,
and it really made me happy to see the normalization of family when I feel a little
pressure to pretend like I’m in a home office cocoon decontextualized from the
next room where my family are. (1:990)
Fathers of color described challenges similar to those of White women in the sense that
they primarily described concern over the well-being of their kids and desire to create a sense of
consistency and normalcy for them. Fathers of color described needing to step in to help their
children throughout the day, whether with meals or remote learning support, and shared that
those activities took time away from their “ability to focus on the task I was working on”
(12:589) and “made for interesting transitions to the day” (12:1404). Fathers of color also
expressed frustration with management’s insensitivity towards caretakers, though less often than
White women.
Mothers of color also described needing to make themselves available for homeschooling
during the day while working, but more than any other group, they described sharing those
responsibilities with spouses (2:20) and adult family members (2:17), some of whom lived in the
home. Particularly among Black- and Latina/Hispanic-identifying women, there was mention of
multigenerational living, which was a source of childcare support but also additional
interpersonal stress. One woman shared she felt pressure from adult family members, saying,
“my job makes me too tired and sore to keep up with my parenting and housework duties at
home, which causes A LOT of tension with the elders” (12:37). Similar to White women and
men of color, women of color described worrying about the well-being of their children and guilt
over time spent working, with one woman adding, “I feel so torn between work and family”
(14:1830). Like men of color, women of color often shared they were “trying to be positive in
front of my family” (14:739), suggesting that emotional effort occurred on and off the clock.
External Relationships. Interestingly, White men described external relationships with
the public more than any other group, accounting for 13.8% of their diary entries describing
antecedents, followed by White women, women of color, and lastly men of color. All groups
shared that engaging with confrontational, rude, and argumentative clients was frustrating and
168
required a lot of patience. White women were more likely than other groups to describe residents
as “condescending” (4:716) and “entitled” (5:1125), with one woman in particularly sharing her
biggest sources of stress were “to be frank, rich entitled White male constituents” (17:515).
Another shared a stressful experience trying to help a resident:
I answered a phone call from a man who started out simply attacking [the]
government in general. I basically let him rant in my ear for several minutes.
When I initially attempted to provide him an answer, he kept cutting me off.
(5:1109)
Women of color also described similar forms of gendered mistreatment, but more often than any
other group described their public service motivation as something intrinsic that made them
“happy” and “enthusiastic” when they were able to help the community (5:1914). One woman
wrote, “it’s wonderful working with the community and changing lives” (8:132). Another
woman shared she experienced a “healthy anxiety” when interacting with residents because she
cared so deeply about helping them, writing, “I want to make sure I do right by them” (5:2005).
Despite the self-described prosocial motivation, interactions with residents were at times
described as a source of stress, particularly when employees were subjected to mistreatment.
Several women of color described residents who became upset and began using expletives, and
they shared they felt their job was to remain calm and positive as part of their service role. As
examples, one woman of color shared her biggest challenges were “clients who are difficult,
verbally abusive, and/or manipulative” (17:331), and another added the difficulty was
“continuing to be polite and nice to someone, even though they are yelling at you for no reason”
(4:1485).
Consequences of Emotional Events
Only 793 diary entries described the consequences of emotional events. As shown in
Table 2E.2, 40% of those entries described feelings of reduced self-efficacy or loss of personal
accomplishment, 31% described developing a cynical and detached attitude about work, 24%
described being emotionally exhausted, and 5% mentioned burnout and job dissatisfaction
generally.
Table 2E.2: Consequences of Emotional Events
White Men
(176)
Men of Color
(76)
White Women
(256)
Women of Color
(285)
40% Reduced Self-Efficacy 46.6% 48.7% 39.8% 35.1%
31% Cynicism 35.8% 25.0% 31.6% 28.1%
24% Emotional Exhaustion 15.3% 22.4% 22.7% 30.2%
5% General Burnout 2.2% 3.9% 5.9% 6.7%

Reduced Self-Efficacy. Across all groups, feelings of reduced self-efficacy stemmed
from the combination of challenging interpersonal interactions, the strain of balancing work–
home responsibilities, and changes to their workload due to COVID. Descriptively, all four
groups expressed similar concerns about “slow progress on tasks that need to get done” (4:335)
and “not feeling I did enough” (17:938). Men of color described a reduced sense of
accomplishment and belief in their ability to do their job more frequently than any other group,
169
at 48.7% of their entries mentioning consequences, followed by White men, White women, and
lastly women of color. Interestingly, women were more likely than men to frame their lack of
personal accomplishment in comparison to others. White women described needing to exceed
performance expectations without making their male colleagues feel threatened to be taken
seriously in departments where women were underrepresented. One woman shared she felt she
could not openly express her irritation with a colleague for fear of retribution from her male
supervisors or being perceived as “overly” emotional instead of focused primarily on
performance outcomes. As examples, a White woman shared she was “feeling discouraged by
how much some other folks are accomplishing, and feeling like we aren’t doing enough” (4:462),
while a woman of color added she was stressed out by “seeing other coworkers accomplishing
more than me” (4:2053). Women again echoed a fear of being replaceable, writing, “anxiety
about whether or not I am ‘useful’ or if they will decide to send me on an unpaid furlough”
(17:671) and “I worry a great deal about someone complaining about my work performance to a
supervisor” (17:848).
Cynicism & Disengagement. White men appeared to be developing the most cynicism
and disengagement, with 35.8% of their entries mentioning consequences describing feeling
detached from their work, followed by White women, women of color, and lastly men of color.
Across all groups, participants wrote about it being “hard to get motivated to do work today”
(4:1057) and “not being able to self-start” (4:2207). Among White men, this lack of motivation
was attributed to feeling overworked, underappreciated, and unable to make a real impact. As
examples, White men wrote, “feeling like a pawn in someone else’s chess game” (4:35), “we
should be replaced by computers within 10 years” (20:42), and “management is pretty much
ready to sacrifice us whenever it suits them and enhances their position” (20:292).
Employees of color also described inequity and mistreatment in the workplace as a
source of their cynicism and disengagement. As an example, a man of color wrote,
A lot of people try to tell me what to do and not ask. What I mean by that is that a
lot of people that are not my boss in any shape or form tell me they need
something done right away and they don’t. So when I tell them no, they always
say they will go above my head. (4:2075)
Other employees of color shared sentiments of mistreatment and lack of appreciation,
writing things like “the hardest-working employees are the lowest-paid” (8:58), “there’s
always winners and losers, leadership picks sides, consensus is talked about, but only
rarely really sought” (9:129), and “we (the workforce) always suspected that management
viewed us as pawns, and unfortunately this pandemic has verified that assertion” (8:131).
Several women of color noted the futility of voicing their concerns:
They are open to listening to criticism and suggestions, but it never seems they act
to address them, so it becomes a fruitless effort that threatens to be used against
you later. You seem to become the problem person. (9:52)
The consequence of this cynicism was that employees of color described themselves as
increasingly “closed off” (12:990), “withdrawn” (12:992), and “teetering on the edge of
burnout” (17:1654). Employees of color described feeling resigned from the work
environment, with one woman of color writing, “I’ve steeled myself to the level of
nonsense I have to deal with daily—while at work, I feel very numb” (5:1546).
170
White women in particular described a heightened sensitivity to performance
expectations and fear of negative evaluations. This may stem from gendered assumptions
about competence that create the stereotype threat anxiety described by Roberson and
Kulik (2007). It is curious to see that White men were the group that described the most
cynicism and disengagement. As the group most privileged in society, it might be
expected they would have the fewest complaints. Alternatively, their relative status
advantage may lead White men to have higher expectations of their own treatment within
organizations, perhaps creating an entitlement that is more sensitive to perceived slights.
Among employees of color, incidents of inequity and mistreatment at work were
described as a major source of their cynicism and disengagement.

171
Appendix 3A: Full Diary Protocol
Survey Items and Factor Loadings
Dependent Variables: Burnout Factors
Overall Burnout
(Eigenvalue: 5.79)
All 16 items listed below Maslach &
Jackson, 1986
Cynicism/
Depersonalization
(Eigenvalue: 2.53)
1. I have become less interested in my work since I started
this job
2. I have become less enthusiastic about my work
3. I have become more cynical about whether my work
contributes to anything
4. I doubt the significance of my work
Emotional
Exhaustion
(Eigenvalue: 2.96)
5. I feel emotionally drained from work
6. I feel used up at the end of the workday
7. I feel tired when I get up in the morning and have to face
another day on the job
8. Working all day is really a strain for me
9. I feel burned out from my work
Loss of Personal
Accomplishment
(Eigenvalue: 2.33)
10. I can effectively solve the problems that arise in my work
(R)
11. I feel I am making an effective contribution to what this
organization does (R)
12. In my opinion, I am good at my job (R)
13. I feel exhilarated when I accomplish something at work
(R)
14. I have accomplished many worthwhile things in this job
(R)
15. At my work, I feel confident that I am effective at getting
things done (R)
Uncategorized 16. I just want to do my job and not be bothered
Independent Variables of Interest
Intersectional
Identity
Gender x Race Interaction Term
Suppressing Discrete
Emotions
How often did you keep feelings of ___ to yourself when you
really felt that way today?
Glomb & Tews,
2004
Amplifying Discrete
Emotions
How often did you express feelings of ___ on the job when you
really did not feel that way today?
Individual Controls
Demographics 1. Age
2. Education
3. Political Ideology

172
Job-Specific Perceptual Controls
Prosocial Motivation
Factor
(Eigenvalue: 2.02)
Why are you motivated to do your work?
1. Because I care about benefiting others through my work
2. Because I want to help others through my work
3. Because it is important to me to do good work for others
through my work
4. Because I want to have a positive impact on others

Red Tape Factor
(Eigenvalue: 1.22)
According to the following characteristics, how would you
describe policies and procedures in your department?
1. Burden (0 “not burdensome”–7 “burdensome”)
2. Necessity (0 “necessary”–7 “not necessary”)
3. Effectiveness (0 “effective”–7 “ineffective”)
Borry, 2016
Workload Factor
(Eigenvalue: 1.80)
1. I do not have time to do the work that must be done
2. I work intensely for prolonged periods of time
3. I have so much work to do on the job that it takes me
away from my personal interests
4. I have enough time to do what’s important in my job (R)
Leiter & Maslach,
2011
Organization-Specific Perceptual Controls
Climate of
Authenticity Factor
(Eigenvalue: 4.17)
1. If you show anxiety or distress with this team, it is held
against you (R)
2. Members of this team are able to discuss how they feel
about problems and issues
3. People on this team reject others for showing irritation or
frustration in the team (R)
4. It is safe to show how you really feel with this team
5. It is uncomfortable for team members to show sadness or
disappointment with each other (R)
6. No one on this team would deliberately act in a way that
disrespects another member’s feelings
7. Working with members of this team, expressions of
feelings are respected
Grandey et al.,
2012
Social Support I am a member of a supportive work group Leiter & Maslach,
2011
Work–Home Boundary Controls
Segmentation/
Integration Factor
(Eigenvalue: 2.17)
1. I don’t like to have to think about work while I’m at home
(R)
2. I prefer to keep work life at work (R)
3. I like to be able to leave work behind at the end of the
workday (R)
Kreiner, 2006
4. It is often difficult to tell where my work life ends and my
family life begins
5. I tend to integrate my work and family duties when I
work at home
6. In my life, there is a clear boundary between my career
Desrochers et al.,
2005
173
and my role as a parent or family member (R)
7. Because of my job, I can’t involve myself as much as I
would like in maintaining close relations with my family
8. Things I want to do at home do not get done because of
the demands my job puts on me
9. My home life interferes with my ability to fulfill
responsibilities at work, such as accomplishing daily tasks
and projects
10. Things I want to do at work do not get done because of
the demands of my family
Netemeyer et al.,
1996
11. Throughout the workday, I deal with personal and work
issues as they occur
12. I tend to not talk about work issues with my family (R)
13. I actively strive to keep my family and work life separate
(R)
14. I tend to not talk about personal issues with most people I
work with (R)
Kossek et al.,
2006
WFH Status Are you working from home?
Children at Home Dummy variable

Cadence for Open-Ended Diary Prompts
CONSTRUCT PROMPT CADENCE
Emotional
Experience
Please describe the events that led to some of your emotional
experiences today.
Daily
What were some of the main sources of stress in your work
environment today?
Daily
What were some of the hardest things about your job today? Daily
Work–Home
Boundary
Did your job affect your home life (or your home life affect your job)
today? If so, how?
Daily
Please describe how your childcare responsibilities/arrangements
have changed since the outbreak of COVID-19.
34

Onboarding
COVID-19 How are you doing personally with the “stay at home” order?
(physical health, mental health, relationships, etc.)
Daily
Job Context Is there anything you would like to share about your workload or
work group? Please describe.
Midpoint & Final
Are there any thoughts you would like to share about the
management/leadership environment? Please describe.
Midpoint & Final

34
Note: This question was only shown to those respondents who had previously indicated they
had children.
174
Is there anything you would like to share about your overall job
satisfaction? Please describe.
Midpoint & Final
General Is there anything else you would like to share about your day that we
did not ask about directly?

Daily

175
Appendix 3B: Eigenvalues for Burnout Factors Over Time
Overall Burnout Cynicism Exhaustion
Loss of
Accomplishment
Total 5.79 2.53 2.96 2.33
Onboarding 7.79 2.87 2.95 3.32
Midway 5.72 2.19 3.03 2.16
Final 6.33 2.69 3.16 1.97

176
Appendix 3C: Unconditional Models Predicting Variation in Burnout
Several unconditional models predicting aggregated and disaggregated burnout,
measured at three time points per person, were used to estimate intraclass correlation and assess
the extent to which variation in burnout was attributed to differences within versus between
individuals. In total, 87.8% of the variation in overall burnout was explained by between-person
differences (see Table 3.1). When disaggregated into the three discrete dimensions of burnout,
81.5% of the variation in cynicism, 79% of the variation in emotional exhaustion, and 77.6% of
the variation in loss of personal accomplishment was explained by between-person differences.
This suggests that the majority of burnout in all its forms had to do with differences between
people (e.g., demographics, job characteristics, home-life characteristics). Loss of personal
accomplishment seemed to be most sensitive to day-to-day changes for a single individual,
followed by emotional exhaustion and, to a lesser extent, cynicism.
Table 3.1: Percent Variation of Burnout Explained
Between Within
Overall Burnout 87.8% 12.2%
Cynicism 81.5% 18.5%
Emotional Exhaustion 79.0% 21.0%
Loss of Personal Accomplishment 77.6% 22.4%

177
Appendix 3D: Intersectionality & EL Interactions for Burnout
(1)
Burnout All
(2)
Cynicism
(3)
Exhaustion
(4)
Accomplishment
White women -2.213***
(0.258)
-1.899***
(0.389)
-2.196***
(0.296)
-0.944***
(0.331)
men of color 0.377
(0.296)
-0.414
(0.404)
-0.603*
(0.349)
0.985***
(0.351)
women of color 0.432**
(0.215)
0.149
(0.330)
0.253
(0.252)
0.370
(0.279)
genuinely expressed
positive emotions
-1.093***
(0.120)
-0.611***
(0.169)
-0.844***
(0.141)
-0.775***
(0.144)
genuinely expressed
negative emotions
-0.0744
(0.188)
-0.149
(0.253)
-0.655***
(0.212)
0.699***
(0.216)
amplified
positive emotions
-0.628*
(0.363)
-0.268
(0.432)
1.188***
(0.431)
-1.539***
(0.357)
amplified
negative emotions
-0.438***
(0.137)
-0.294
(0.213)
-0.458***
(0.162)
-0.0625
(0.177)
suppressed
positive emotions
0.279
(0.230)
-0.225
(0.276)
1.150***
(0.260)
-0.770***
(0.232)
suppressed
negative emotions
-0.393*
(0.230)
0.0277
(0.263)
-1.473***
(0.262)
0.214
(0.221)
Interaction Terms: Intersectional Identity * Amplification of Positive Emotions
White women -0.0277
(0.498)
-0.298
(0.431)
-1.937***
(0.583)
1.280***
(0.354)
men of color 0.778**
(0.333)
0.333
(0.426)
-0.754*
(0.394)
1.778***
(0.352)
women of color -2.576***
(0.665)
-0.908
(0.745)
-4.780***
(0.737)
-0.773
(0.638)
Interaction Terms: Intersectional Identity * Suppressed of Negative Emotions
White women 0.452*
(0.263)
0.360
(0.338)
1.469***
(0.308)
-0.387
(0.281)
men of color 0.344
(0.302)
0.590*
(0.319)
0.969***
(0.362)
0.716***
(0.274)
women of color 2.769***
(0.583)
1.823***
(0.608)
3.427***
(0.645)
1.238**
(0.521)
baby boomers -0.0275
(0.103)
-0.511***
(0.145)
0.0988
(0.110)
0.529***
(0.133)
millennials 0.745**
(0.341)
0.414
(0.481)
-0.800*
(0.411)
1.712***
(0.392)
education -0.0617
(0.127)
-0.0209
(0.123)
-0.510***
(0.140)
0.479***
(0.106)
political ideology -0.106
(0.0778)
-0.0596
(0.0734)
0.0334
(0.0846)
-0.235***
(0.0632)
prosocial motivation 0.253*** -0.0562 0.303*** 0.221**
178
(0.0762) (0.118) (0.0913) (0.0974)
freq. colleagues 1.440***
(0.558)
1.227*
(0.700)
1.862***
(0.599)
-0.176
(0.612)
freq. residents 0.782***
(0.153)
0.477***
(0.173)
0.497***
(0.169)
0.303**
(0.145)
lives with children -0.945***
(0.203)
-0.377
(0.262)
-1.694***
(0.238)
0.241
(0.222)
work–home preference 0.0247
(0.114)
-0.276**
(0.125)
-0.0454
(0.136)
0.0359
(0.108)
remote worker 1.900***
(0.505)
0.448
(0.728)
3.086***
(0.597)
0.470
(0.594)
hybrid (mostly home) 0.936***
(0.302)
0.0878
(0.415)
1.055***
(0.355)
0.485
(0.342)
hybrid (mostly in-
person)
-1.110**
(0.561)
-1.007*
(0.568)
0.278
(0.665)
0.470
(0.470)
workload 0.101*
(0.0611)
0.0114
(0.0903)
0.272***
(0.0703)
0.0378
(0.0763)
Red tape -0.316***
(0.105)
-0.447***
(0.151)
-0.352***
(0.122)
-0.0219
(0.128)
authentic climate -0.412***
(0.0664)
-0.432***
(0.0911)
-0.129*
(0.0737)
-0.353***
(0.0819)
social support -0.0783**
(0.0306)
-0.170***
(0.0438)
0.0293
(0.0358)
-0.0661*
(0.0365)
Obs
# Groups
118
61
136
63
127
61
128
63
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

179
Appendix 3E: Primary and Secondary Coding Scheme Examples
Primary Code: Cynicism/Disengagement

Personal Life
“I should be happy being home with the kids, but just the fact that we can’t leave makes it
seem like imprisonment.”

“Mentally, this is getting harder. I am beginning to think the isolation will not end soon. I am
beginning to feel my personal rights as guaranteed by the Constitution are being infringed
upon.”

Job Design
“Not having more control, not having any input. Feeling like a pawn in someone else’s chess
game.”

“The fact that it’s pointless and we’re working this hard for nothing.”
Emotional
Experience
“An emotionally bankrupt letter from our general manager, where he tried to convince us
that we are a family and that he had our best interests at heart.”
Primary Code: Emotional Exhaustion

Personal Life
“I REALLY miss hugs. I miss going to parties outside of Zoom. My depression and anxiety
are still looming, but I don’t have the time/space to do much about them.”

“I find myself to be more sluggish and irritable. If we are able to telecommute after this is
over [for] a week, I’d like it. Right now I don’t because there is no other outlet.”

Job Design
“The stress of such an overwhelming job with not enough resources and an unhelpful
supervisor made me exhausted.”

“It’s starting to wear me down a bit. The days all run together, I have many responsibilities
that must be fulfilled at the same time, and it’s getting stressful.”

“It’s hard when people just want to vent to you but you have 15 other calls in the queue that
you need to respond to.”
Emotional
Experience
“Had to take corrective action on a few items at work and guide discussion to the appropriate
solutions. Finding the constructive language to do so was challenging and emotionally
draining.”
Primary Code: Loss of Personal Accomplishment

Personal Life
“I feel like I have no structure and nothing or not enough is being completed.”

“I should like being home, but it doesn’t feel right. When it’s slow at work, I feel like I’m sort
of useless”

Job Design
“It drives me crazy when work issues are left undecided and you have to wait till the next day
to find out what the next steps are.”

“Very long meetings that interrupted my regular work assignments, so I’ve fallen behind on
work.”
180

Emotional
Experience
“Dealing with angry people with no authority to actually help them.”

“I think the hardest is feeling helpless that my staff is having such a hard time.”

Abstract (if available)

Abstract Emotional labor research in public administration lags behind other fields, is often omitted from discussions of representative bureaucracy, and rarely looks at its gendered and racialized dimensions. The existing scholarship fails to consider the dynamic nature of emotions and that different emotions (e.g., happiness versus anger) might warrant different emotional labor techniques for different groups. Meanwhile, scholars from sociology, applied psychology, and organizational behavior widely recognize the importance of emotional labor, but few have used an intersectional lens to study the well-recognized phenomenon.
This dissertation uses an intersectional approach to codify the difficult-to-measure and often unobserved emotional labor that can institutionalize inequity within public organizations. An intersectional approach is essential to make visible the experiences of those at the intersection of multiple marginalized identities, and this dissertation describes in detail how the antecedents, experiences, and consequences of emotional labor differ based on the employee’s combination of gender and racial identity. Using a mixed-methods research design that combines daily diary entries and semi-structured interviews, this work (1) describes and measures the emotional labor embedded in both service encounters with the public and internal interactions among colleagues, (2) looks at subgroup differences in the emotional effort at the intersection of race and gender, and (3) assesses the relationship between emotional labor and burnout to inform our understanding of the well-being of a diverse public sector workforce.
I find meaningful differences within and between individuals in the emotions needed to effectively engage the public and navigate public institutions. The results reveal that, compared to their peers, women of color engage in more taxing forms of emotional labor, feel more emotionally constrained by organizational rules, are more cognizant of managing gendered and racialized stereotypes, and are more sensitive to whether the climate allows for authentic expression. I also show that public employees experienced heightened burnout during the pandemic, and the suppression of emotion contributed to that burnout, but in different ways for different groups. In particular, women of color who suppressed negative emotions were more likely to experience a reduced sense of personal accomplishment, increased cynicism, disengagement from their work, and more emotional exhaustion.
This project reveals important distinctions in the type of emotional labor demanded of public employees and how those emotional demands differ across gender and racial identities. The results make visible the experiences of those at the margins of multiple lived experiences of oppression, allowing women of color to articulate their own emotional experiences in ways that center their voices. Importantly, this work highlights the importance of factoring emotional labor into the experience of burnout at work while emphasizing that the relationship between the two varies for individuals of different backgrounds. I provide concrete proof that there is an uneven distribution of emotional labor in public organizations, and it falls predominantly on women of color.
Measuring a construct as complex and dynamic as emotional labor lays the groundwork for important reform. By codifying, measuring, and describing the differential emotional burdens embedded in public organizations, I quantitatively demonstrate the need for equitable human resource management practices that address how organizations structurally reinforce inequity.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Technological innovation in public organizations

PDF

Emotional counterpublics: the feeling of racial justice

PDF

Diverging pathways to citizenship and immigrant integration in the U.S.

PDF

Participation in higher education diversity, equity, and inclusion work: a relational intersectionality of organizations analysis

PDF

The role of social media for community-based organizations focused on environmental justice in southeast Los Angeles

PDF

Black, Muslim, and gay/queer male allies: an intersectional analysis of men’s gender justice activism

PDF

How can metrics matter: performance management reforms in the City of Los Angeles

PDF

Affordable south Los Angeles: survival, support, and different futures

PDF

Broken windows on campus: Policing and racism in higher education

PDF

Environmental justice in real estate, public services, and policy

PDF

Goal setting and performance in federal agencies

PDF

Care and social-emotional well-being: organizational conditions in policy and practice

PDF

The color of success: African American and Japanese American physicians in Los Angeles

PDF

Changing workplace and inequality in work-family policies

PDF

Who will watch my children? How family and work structures shape low-income, single mothers' childcare use in Los Angeles

PDF

Beyond commitments: a qualitative examination of the persistent disparities faced by Black women in executive leadership roles post the 2020 crisis and beyond

PDF

More knowledge, better plans: a study of heterogeneity in dementia prevalence and mid-life cognitive changes

PDF

Outcomes-based contracting through impact bonds: ties to social innovation, systems change, and international development

PDF

Good intentions are not enough: exploring implicit racial bias in philanthropy

PDF

Crowded with potential: housing and social mobility strategies among China's educated migrants

Asset Metadata

Creator Barboza-Wilkes, Cynthia Jane (author)

Core Title Reproducing inequity in organizations: gendered and racialized emotional labor in pubic organizations

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Public Policy and Management

Degree Conferral Date 2022-08

Publication Date 07/19/2022

Defense Date 04/11/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag diversity management,emotional labor,intersectionality,OAI-PMH Harvest,Public Management

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Schweitzer, Lisa (committee chair), Beckman, Christine (committee member), Resh, William (committee member)

Creator Email cbarboza@usc.edu,cynthia.j.wilkes@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111373230

Unique identifier UC111373230

Legacy Identifier etd-BarbozaWil-10865

Document Type Dissertation

Format application/pdf (imt)

Rights Barboza-Wilkes, Cynthia Jane

Type texts

Source 20220719-usctheses-batch-956 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu