Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Strategic audience partitioning: antecedents and consequences
(USC Thesis Other)
Strategic audience partitioning: antecedents and consequences
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
STRATEGIC AUDIENCE PARTITIONING: ANTECEDENTS AND CONSEQUENCES
by
Francesca Valsesia
________________________________________________________________________
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
(BUSINESS ADMINISTRATION)
May 2018
Copyright 2018 Francesca Valsesia
i
ACKNOWLEDGMENTS
There are many people I am thankful to. First, my advisor Joe. You have been there for me
since day one, and I consider myself lucky for being your student. Thank you for giving me so much
of your time, for your invaluable advice and for always being there to encourage me when I needed it
the most. You thought me what being a professor means and to love this profession deeply.
An immense thank you also goes to my co-chair, Kristin. You are one of the most passionate
and dedicated scholars I know and an exceptional role model. I have learned a lot from you. Thank
you for all the time and the all-rounded advice you gave me. You always pushed me to become the
best scholar I can be and I will always be grateful for it.
Thanks to the rest of my committee, Norbert, Steph, and Dina. I have learned a lot from all of
you. Norbert, thank you for opening the doors of your lab to me, for your always being encouraging
and for teaching me to put things into the right perspective. Steph, your door was always open and
your feedback has been extremely valuable. Dina, thank you for pushing me to always look at things
in a different perspective.
Thank you to my co-author Andrea, who was the first to make me realize that a career in
academia was the right choice for me.
Thank you to all my friends in the PhD program. I can’t wait to see where we all end up in 5,
10, 50 years.
Thank you to my husband, Marco. I know it has not always been easy to be at my side in the
past 5 years. You took a leap of faith, followed me in this crazy adventure and always supported me
unconditionally. I love you.
Thank you to my parents, and my brother Riccardo. You were far away but I always felt your
love and support. Knowing that you believed in me gave me the strength I needed in times of difficulty.
ii
Table of Content
Acknowledgments ................................................................................................................................. i
List of Tables ...................................................................................................................................... vii
List of Figures .................................................................................................................................... viii
Abstract ................................................................................................................................................ ix
Chapter 1: INTRODUCTION AND OUTLINE ............................................................................... 1
1.1. Strategic Audience Partitioning ...................................................................................................... 1
1.2. Overview ........................................................................................................................................ 2
Chapter 2: THE SIDELINE EFFECT – HOW PARTITIONING AN AUDIENCE
FACILITATES SELF-PRESENTATION OBJECTIVES .............................................................. 5
2.1. Chapter Introduction ....................................................................................................................... 5
2.2. Chapter Overview ........................................................................................................................... 8
2.3. Literature Review ............................................................................................................................ 9
2.3.1. Impression Management and Self-Promoter Dilemma ................................................... 9
2.3.2. Self-presentation and the Audience .......................................................................... 12
2.4. The Current Investigation ........................................................................................... 13
2.5. Empirical Analysis – Pre-test .................................................................................................. 14
2.5.1. Overview ....................................................................................................................... 14
2.5.2. Method ........................................................................................................................... 15
2.5.3. Results .......................................................................................................................... 16
2.5.4. Discussion .................................................................................................................... 17
2.6. Empirical Analysis – Part I ..................................................................................................... 17
2.7. Study 1A ................................................................................................................................. 18
2.7.1. Overview .................................................................................................................. 18
iii
2.7.2. Method ..................................................................................................................... 18
2.7.3. Results ...................................................................................................................... 19
2.8. Study 1B ................................................................................................................................. 19
2.8.1. Overview .................................................................................................................. 19
2.8.2. Method ..................................................................................................................... 19
2.8.3. Results ...................................................................................................................... 20
2.9. Study 1C ................................................................................................................................. 20
2.9.1. Overview .................................................................................................................. 20
2.9.2. Method ..................................................................................................................... 20
2.9.3. Results ...................................................................................................................... 21
2.10. Discussion of Studies 1A, 1B and 1C ................................................................................... 22
2.11. Study 2A ............................................................................................................................... 23
2.11.1. Overview ................................................................................................................ 23
2.11.2. Method ................................................................................................................... 23
2.11.3. Results .................................................................................................................... 24
2.12. Study 2B ............................................................................................................................... 24
2.12.1. Overview ................................................................................................................ 24
2.12.2. Method ................................................................................................................... 25
2.12.3. Results .................................................................................................................... 25
2.13. Discussion of Studies 2A and 2B.............................................................................. 26
2.14. Study 3 .................................................................................................................................. 26
2.14.1. Overview ................................................................................................................ 26
2.14.2. Determining Self-Enhancing Tweets ..................................................................... 26
iv
2.14.3 Method .................................................................................................................... 27
2.14.4. Results .................................................................................................................... 28
2.14.5. Discussion of Study 3 ............................................................................................ 28
2.15. Empirical Analysis – Part II .................................................................................................. 29
2.16. Study 4 .................................................................................................................................. 30
2.16.1. Overview ................................................................................................................ 30
2.16.2. Method ................................................................................................................... 30
2.16.3. Results .................................................................................................................... 31
2.16.4. Discussion .............................................................................................................. 31
2.17. Study 5 .................................................................................................................................. 32
2.17.1. Overview ................................................................................................................ 32
2.17.2. Method ................................................................................................................... 32
2.17.3. Results .................................................................................................................... 33
2.17.4. Discussion .............................................................................................................. 35
2.18. Study 6 .................................................................................................................................. 36
2.18.1. Overview ................................................................................................................ 36
2.18.2. Method ................................................................................................................... 36
2.18.3. Results .................................................................................................................... 37
2.18.4. Discussion .............................................................................................................. 39
2.19. Empirical Analysis – Part III ................................................................................................ 40
2.20. Study 7 .................................................................................................................................. 40
2.20.1. Overview .................................................................................................................... 40
2.20.2. Method ................................................................................................................... 41
v
2.20.3. Results .................................................................................................................... 42
2.20.4. Discussion .............................................................................................................. 42
2.21. Study 8 .................................................................................................................................. 43
2.21.1. Overview ................................................................................................................ 43
2.21.2. Method ................................................................................................................... 44
2.21.3. Results .................................................................................................................... 45
2.21.4. Discussion .............................................................................................................. 46
2.22. Chapter Discussion ............................................................................................................... 46
Chapter 3: PERSUADING THE BYSTANDER ............................................................................ 51
3.1. Chapter Introduction ..................................................................................................................... 51
3.2. Chapter Overview ......................................................................................................................... 53
3.3. Literature Review .......................................................................................................................... 53
3.3.1. Persuasion, Persuasion Knowledge and Resistance to Persuasion ................................ 53
3.3.2. Targeted Communication and Persuasion ............................................................... 55
3.4. The Current Investigation ....................................................................................................... 56
3.5. Study 1 .................................................................................................................................... 58
3.5.1. Overview .................................................................................................................. 58
3.5.2. Method ..................................................................................................................... 59
3.5.3. Results ...................................................................................................................... 60
3.5.4. Discussion ................................................................................................................ 61
3.6. Study 2 .................................................................................................................................... 62
3.6.1. Overview .................................................................................................................. 62
3.6.2. Method ..................................................................................................................... 62
3.6.3. Results ...................................................................................................................... 63
vi
3.6.4. Discussion ................................................................................................................ 64
3.7 Chapter Discussion .................................................................................................................. 64
Chapter 4: CONCLUSIONS ............................................................................................................. 67
4.1. Chapter Introduction ..................................................................................................................... 67
4.2. Main Findings and Implications ................................................................................................... 67
4.3. Suggestions for Future Research .................................................................................................. 69
References ........................................................................................................................................... 72
Appendices .......................................................................................................................................... 83
APPENDIX A—Pre-test S1 and S2 (Chapter 2) .............................................................................. 83
APPENDIX B—Stimuli S1B and S5 (Chapter 2) ......................................................................... 84
APPENDIX C—Facebook Posts Used S2 (Chapter 2) ................................................................. 85
APPENDIX D—IQ Test S4 (Chapter 2) ....................................................................................... 86
APPENDIX E—Twitter Posts Used S6 (Chapter 2) ..................................................................... 87
APPENDIX F—Additional Results S6 (Chapter 2) ...................................................................... 88
APPENDIX G—Stimuli S8 (Chapter 2)........................................................................................ 90
APPENDIX H—Stimuli S2 (Chapter 3)........................................................................................ 91
vii
LIST OF TABLES
TABLE 2-1. Strategies Used to Solve the Self-Promoter Dilemma ..............................................16
TABLE 2-2. Effect of Content on Likelihood to Tag in Twitter Post ...........................................28
viii
LIST OF FIGURES
FIGURE 2-1. Conceptual Model Tested .........................................................................................8
FIGURE 2-2. Choice Study 1C .....................................................................................................22
FIGURE 2-3. Model Tested in Study 6 .........................................................................................37
FIGURE 3-1. Conceptual Model Tested .......................................................................................58
ix
Abstract
Interpersonal communication serves a variety of functions for individuals in everyday life
including helping to maintain the desired image in the eyes of others, seeking to persuade others
to see one’s point of view, or attempting to convince them to take a desired action.
Communicators can use a variety of strategies to increase the effectiveness of their
communication. In my dissertation, across 2 essays, I introduce one such strategy, audience
partitioning, and investigate its antecedents and consequences. I define strategic audience
partitioning as the decision by the sender of a one-to-many communication to change the
participation structure of his or her audience by dividing recipients into two distinct groups:
addressed and non-addressed recipients. I focus on both the anticipated and actual effects that
audience partitioning has on non-addressed recipients in an audience, whom I refer to as
bystanders. In essay 1, I focus on the use of audience partitioning when the goal of the
communicator is to effectively manage one’s image in the eyes of others (i.e. impression
management), while in essay 2, I investigate the use of audience partitioning in the context of
persuasive communication.
1
CHAPTER 1: INTRODUCTION AND OUTLINE
1.1. Strategic Audience Partitioning
Consumers are social beings (Fiske 2009), and therefore dedicate a lot of time to
communicating with one another. Interpersonal communication serves a variety of functions for
individuals in everyday life including helping to maintain the desired image in the eyes of others
(Goffman 1981), seeking to persuade others to see one’s point of view, or attempting to convince
them to take a desired action (Bochner 1984). The extent to which a particular message is
effective in achieving the goals of the communicator is a function of the rhetorical tools used by
the speaker, as well as his or her ability to tailor the message to a specific target audience (Bell
1984; Tice et al 1995).
Consumers engage both in one-to-one communication–in which case their target
audience consists of a single individual–and in one-to-many communication, also referred to as
broadcasting (Barasch and Berger 2014) –in which their target audience is more than one
individual. This happens, for example, anytime an individual shares a story with an entire group
of friends around the dinner table. Notably, one-to-many communication is especially prevalent
in the context of social media, where users can share content seen by a large number of friends
(as in the case of Facebook) and followers (as in the case of Twitter). One-to-many
communication presents distinct communicative challenges (often including the impossibility of
tailoring the message to each recipient), but also opens the door for the communicator to use new
communication strategies to achieve his or her goals.
My work introduces one such communication strategy, audience partitioning. I define
strategic audience partitioning as the decision by the sender of a one-to-many communication to
change the participation structure of his or her audience by dividing recipients into two distinct
2
groups: addressed and non-addressed recipients. Audience partitioning can be implemented
online, for instance, by tagging specific individuals in public social media posts. In the offline
world, audience partitioning can be accomplished simply by addressing a specific audience
member (or members) during group interactions. For instance, recall our dinner example.
Imagine an individual sitting at the dinner table hearing a fellow diner address another person
sitting next to him. Would the individual’s opinion of the speaker and attitude towards the
message be affected by the mere fact the message was addressed to someone else? Critically, I
argue that speakers can choose whether to partition their audience, and do so strategically by
including the real intended recipients of the message but not addressed them directly. These
recipients still receive the message, but do so as non-addressed members of a partitioned
audience.
In my dissertation, across 2 essays, I document the phenomenon of strategic audience
partitioning and investigate both the anticipated and actual effect that audience partitioning has
on non-addressed recipients in an audience, whom I refer to as bystanders. In essay 1, I focus on
the use of audience partitioning when the goal of the communicator is to effectively manage
one’s image in the eyes of others (i.e. impression management), while in essay 2, I investigate
the use of audience partitioning in the context of persuasive communication.
1.2. Overview
Essay 1 (chapter 2), The Sideline Effect: How Partitioning an Audience Facilitates Self-
Presentation Objectives, asks whether and, if so, why consumers use audience partitioning and
whether audience partitioning can promote WOM communication. A priori, I predict audience
partitioning is a distinct self-presentation strategy that consumers use when broadcasting self-
enhancing content. They do so in an attempt to solve what is known as the self-promoter
dilemma (i.e., sharing self-enhancing content without coming across as overtly attention-
3
seeking). Across eleven studies, including the analysis of secondary data, I show consumers do
in fact use audience partitioning strategically. First, I show how individuals are more likely to
use audience partitioning when sharing self-enhancing (vs. not self-enhancing) content, which is
when they would face the self-promoter dilemma. Further, I document why this happens. It
occurs because, when self-promoters partition their audience, they expect non-addressed
recipients to be less prone to view their message as attention-seeking (compared to members of a
non-partitioned audience). By expecting to come off as less attention-seeking, people generally
believe non-addressed recipients will form a more favorable impression of them.
Importantly for marketing, I show the use of audience partitioning promotes WOM in
contexts where the self-promoter dilemma might otherwise inhibit consumers from talking about
their purchases and consumption experiences. Further, I investigate the efficacy of this strategy. I
do so by examining whether those who have been placed in the role of non-addressed recipients
through audience partitioning indeed form better impressions of the self-promoter. This is an
important question to ask, because individuals are often miscalibrated with respect to how their
self-promotional efforts are received by others (Scopelliti, Vosgerau, and Loewenstein 2015).
I also examine the effects of audience partitioning from the recipient’s side. I find that
including an explicit addressee when sharing a self-enhancing message generally, but not always,
improves the impressions non-addressed recipients form. Whether audience partitioning is a
successful self-presentation strategy depends on the inferences non-addressed recipients make
regarding the reason why one or more addressee is being addressed. If non-addressed recipients
believe there is a reason for others to be addressed (e.g., the self-promoter and the addressees
share common ground), they consequently attribute the message to this shared connection rather
than to a desire to self-promote and form more positive impressions. Without such a reason,
4
audience partitioning does not have any effect on impression formation.
In Essay 2 (chapter 3), Persuading the Bystander, I examine audience partitioning in the
context of persuasive communication, such as personal selling or soliciting support for a cause. I
ask whether individuals who are exposed to persuasive communication addressed at someone
else are more persuaded than if they had been addressed directly. Past work on persuasion
knowledge tells us that people’s knowledge about persuasion agents’ goals, intentions and tactics
affects the efficacy of a persuasion attempt. I show that such knowledge is more likely to be
activated when consumers are addressed directly, but less likely if they are not addressed
directly. In this work, I find, across a range of persuasion contexts, that non-addressed members
of a partitioned audience end up being more susceptible to persuasion than members of a non-
partitioned audience.
Taken together, these two essays introduce the phenomenon of strategic audience
partitioning, document its prevalence and start uncovering its antecedents and consequences.
5
CHAPTER 2: THE SIDELINE EFFECT – HOW PARTITIONING AN AUDIENCE
FACILITATES SELF-PRESENTATION OBJECTIVES
2.1. Chapter Introduction
Consumers commonly use products and brands in an attempt to manage the impressions
they make on others. Sometimes consumers passively let their products and brands do the talking
by, for example, displaying conspicuous brands (Erdem and Swait 1998; Han, Nunes, and Drèze
2010), while other times they actively talk about products and brands. One of the critical drivers
of word of mouth communication (WOM) is indeed a desire to self-enhance, that is, share
content aimed at making the individual look good in the eyes of others (Chung and Darke 2006;
De Angelis et al 2012; Lampel and Bhalla 2007). This is especially true online and in particular
on social media (Schau and Gilly 2003; Lovett, Peres, and Shachar 2013). Sharing self-
enhancing content is an instantiation of what research in social psychology calls self-promotion
(Jones and Pittman 1982), an impression management strategy intended to elevate other peoples’
opinions of the person.
The self-promoter, however, faces a dilemma: how does he or she share self-enhancing
content without coming across as overtly attention-seeking? If an individual sharing self-
enhancing content is perceived as overtly attention-seeking, he or she runs the risk of making a
negative impression, appearing arrogant and boastful (Berman et al. 2015; Reeder 2009;
Schlenker and Leary 1982; Sekhon et al. 2014). Consider how the Twitter community mocks
blatantly self-serving status updates on the website repository Twouchebags
(twitter.com/twouchebags). Critically, for marketers interested in leveraging WOM, consumers
who fear negative repercussions may hesitate to share information about products, services and
6
brands, particularly when what they mention is expensive and hence may come across as
boastful (Berger 2014).
In this research, I identify a distinct self-promotion strategy consumers employ in an
attempt to solve the self-promoter dilemma, referred to as audience partitioning. I define as the
decision by the senders of one-to-many communications to change the participation structure of
their audience by dividing recipients into addressed and non-addressed recipients. Audience
partitioning can therefore be employed when sharing content with more than one recipient and
involves altering the participation structure of the audience (Goffman 1959, 1981; McGregor
1986).
To illustrate how this process works, consider Facebook. Someone can post a self-
enhancing message-such as Wow, this is the third year in a row that my work has received notice
in the Spider Awards!-on his or her wall without addressing any designated recipients.
Alternatively, the individual can choose to address the same message to one or more recipients
by tagging them. Operationally, tagging occurs when a person posts a message using the @
symbol followed by the name of a specific individual or individuals- such as @John, wow, this is
the third year in a row that my work has received notice in the Spider Awards!. Similarly,
offline, an individual chatting with a group of people can chose to tell a self-enhancing anecdote
to the whole group or address a specific person, knowing everyone else in the group will hear. It
is my assertion that non-addressed recipients are often the true target of self-enhancing messages.
They are relegated to the role of bystanders in an attempt to suppress potential negative
responses to overt self-promotion. The self-promoter’s expectation is that, if non-addressed
members of a partitioned audience believe a message was not directed at them, they will perceive
7
the message as less attention-seeking and ultimately form a more positive impression of the
person in light of the new information.
My work is the first to investigate how individuals seeking to self-promote use audience
partitioning strategically. While past research has proposed solutions to the self-promoter
dilemma based on what self-promoters share (message) and with whom (audience) (see Pfeffer,
Fong, and Cialdini 2006; Sezer, Gino, and Norton 2018; Sekhon et al. 2014), in this research,
unlike past work, both the message and the composition of the audience stay exactly the same.
What changes is the perceived role of different audience members. This work provides a novel
contribution to the literature on self-presentation and impression management by documenting
an un-researched self-presentation strategy used in one-to-many self-promotion, both online and
offline. Moreover, by showing how audience partitioning can influence whether consumer share
information publicly about purchases and consumption experiences, this work also contributes to
the literature on word-of-mouth communication. Finally, I contribute to the literature on social
cognition and communication, which has provided a wide variety of audience taxonomies but
has yet to investigate whether, when and how people manage the participation structure of an
audience empirically.
8
FIGURE 2-1. Conceptual Model Tested
2.2. Chapter Overview
The remainder of this chapter proceeds as follows. I begin by reviewing the literature on
impression management and self-presentation strategies. I then summarize the work in sociology
and communication that has looked at the participation structure of an audience. My
conceptualization links these two literature streams together to explain when and why consumers
would expect strategic audience partitioning to facilitate their own impression management
goals. I present eleven studies, including experiments and the analysis of secondary data that,
taken together, provide substantial evidence in support of my conceptualization (see figure 1).
For ease of exposition, I divide my empirical package into two parts.
In Part I, I document the basic phenomenon by showing how consumers are more likely
to use audience partitioning when the content to be shared is considered self-enhancing.
Accordingly, in the initial set of studies the independent variable is the nature of the content
9
being shared and the dependent variable is the choice between partitioning one’s audience or not.
Finding that audience partitioning is more likely when the content shares is self-enhancing is
consistent with the notion that audience partitioning is used strategically.
In Part II, I examine the reasoning underlying the strategic use of audience partitioning,
demonstrating how its use is expected to solve the self-promoter dilemma. In these studies, I
show how self-promoters who have partitioned an audience expect to be perceived as less
attention-seeking, and as a result expect to make a better impression on non-addressed audience
members.
While Part II focuses on the expected consequences of audience partitioning, in Part III, I
examine the actual consequences of audience partitioning in terms of impressions. In other
words, I ask whether partitioning is indeed a successful self-presentation strategy and whether
others form different opinions of a self-promoter as a function of whether they are members of a
non-partitioned audience as opposed to non-addressed audience members in a partitioned
audience.
I conclude the chapter with a discussion of the implications of my findings for both
researchers and practitioners, as well as future avenues of research.
2.3. Literature Review
2.3.1. Impression Management and Self-Promoter Dilemma
Everyone, at least at times, is concerned with self-presentation and attempts to manage
other people’s impressions of themselves. Self-presentation is an inescapable feature of life
because so much time is spent in the presence of others whom we believe to be constantly
evaluating us (Jones and Pittman 1982). Attempts to convey a positive image are so pervasive
they often become habitual and even automatic (Brown 2007). Self-presentation efforts have
been categorized broadly as falling into one of two types: protective or acquisitive (Arkin 1981).
10
Protective self-presentation refers to individuals trying to avoid associating the self with content
that might make a negative impression on others while acquisitive self-presentation refers to
actively associating the self with content expected to make a positive impression on others. The
focus of this research is acquisitive self-presentation.
A common acquisitive self-presentation strategy is to share self-enhancing information.
The desire to be seen by others as socially worthy is one of the most central human motivations
and results in the tendency to self-enhance by highlighting one’s strengths, positive traits,
accomplishments and possessions (Fiske 2001; Schwartz 1992). In the context of consumption
specifically, the desire to make a good impression on others has been shown to be a critical
driver of what consumers purchase as well as which purchases and consumption experiences
they choose to discuss with others (Chung and Darke 2006; De Angelis et al 2012; Lampel and
Bhalla 2007). It is critical to point out that the act of sharing self-enhancing information has been
labeled alternately as bragging (Berman et al. 2015), self-promotion (Jones and Pittman 1982;
Rudman 1998; Scopelliti, Loewenstein, and Vosgerau 2015), boasting (Levine and West 1976),
self-praise (Dayter 2014), positive self-disclosure (Miller et al. 1992), and positive self-
description (Holtgraves and Srull 1989). As evidenced by some of these descriptors, sharing self-
enhancing information is not always looked upon favorably, leading individuals to attempt to
manage self-promotion efforts strategically.
The extent to which an individual’s self-promotional efforts are successful depends on
one’s ability to mask the intention of generating positive impressions (Eastman 1994; Giacalone
and Rosenfeld 1986; Jones and Pittman 1982). When seen as intentionally attention-seeking, the
person usually misses the mark in terms of making a good impression (Berman et al. 2015,
Scopelliti, Vosgerau, and Loewenstein 2015). This is because people generally value modesty,
11
and a person who is seen as deliberately self-enhancing runs the risk of appearing overtly
attention-seeking and thus arrogant and/or boastful (Godfrey, Jones, and Lord 1986; Powers and
Zuroff 1988; Schlenker and Leary 1982; Tal-Or 2010; Tice et al. 1995).
People are generally aware of the negative consequences that can accompany being
perceived as actively self-promoting. The desire to self-enhance while not being harmed by the
fact that self-promoters are viewed negatively is what has been dubbed the self-promotion
dilemma (Pfeffer et al. 2006), or the braggart’s dilemma (Berget et al. 2015). A number of
strategies used by consumers for delivering self-enhancing content with less conceit have been
documented in the literature. One strategy, for instance, is that of changing who delivers the
message by having a third party sing one’s praises (Pfeffer et al. 2006; Scopelliti, Vosgerau and
Loewenstein 2018). Alternatively, one could brag indirectly by praising others closely associated
with the self, a behavior commonly known as basking in reflected glory (Cialdini et al. 1976).
Other strategies involve altering the message, which one can do by explicitly attributing one’s
success to external factors such as luck (Hareli and Weiner 2000), or balancing positive with
negative information, as in the case of humblebragging (Sezer, Gino, and Norton 2018). Since
positive self-statements in response to a specific question are better received than unprompted
self-presentation (Holtgraves and Srull 1989), an alternative strategy would be trying to get
others to ask questions about things the individual wants to self-promote about. Finally, a
somewhat related strategy involves tailoring the self-enhancing message in light of an audience’s
expectations and characteristics (Tice et al. 1995), what is referred to as audience design (Bell
1984). Again, all of these strategies focus on who shares what with whom, while in my research,
all of these aspects stay exactly the same. What changes with audience partitioning is the role
assigned to different audience members.
12
2.3.2. Self-Presentation and the Audience
By definition, all forms of self-presentation require an audience (Goffman 1959; Hogan
2010). Indeed, the mere presence of an audience influences behavior by making individuals
conscious of how others perceive and therefore might evaluate them (Argo, Dahl, and
Manchanda 2005; Latané and Nida 1981; Petty et al. 1977; Zajonc 1965). One fundamental
characteristic of an audience is its size. On some occasions, communication (as a form of self-
presentation) involves transferring information to a single recipient, referred to as narrowcasting,
while on other occasions, an individual transfers information to more than one recipient, referred
to as broadcasting (Barasch and Berger 2014).
In an era of expansive digital networks, individuals frequently find themselves in the
position of broadcasting to audiences comprised of diverse constituencies (Marwick and boyd
2011; Nadkarni and Hofmann 2012; Schau and Gilly 2003). When broadcasting, an audience
comprised of multiple diverse constituencies can complicate self-promotional efforts,
particularly when each might have different information and expectations with respect to the
self-promoter. While we know self-presentational efforts are often adapted to meet the
expectations of different audience members to be effective (Bell 1984; Goffman 1959),
contemporary forms of one-to-many communication (e.g., social media) are making audience
design—tailoring one’s message to a specific audience—more difficult, if not altogether
impossible (Marder et al. 2016). Simultaneously, many of the forms of contemporary one-to-
many communication (e.g., social media) make altering the participation structure of one’s
audience (e.g., through tagging) very easy.
Past research in social cognition and communication discusses various audience
participation structures and the associated roles individuals might play. Goffman (1959) notes
13
that an important distinction in roles is between addressed and non-addressed recipients of a
particular message. The role of an addressed recipient differs from that of a non-addressed
recipient who is on the sideline. While the former is expected to play an active role in the
conversation (for instance, by responding to the remarks made), the latter’s role (as a result of the
designation of one or more addressees) is to listen, yet feel somewhat inhibited with respect to
being actively involved (e.g., responding). Importantly, non-addressed recipients are still
members of the audience and thus privy to the message. In fact, the speaker is fully aware that
most non-addressed recipients are listening to what is being said
1
.
2.4. The Current Investigation
To date, what has not been examined by the literature are deliberate attempts to change
the structure of one’s audience thereby actively assigning different recipients to different roles.
The central premise of my research is that self-promoters will actively manage the participation
structure of their audience when sharing self-enhancing content in one-to-many communication.
They do so by means of audience partitioning, that is, by dividing an audience into addressed
recipients (addressees) and non-addressed recipients (bystanders).
I further propose self-promoters regularly want bystanders’ attention yet do not want this
desire to be transparent. The self-promoter expects non-addressed members of a partitioned
audience—those relegated to the sidelines—to believe the communication was not really
intended for them, and consequently to view the message as less attention-seeking. In what
follows, I set out to test that: i) audience partitioning is disproportionally used when consumers
share self-enhancing content, ii) this happens because self-promoters expect non-addressed
recipients assigned to the role of bystander to perceive them as less attention-seeking, resulting
1
With the exception of eavesdroppers who listen in surreptitiously (Clark and Carlson 1982; Clark and Schaefer 1987).
14
in a more favorable impression, and iii) audience partitioning is a successful self-presentation
strategy since bystanders do indeed form a more positive impression of the self-promoter by
perceiving him/her as less attention-seeking.
For the proposed sideline effect to occur only a few conditions are necessary. An
individual must communicate with an audience greater than one and do so in a context in which
addressing an individual directly is both possible and appropriate. At the same time, the other
audience members must find it possible and easy to attend to the message. These conditions are
easily met both online and offline. Imagine one diner addressing another diner at a table while
delivering a self-enhancing message, fully realizing that s/he is easily heard by others at the table
(i.e., bystanders). In a parallel fashion, an individual can choose to post a message—on any one
of a number of social media platforms such as Facebook, Twitter and Instagram—visible to
anyone in his or her network but address it to one or more recipients by way of a tag.
2.5. Empirical Analysis – Pre-test
2.5.1. Overview
The current study was designed to provide initial evidence that audience partitioning is
seen as a solution to the self-promoter dilemma and is indeed used as a self-presentation strategy
to solve such dilemma.
I asked respondents whether they have ever felt the self-promoter dilemma and, in case of
a positive answer to this question, how they chose to solve such dilemma. Respondents were
presented with a set of self-promotional strategies that previous literature has highlighted as
being used to solve the self-promoter dilemma. Among these, was included a description of
audience partitioning. Finding that audience partitioning is indeed used as a solution to the self-
promoter dilemma, and that this usage compares to that of other well established self-
15
presentation strategies would provide an idea of the importance of the phenomenon under
consideration.
2.5.2. Method
One hundred and fifty-one college students (48.0% female, Mage = 20.0) completed the
survey for partial course credit. First, respondents read the following paragraph about the self-
promoter dilemma: Often times, people want to share positive information about
themselves such as getting good grades, buying a new car, going on an amazing vacation, or
winning some award or competition. However, they often worry about coming across
as braggy and attention-seeking. This conflict is known as the self-promoter dilemma because
the individual is conflicted about whether and how to share information that reflects positively
on them without incurring these negative consequences.
After reading the paragraph, respondents were asked if the ever experienced the self-
promoter dilemma with the following question: Can you think of a time when you experienced
the self-promoter dilemma? (yes, no). Those who reported remembering facing the dilemma
were asked to write a short paragraph describing the instance they recalled.
Next, they were presented with a list of 6 strategies used to solve the self-promoter
dilemma listed in a random order and were asked to select all the ones they remembered using
themselves. Five of these strategies were selected based on previous literature and included:
using humility, using a complaint, using questions, using others to disseminate, and using
indirect brags (see complete description in Appendix A). The first two items are intended to
assess the use of humblebrags - humility and complaints are the two main forms that
humblebrags can take (Sezer, Gino, and Norton 2018). The third item refers to individual’s
attempts to get others to ask them questions about things they want to self-promote about, since
16
positive self-statements in response to a specific question are better received than unprompted
self-presentation (Holtgraves and Srull 1989). The forth item refers to the behavior that
Scopelliti, Vosgerau and Loewenstein (2018) refer to as bragging through an intermediary,
asking others to share positive information about the self (see also Pfeiffer et al. 2006). Finally,
the fifth item refers to the behavior commonly known as basking in reflected glory (Cialdini et
al. 1976), that is bragging about someone else the individual is associated with, so that it
indirectly reflects well on the individual. The sixth strategy was the new strategy presented in my
work, audience partitioning, described to respondents as using overhearing - Sharing the positive
information with one person while intentionally having others overhear the conversation.
2.5.3. Results
I find that 137 respondents (90.73%) reported remembering experiencing the self-
promoter dilemma. Of these 44 (32.12%) reported having used audience partitioning in an
attempt to solve the dilemma. For a representation of how the use of this strategy compares with
other commonly used strategies described in the literature, see Table 1.
TABLE 2-1. Strategies Used to Solve the Self-Promoter Dilemma.
Strategy
Number of
respondents who used
the strategy
% of those who felt the
self-promoter dilemma
Use humility 113 82.48%
Use questions 83 60.58%
Use overhearing 44 32.12%
Use complaints 44 32.12%
Use indirect brags 39 28.47%
Use others to disseminate 22 16.06%
17
2.5.4. Discussion
These findings suggest that experiencing the self-promoter dilemma is indeed a common
occurrence for individuals. Most importantly, a third of those who feel the dilemma report
having used audience partitioning in an attempt to solve the dilemma. The use of this strategy is
as common as complaint-based humblebrags and more common than other important strategies
previously discussed in the literature. This provides initial evidence of the frequency of audience
partitioning as a solution to the self-promoter dilemma.
2.6. Empirical Analysis – Part I
In the first part of my empirical analysis I set out to document how the nature of the
content that an individual intends to share—self-enhancing or not—affects the likelihood of
partitioning the audience. In study 1A, I vary the nature of the content respondents imagine
sharing with dinner companions, while in study 1B, I ask respondents to generate content that
was either explicitly self-enhancing or simply self-relevant. In study 1C I replicate the findings
of study 1B while giving respondents the opportunity to avoid sharing the content they self-
generated. In study 2A, I test the effect of the nature of the content to be shared on audience
partitioning in a different domain, social media. Relatedly, in study 2B I examine how an
individual difference—public self-consciousness—impacts the propensity to partition one’s
audience when sharing self-enhancing content. I move outside the lab in study 3, analyzing
Twitter data to explore how the nature of the content affects audience partitioning in the real
world.
18
2.7. Study 1A
2.7.1. Overview
In study 1A, I document in a controlled laboratory experiment how the nature of the
content consumers intend to share impacts the likelihood of partitioning their audience. I ask
respondents to choose how they wish to share information about a recent experiential purchase
(either by addressing a single member of the audience or by addressing no one in particular). I
vary the extent to which the content is self-enhancing while holding constant the type of
purchase (a prior dining experience) and the context in which respondents are sharing the
content. In line with my conceptualizing, I expect respondents to be significantly more likely to
partition their audience when sharing self-enhancing content.
2.7.2. Method
I recruited participants through Amazon Mechanical Turk (mTurk) and restricted eligible
respondents to U.S. residents with a 95% or higher approval rate and a completion rate of at least
50 hits. Participants were compensated 0.50 USD. I allowed only one response per IP address
(Goodman, Cryder, and Cheema 2013). I used the same criteria to select participants in all
studies employing mTurk workers, unless otherwise specified. A total of 410 participants (49.5%
female, M
age
= 36.7) completed the survey.
Respondents read a short scenario about sitting at a table with a group of people they
know. The group included their friend Alex, the potential addressee. The scenario asked
respondents to imagine they had decided to share a prior restaurant dining experience. I
manipulated the nature of the Content (Self-Enhancement vs. Control) by telling respondents
they either wanted to say: Went to Spago the other day. Not too shabby for a 200-dollar dinner!
or Went to Shakey’s the other day. Not too shabby for a 20-dollar dinner! Respondent reported
19
how they preferred to share this content by indicating whether they preferred to: Tell Alex, who
is sitting next to you, knowing that everyone at the table will be able to hear you, (i.e., using
audience partitioning) or Tell everyone at the table simultaneously.
2.7.3. Results
In the Self-Enhancement condition, 52.4% of respondents chose to partition their
audience by addressing Alex directly. In line with expectations, this was significantly greater
than the 39.6% of participants who chose to address Alex directly in the Control condition (χ2 =
6.76, p < .01).
2.8. Study 1B
2.8.1. Overview
In study 1B, I set out to test the effect observed in 1A when respondents generate the
content to be shared themselves. Respondents are asked to write something either self-enhancing
or simply self-relevant and then are asked to choose how they would rather share it. Again, ex
ante, I expect respondents to be more likely to partition their audience (i.e., address a single
member of the audience) when the content they generate is self-enhancing despite the fact that
the entire audience would be exposed to whatever they say irrespective of their choice.
2.8.2. Method
One hundred and eighty-eight college students (61.0% female, M
age
= 20.0) completed
the study for partial course credit. First, respondents were asked to recall and summarize a self-
relevant fact/story. They did not know at the time of writing that subsequently they would choose
how to share this fact/story with others. I manipulated the nature of the Content (Self-
Enhancement vs. Control) in the following way. In one condition, respondents were instructed to
write something about themselves others would find impressive (self-enhancing condition). In
20
the control condition, respondents were instructed to write about something interesting about
themselves (see stimuli in appendix B).
Next, respondents were asked to imagine they were having dinner with a group of
friends, and, just as in study 1A, were asked how they would prefer to share the content they just
wrote about. Finally, as a manipulation check, respondents were instructed to think back to the
story/fact they wrote about and answer: To what extent was it self-enhancing (a fact/story that
would make you look good in the eyes of others)? (1 = not at all, 9 = a great deal).
2.8.3. Results
Manipulation Check. A between-subject ANOVA indicates respondents considered the
content they wrote about in the Self-Enhancement condition more self-enhancing (M
SelfEnhancement
= 6.64, SD = 1.71) compared to the Control condition (M
Control
= 4.59, SD = 2.19, F(1,186) =
51.18, p < .01, ωp² = .211).
Choice. In line with my expectations and the results of study 1A, significantly more
respondents chose to partition their audience (address Alex directly) in the Self-Enhancement
condition than in the Control condition (51.6% vs. 31.2%, respectively, χ2 = 8.05, p < .01).
2.9. Study 1C
2.9.1. Overview
Both in study 1A and in study 1B respondents were asked to choose how they’d rather
share self-enhancing content. Importantly, one could ask whether forcing respondents to make
such choice, rather than giving them the opportunity of choosing not to share such self-
enhancing content, had any influence on my previous results. In other words, perhaps
respondents chose audience partitioning because they saw it as the lesser of two evils but would
actually not use that strategy if given the opportunity to simply not share the self-enhancing
21
information. Notably, while such a criticism seems to contradict the results of the pre-test (where
respondents reported having chosen to use audience partitioning to solve the self-promoter
dilemma even if they could have simply avoided self-promoting), it is still importantly to address
it directly. In study 1C, respondents are therefore also given the opportunity to simply avoid
sharing any self-enhancing information altogether.
2.9.2. Method
Three hundred and seventy college students completed the study for partial course credit.
The procedure of this study mirrors those of study 1B, with one important exception. Instead of
being asked how they would share the content they wrote about, they were told they were
considering sharing such content and they were given three options to choose from: Tell Alex,
who is sitting next to you, knowing that everyone at the table will be able to hear you, Tell
everyone at the table simultaneously, or I would not tell this story.
2.9.3. Results
Manipulation Check. A between-subject ANOVA indicates respondents considered the
content they wrote about in the Self-Enhancement condition more self-enhancing (M
SelfEnhancement
= 6.59, SD = 1.53) compared to the Control condition (M
Control
= 4.50, SD = 2.03, F(1,368) =
126.16, p < .01, ωp² = .253).
Choice. Respondents’ choices are reported in Figure 2. Importantly, in line with my
expectations and the results of study 1A and study 1B, even when respondents are given the
opportunity of not sharing the content they generated, significantly more respondents chose to
partition their audience (address Alex directly) in the Self-Enhancement condition than in the
Control condition (35.1% vs. 25.4%, respectively, χ2 = 4.15, p = .04).
22
FIGURE 2-2. Choice Study 1C.
2.10. Discussion of Studies 1A, 1B and 1C
The results from study 1A, 1B, and 1C provide evidence that individuals are more likely
to partition their audience, turning most recipients into bystanders, when they intend to share
self-enhancing content. These results are consistent with audience partitioning as a behavior
intended to facilitate self-presentation objectives. I show this to be the case when the content
shared describes an experiential purchase (more vs. less expensive) as well as when respondents
generate their own content (self-enhancing vs. simply interesting). Importantly, demonstrating
the effect with content that is personally relevant to respondents, as I did on study 1B, lends
external validity to my findings. Moreover, these results hold even when respondents are given
the opportunity to avoid sharing the self-enhancing content altogether. In study 1C, respondents
were given the opportunity to not share anything with the table, yet only 23.2% of those who
generated self-enhancing content chose to do so. The vast majority still chose to share such
information, in line with previous literature suggesting that individuals want to share information
that reflects well on them. Almost half of those who shared self-enhancing content chose to do so
using audience partitioning. I argue this is because they see audience partitioning as a solution to
23
the self-promoter dilemma. As I will test more definitively in later studies, respondents who fear
appearing attention-seeking believe they will make a more favorable impression if they choose to
address a single audience member.
In study 2, I set to replicate my findings in a different domain with its own idiosyncratic
communication norms, that is to say, social media. Replicating the effect on social media
magnifies the importance of the phenomenon, especially given that on Facebook, Twitter, and
Instagram audience partitioning is especially easy to implement with the use of tags.
2.11. Study 2A
2.11.1. Overview
The primary objective of study 2A is to test the effect observed in face-to-face
interactions in study 1A and 1B in a social media context. Online as well as offline, I expect
consumers to be more likely to partition their audience when sharing self-enhancing content. I
again vary the extent to which the content is self-enhancing in nature across conditions.
Recall the content shared in study 1A involved an experiential purchase, while in 1B and
1C the content was self-generated. Self-enhancing content can refer to brands and purchases, but
it can also refer to personal skills and achievements (Scopelliti et al. 2015). Here, I test the effect
across different types of self-enhancing content individuals share when trying to impress others. I
vary the topic to reference either a car purchase (material possession), or the results of an IQ test
(achievement). These are simply included as replicates and, a priori, I do not expect topic to
matter.
2.11.2. Method
I recruited participants through Amazon Mechanical Turk (mTurk) and applied the same
selection criteria described in study 1A. Moreover, only respondents prescreened to be regular
24
Facebook users were allowed to complete the survey. Three hundred and twelve participants
(58.0% female, M
age
= 35.3) completed the survey.
Respondents read a short scenario either about taking an IQ test or driving a new car
(Topic: IQ vs. Car). I manipulated Content (Self-Enhancement vs. Control) by informing
respondents that they either performed exceptionally well or exceptionally poorly on an IQ test,
or that their car was either a new BMW Series 5 or an old Toyota Corolla. Next, I presented
respondents with a Facebook post about the topic (see appendix C) and they chose whether they
would post it on their Facebook wall as is or to address a good friend by tagging him/her.
2.11.3. Results
As expected, the Topic of the message (IQ vs. Car) did not affect any of my results.
Therefore, I collapsed the data across these conditions. In line with my prediction, significantly
more respondents chose to partition their audience by tagging a friend in the Self-Enhancement
condition than in the Control condition (29.6% vs. 18.8%, respectively, χ2 = 5.03, p = .02).
2.12. Study 2B
2.12.1. Overview
Thus far, I found that the nature of the content one intends to share predicts the use of
audience partitioning-more self-enhancing content that may come across as attention-seeking is
associated with a greater use of audience partitioning. In this study, I approach the phenomenon
in a different way. Instead of manipulating the content to be shared, I measure public self-
consciousness, an individual difference indicating the extent to which consumers are aware of
how attention-seeking they might appear. If consumers use audience partitioning to avoid
appearing attention-seeking, I would expect individuals higher in public self-consciousness
(Fenigstein, Scheier, and Buss 1975), those who are more aware of how they appear to others, to
25
be more sensitive to how others perceive them when sharing self-enhancing content. Thus, I
expect public self-consciousness to predict respondents’ likelihood of partitioning their audience
(choosing to tag a friend) for a self-enhancing social media post. Finding an effect of public self-
consciousness would lend additional support to my conceptualization.
2.12.2. Method
I recruited participants through Amazon Mechanical Turk (mTurk) and applied the same
selection criteria used in previous studies. Moreover, only respondents prescreened to be regular
Facebook users were allowed to complete the survey. Three hundred and eight participants
(60.4% female, M
age
= 36.4) completed the study.
Respondents read one of two short scenarios before deciding how they would prefer to
post the content on Facebook. I used the positive (self-enhancing) versions of both the car and
the IQ scenarios employed in study 2A. Respondents simply chose whether to address a friend
by tagging him/her or not address anyone. Public Self-Consciousness was measured using the
seven items from the public self-consciousness scale (α = .88) developed by Fenigstein et al.
(1975). A priori, I expected Public Self-Consciousness to be associated with a greater likelihood
of partitioning the audience (i.e., tagging a friend).
2.12.3. Results
The Topic of the message (Car vs. IQ) did not affect any of my results. Therefore, I
collapsed the data across topics. A binary logistic regression model predicting audience
partitioning with respondents’ individual level of Public Self-Consciousness reveals a positive
and significant coefficient (b = 0.38, Wald χ2 = 4.58, p = .03). As expected, this result indicates
the likelihood of audience partitioning (i.e., tagging someone in the Facebook post) increases
with higher levels of Public Self-Consciousness.
26
2.13. Discussion of Studies 2A and 2B
The results of study 2A and 2B provide further evidence in support of audience
partitioning as behavior that is driven by the self-enhancing nature of the content shared and do
so in the context of social media interactions. Moreover, study 2B reveals that individuals’
sensitivity to how the content they share can be perceived by others also matters. I find
individuals more concerned about self-presentation (i.e., higher in public self-consciousness) are
more inclined to partition their audience when sharing self-enhancing content. While all my
previous studies were scenario studies, in study 3 I set to test whether consumers are more likely
to partition their audience when sharing self-enhancing content in the real world using archival
data reflecting actual behavior on social media.
2.14. Study 3
2.14.1. Overview
In this study, I examine the real-world applicability of my previous findings. If people are
more likely to partition their audience when sharing self-enhancing content, I should find
evidence of this in the real world. To test whether this is the case, I analyze the content of
420,797 real tweets. If my theorizing is correct, the chance of someone being tagged in a tweet
should increase when the tweet contains self-enhancing content.
2.14.2. Determining Self-Enhancing Tweets
To determine whether the content of a tweet is self-enhancing, I relied on LIWC, a
linguistic processing software designed for automated text analysis (Pennebaker et al. 2015).
Based on Schwartz’s theory of basic values (1992), power and achievement are the two core
values underlying self-enhancement. I therefore identify self-enhancing posts using two standard
LIWC dictionaries, power and achievement.
27
To validate the use of power and achievement to identify self-enhancing posts, I asked
1,000 Twitter users recruited on mTurk to each write two tweets with different content. I
manipulated the nature of the Content (Self-Enhancement vs. Control) within-subjects as
follows. Respondents were first instructed to write a tweet about anything that comes to your
mind (e.g., you could write about what you are doing right now). Next, they were instructed to
write a self-enhancing tweet (a tweet written with the specific goal of impressing those who read
it). Each tweet was judged as self-enhancing or not based on whether it contained any power or
achievement-related words according to the standard LIWC dictionaries. A test on equality of
proportions reveals the percentage of posts coded as self-enhancing using this approach is
significantly greater when respondents were indeed asked to write a self-enhancing post (54.3%
vs. 40.1%, z = -6.36, p < .01). Notably, those writing about anything that was on their mind
could have still included self-enhancing content, which makes this is an especially conservative
test.
Next, I set out to test whether a relationship exists between the content of actual tweets
(classified as self-enhancing based on the LIWC dictionary) and the use of audience partitioning
by way of tags.
2.14.3. Method
I collected all original tweets posted in the Greater Los Angeles area for the day of
September 16, 2016. Original tweets exclude retweets and replies to others’ tweets. I also
excluded tweets by verified users, typically accounts of public interest that are generally
associated with companies or celebrities. My final sample consisted of 420,797 tweets written by
129,718 users. Of these tweets, 63,326 employed audience partitioning. In other words, 15.0% of
the tweets in my sample included a tag addressing one or more other Twitter users.
28
For each tweet, I collected two covariates that assess the size of the social network of the
Twitter users, namely their number of followers (M = 4,520.14, SD = 24,785.79) and the number
of other users they follow (M = 1,962.82, SD = 12,216.83).
2.14.4. Results
Some users provided more than one tweet during the period under investigation.
Therefore, I conducted a logistic regression with robust standard errors clustered at the user level
to account for the non-independent nature of my data. As expected, the presence of self-
enhancing content in a tweet significantly and positively predicts the presence of at least one tag
(OR = 1.09, z = 3.78, p < .01). Moreover, this result holds when including the covariates that I
expected could influence the decision of whether to tag someone in the tweet (OR = 1.06, z =
2.45, p = .01), as shown in table 1 (left part).
TABLE 2-2. Effect of Content on Likelihood to Tag in Twitter Post.
Logistic Regression w/clustered errors
Variable Odds Ratio S.E. z p
Self Enhancement 1.06 (.02) 2.45 .01
Followers (log) .94 (.02) -2.19 .03
Following (log) 1.22 (.02) 10.56 .00
Constant .65 (.00) -35.67 .00
N tweets = 420,797; Twitter users = 129,718.
2.14.5. Discussion of Study 3
The results from study 3 suggest Twitter users consider the nature of the content they
share in a tweet when making decisions regarding audience partitioning; the inclusion of self-
enhancing content increases the likelihood of a tag within a tweet. While correlational in nature,
29
these results support the proposition that individuals sharing self-enhancing content use audience
partitioning as a self-presentational tool and do so in a real-world context.
I now switch from demonstrating a phenomenon to investigating the underlying process
driving that phenomenon. The studies in part II of my empirical package document the self-
promotional motivation for audience partitioning.
2.15. Empirical Analysis – Part II
In part I, I demonstrated how the nature of the content consumers intend to share—self-
enhancing or not—affects the likelihood of partitioning an audience. Consistent across studies, I
observe the likelihood of partitioning an audience is greater when sharing self-enhancing
content. The primary goal of the next set of studies is to show how this effect is driven by self-
presentation concerns. In other words, in what follows, I show how audience partitioning is
employed strategically to solve the self-promoter dilemma.
Consequently, instead of offering respondents a choice regarding audience partitioning,
which previously served as my dependent variable, I now manipulate whether the audience
already has been partitioned or not. I do so to test the relationship between audience partitioning
and several downstream consequences. Specifically, I establish that self-promoters expect to be
perceived as less attention-seeking when an audience is partitioned and, in turn, expect to make a
better impression on non-addressed audience members, those relegated to the sideline as
bystanders. Moreover, I show the use of audience partitioning increases consumers’ likelihood of
sharing self-enhancing content, which ultimately promotes word-of-mouth communication.
30
2.16. Study 4
2.16.1. Overview
Study 4 tests the basic proposition that audience partitioning has a meaningful influence
on expected impressions. I posit that individuals sharing self-enhancing content expect to make a
better impression on non-addressed audience members when the audience is partitioned
compared to the impression they would make on them if the audience was not partitioned. In this
study, I use a consequential set up by having respondents post a real self-enhancing message on
their Facebook wall. What varies is whether they are instructed to tag someone or not (audience
partitioning). A priori, I predict that the act of tagging a friend on Facebook will affect how
respondents expect their social network is going to respond. Specifically, I predict respondents
who partition their audience will anticipate a more favorable response from friends cast in the
role of bystanders.
2.16.2. Method
I recruited 201 respondents (45.3% female, M
age
= 34.1) on the online platform Prolific
Academic. Respondents were prescreened to be native English speakers (58.7% were Great
Britain citizens, 36.8% were US citizens, 3.0% were Canadian, and the remaining 1.5% were
from Ireland, Australia and New Zealand). Each participant was compensated 1.20 USD. I
allowed only one response per IP address (Goodman et al. 2013). At the beginning of the study,
respondents were informed that they would be required to write a public Facebook post on their
wall to complete the survey which ensured respondents were Facebook users.
In the first part of the study, respondents answered three relatively easy logic questions
purportedly part of an IQ test (the correct response rate averaged 94.9%; see appendix D for
questions, results, and additional details). Regardless of their answers, all respondents were
31
informed that they performed better than 80% of those who completed the survey and, although
not diagnostic, this implied an IQ level of up to 160. Next, they were asked to log on to their
Facebook account and post the following message about these results: Aced an IQ test today! I
knew I was a genius J! I manipulated whether or not respondents were instructed to tag a
Facebook friend, albeit anyone of their choosing (Audience Partitioned: Yes vs. No). In order to
ensure compliance, respondents were asked to upload a screenshot of their post cropped to avoid
disclosing personal information. Notably, respondent’s drop out rate did not vary across
conditions.
Finally, respondents were asked to name a person they recently befriended on Facebook
(those in the Audience Partitioned condition were explicitly instructed to think of someone other
than the person tagged). While thinking about this person, respondents were asked to answer the
following questions: What is the overall impression that you believe this person will form about
you? (1 = very unfavorable, 9 = very favorable), and To what extent do you believe this person
will like you? (1 = not at all, 9 = a great deal). Responses to these two questions were averaged
to form a single expected Impression measure (α = .88).
2.16.3. Results
A between-subject ANOVA predicting Impression with Audience Partitioned (Yes vs.
No) reveals that, as predicted, respondents expect the non-addressed recipient they thought of to
form a better impression if someone else was addressed explicitly (tagged) in the post (M
Partitioned
= 5.39, SD = 1.74 vs. M
NotPartitioned
= 4.69, SD = 1.91, F(1, 199) = 7.39, p < .01, ωp² = .031).
2.16.4. Discussion
Study 4 shows how those who share self-enhancing content expect to make a better
impression on non-addressed audience members if the audience is partitioned versus not. In other
32
words, audience partitioning appears to be perceived as a potential solution to the self-promoter
dilemma. In study 4, I test the effect of a partitioned audience on impressions but do so only with
content that is clearly self-enhancing. However, as I argued before, the extent to which the
shared content is perceived as self-enhancing should play a role in driving expectations. In study
5, I compare expected impressions when the content shared is either self-enhancing or not self-
enhancing in nature.
2.17. Study 5
2.17.1. Overview
In study 5, I vary the extent to which the content individuals intend to share is self-
enhancing in nature. First and foremost, in this study, I predict a positive effect of audience
partitioning on expected impressions when the content shared is self-enhancing in nature (as
shown in study 4) but not when it is not. Second, I set out to test whether the findings of study 4
would replicate in a different setting (face-to-face interactions) and with self-generated content.
Finally, I examine why individuals sharing self-enhancing content expect to make a better
impression when they use audience partitioning. A priori, my prediction is that, by singling out
one or more recipients as addressees, respondents would expect to come across as less attention-
seeking by everyone else in the audience. Thus, in this study, I delve more deeply into the
underlying process.
2.17.2. Method
In this study, I recruited 251 participants (55.8% female, M
age
= 38.9) through Amazon
Mechanical Turk (mTurk). I manipulated Content (Self-Enhancement vs. Control) by asking
respondents to write a self-relevant fact/story that was either self-enhancing or simply
interesting, as I did in study 1B. Next, respondents were asked to imagine they were having
33
dinner with a group of people they knew, including their friend Alex (the potential addressee)
and another friend they recently made, Scotty (the potential bystander). In this study, audience
partitioning was manipulated within subjects. Respondents were instructed to imagine two
different situations: i) they shared the content they wrote about by addressing Alex directly,
knowing that everyone at the table, including Scotty, would be able to hear them, and ii) they
addressed no one in particular and told everyone at the table simultaneously. The order in which
these two scenarios were presented was counterbalanced.
After reading each scenario (Audience Partitioned: Yes vs. No), respondents were
instructed to think of how Scotty would react in this situation and answer the same set of
Impression questions asked in study 4. They were also asked to report how attention-seeking
they expected to appear in the following way: To what extent do you believe Scotty will think
you told the story to get people's attention? (1 = not at all, 9 = a great deal).
Finally, as a manipulation check, respondents were instructed to think back to the
story/fact they reported and were asked: To what extent was it self-enhancing (a fact/story that
would make you look good in the eyes of others)? (1 = not at all, 9 = a great deal).
2.17.3. Results
Manipulation Check. A between-subject ANOVA indicates the content respondents
chose to share in the Self-Enhancement condition was considered more self-enhancing
(M
SelfEnhancement
= 6.81, SD = 1.87) compared to that shared in the Control condition (M
Control
=
4.17, SD = 2.46, F(1,249) = 91.28, p < .01, ωp² = .265).
Impression. Notably, the order in which the scenarios were presented did not impact any
of my results and will therefore not be discussed further. A mixed ANOVA predicting
Impression with Content (between-subject) and Audience Partitioned (within-subject) only
34
shows a significant interaction (F(1, 249) = 15.15 p < .01, ωp² = .053). Simple contrasts reveal,
as predicted, that in the Self-Enhancement condition respondents expected Scotty to form a
better impression if the audience was partitioned and Alex was addressed (M
Partitioned
= 6.29, SD
= 1.33) as opposed to no one in particular being addressed (M
NotPartitioned
= 6.01, SD = 1.67, F(1,
249) = 4.92, p = .03, ωp² = .02). Critically, for respondents in the Control condition, this was not
the case. In fact, I observe that the opposite holds true (M
Partitioned
= 5.90, SD = 1.46, vs.
M
NotPartitioned
= 6.30, SD = 1.46, F(1, 249) = 10.89, p < .01, ωp² = .04). While I did not predict this
reversal a priori, the result is consistent with respondents believing those not addressed could
feel ignored, which may explain why respondents typically do not want to segment the audience
for non-self-enhancing content.
Attention Seeking. A mixed ANOVA predicting Attention Seeking with Content
(between-subject) and Audience Partitioned (within-subject) reveals the presence of two main
effects. Not surprisingly, respondents felt they would be perceived as more attention-seeking
when sharing self-enhancing content (M
SelfEnhancement
= 5.43, SD = 2.24 vs. M
Control
= 4.38, SD =
2.34, F(1,249) = 19.06, p < .01, ωp² = .067). Importantly, they also felt they would also be
perceived as more attention-seeking if they did not partition the audience and addressed the
whole table (M
NoPartitioned
= 5.55, SD = 2.33) as opposed to having partitioned the audience by
addressing Alex (M
Partitioned
= 4.23, SD = 2.19, F(1, 249) = 89.50, p < .01, ωp² = .261),
independently of the content of the message (F
Interaction
(1, 249) = 0.12, p = .73).
Mediation Analysis. Recall that I propose that the concern with being perceived as
actively seeking others’ attention when delivering self-enhancing content would be perceived as
having a negative effect on the impressions formed by others. In other words, I expect audience
partitioning to positively influence expected impressions through an expectation of being
35
perceived as less attention-seeking. I expected this to be the case when the content shared was
self-enhancing (Self-Enhancement condition) but not in the Control condition.
Given the within-subject nature of my data, computing a mediational model is not
straightforward. Hence, I use Generalized Structural Equation Modeling (GSEM) to assess
whether Attention Seeking mediates the relationship between Audience Partitioned and
Impression for respondents in the Self-Enhancement condition. A model with 5,000 bootstrap
samples reveals the presence of a significant indirect effect (b = .43, SE = .08, 95% CI [.25,
.62]), providing formal evidence of mediation. In contrast, the same model applied to
respondents in the Control condition reveals a non-significant indirect effect (b = .05, SE = .06,
95% CI [-.07, .19]).
2.17.4. Discussion
I designed study 5 with two main objectives in mind. First, this study compares the effect
of audience partitioning on impressions for content deemed self-enhancing versus not self-
enhancing. I show partitioning an audience improves expected impressions only when the
content being shared is deemed self-enhancing; this is consistent with a belief that audience
partitioning helps solve the self-promoter dilemma. Second, study 5 replicates the findings of
study 4 outside the context of social media communication and does so with self-generated
content respondents consider particularly relevant to them.
Finally, the results from study 5 provide evidence that the reason individuals who share
self-enhancing content expect to make a better impression if they partition the audience is how
they expect to come across – namely as less attention-seeking. This, I argue, has important
consequences for marketers as it influences the likelihood consumers generate word-of-mouth
36
about a product or service experience. I test the downstream consequence on word-of-mouth in
study 6.
2.18. Study 6
2.18.1. Overview
Study 6 examines the effect of audience partitioning on word-of-mouth in terms of
consumers’ propensity to share self-enhancing content involving a product or service experience.
I manipulate audience partitioning and test my full conceptual model. I predict that respondents
posting self-enhancing content (in this case, about a vacation in Europe) on Twitter will expect
non-addressed recipients (i.e., bystanders) to see the content as less-attention seeking if the
audience is partitioned (post includes a tag). This in turn should lead to an expectation that
bystanders will form a more positive impression, driving the choice of whether to post the
content on Twitter (i.e., serial mediation). In short, I expect audience partitioning to ultimately
influence the propensity to generate word-of-mouth about an experience.
2.18.2. Method
In this study, 280 college students (47.0% female, M
age
= 20.1) completed the study for
partial course credit. Respondents were asked to imagine they were going on a vacation to
Europe with their friend Max and were considering posting messages (tweets) about the vacation
on Twitter. The content was drawn from real tweets and pretested to ensure it was perceived as
self-enhancing (see appendix E). I manipulated Audience Partitioned (Yes vs. No) between
subjects, and included three within-subject replicates (Post: 1 vs. 2 vs. 3). All three posts were
addressed either to the respondent’s co-traveler (@Max Rooks) or were not addressed to anyone
in particular. I employed three different posts for reassurance the findings did not depend on the
content of any single post.
37
After reading each post, respondents reported how likely they would be to post the
message on Twitter on a 9-point scale. Next, they rated the impression they expected readers of
the post to have of them using the same measures employed in previous studies (α = .89). They
also reported their expectations with regard to being seen as attention-seeking in the following
way: To what extent will they think you wrote the post to get people’s attention? (1 = Not at all;
9 = A great deal).
The model tested in this study is presented in figure 2; solid lines denote the predicted
relationships while dashed lines denote estimated but not hypothesized relationships.
FIGURE 2-3. Model Tested in Study 6.
The model tested includes a Random Effect at the Respondent Level and the variable
Post as a Covariate.
2.18.3. Results
I begin by presenting analyses of the relationships between my independent variable and
the measured variables. I follow this with a concurrent test of the full model presented in figure
2. Notably, while the three posts I showed respondents differed in terms of Likelihood of
Sharing, Attention Seeking and expected Impression, I never observe a significant interaction of
Post with Audience Partitioned. Therefore, while Post is always included in my analyses, I only
discuss the effect of my focal treatment variable Audience Partitioned, while the full results
including those for Post are presented in appendix F.
38
Likelihood of Sharing. A mixed ANOVA predicting Likelihood of Sharing reveals a
significant main effect of Audience Partitioned such that respondents were more likely to post a
message if it was addressed to a specific individual (M
Partitioned
= 2.61, SD = 2.16 vs. M
NoPartitioned
=
2.23, SD = 1.94, F(1, 278) = 5.41, p = .02, ωp² = .016).
Expected Impression. Similarly, a mixed ANOVA predicting the expected Impression
reveals a main effect of Audience Partitioned. The expected impression was more positive if the
audience was partitioned (M
Partitioned
= 3.73, SD = 1.86, vs. M
NoPartitioned
= 3.42, SD = 1.76, F(1,
278) = 6.20, p = .01, ωp² = .018).
Attention Seeking. The pattern of results is similar when I predict Attention Seeking; I
find the expected main effect of Audience Partitioned such that respondents expected to be
perceived as less attention-seeking when the audience was partitioned (M
Partitioned
= 7.00, SD =
2.89 vs. M
NoPartitioned
= 7.37, SD = 1.80, F(1, 278) = 6.15, p = .01, ωp² = .018).
Mediation Analysis. I now turn to testing the full model presented in figure 2. The model
depicted implies sequential mediation with repeated measures and is one variation of two-level
mediation models (Krull and MacKinnon 2001). Because each respondent was asked to consider
three different possible posts, I include Post as a covariate in each equation in the model. I also
include a random effect at the Respondent level to account for the repeated measure structure of
the data. To test this specification, I chose a GSEM approach available in Stata version 14, which
allows sequential mediation paths with covariates and random effects for repeated measures
(Preacher, Zyphur, and Zhang 2010) in the model.
The full results of the GSEM model are presented in appendix F. Critically for me, the
results indicate a statistically significant indirect effect (b = .06; z = 2.37; p = .02) detected for
the expected serial mediation path Audience Partitioned → Attention-Seeking → Impression →
39
Likelihood of Sharing. A bias-corrected bootstrap with 5,000 repetitions further confirms the
significance of the serial mediation indirect effect (95% CI [.03; .10]).
The use of GSEM allows me to check the order of the mediation effect as my two
mediators are measured simultaneously, and reversing their order may lead to two non-
equivalent models (Little et al. 2007). A model in which the indirect effect of Impression comes
before that of Attention Seeking reveals a non-significant serial mediation effect (b = .01; z =
1.63; p = .11), consistent with the causal order proposed in my conceptual model.
2.18.4. Discussion
In study 6, as in study 5, I observe that individuals who use audience partitioning when
sharing self-enhancing content expect their (non-addressed) audience to believe they are less
attention-seeking and thus ultimately form a more positive impression of them in response to the
content being shared. Moreover, I show that these expectations influence their likelihood of
sharing that content.
The effect on likelihood of sharing documented in this study is important to marketers
and our understanding of WOM communication because coming across as overtly self-
enhancing can inhibit consumers from talking about their experiences and purchases (Berger
2014). In this study, I find audience partitioning drives the ultimate choice of whether or not
respondents choose to post tweets about their vacation on Twitter, thus promoting WOM.
Across a number of studies, I have shown that consumers use audience partitioning as a
solution to the self-promoter dilemma, with the expectation that non-addressed recipients will
see them as less attention-seeking and therefore form a more positive impression of them. We
know from previous literature that, at times, self-promoters miscalibrate their self-presentational
efforts (Scopelliti et al. 2015; Sezer, Gino, and Norton 2018). It is therefore important to ask
40
whether audience partitioning is indeed a solution to the self-promoter dilemma. In Part III I take
the perspective of the receiver of a self-enhancing communication and ask whether the use of
audience partitioning does indeed affect perceptions of the self-promoter.
2.19. Empirical Analysis – Part III
In Part III I move the focus of my investigation from the sender to the receiver of a self-
enhancing communication. I ask whether non-addressed recipients (bystanders) in a partitioned
audience form a more positive impression of the self-promoter compared to members of a non-
partitioned audience.
I show that this is indeed the case. In other words, audience partitioning is effective at
mitigating the negative consequences of self-promotion. This happens because, as expected by
the self-promoter, bystanders see him or her as less attention-seeking and this has a positive
effect on the overall impression they form.
The reason why bystanders tend to give the benefit of the doubt to a self-promoter who is
using audience partitioning is their expectation that there is a reason why the self-promoter is
sharing the self-enhancing content specifically with the addressee, other than the desire of getting
everyone’s attention. Importantly, I show that when this is not the case, and bystanders cannot
think of a good reason for the self-promoter to share the message with the addressee, audience
partitioning no longer affecting impressions.
2.20. Study 7
2.20.1. Overview
Study 7 examines the effect of audience partitioning on the actual impressions that
bystanders form of the self-promoter. Respondents were asked to imagine they were sitting at a
dinner table and they heard one of their dinner companions share a message pre-tested to be
41
perceived as self-enhancing. I manipulate audience partitioning by telling them the message was
either addressed to the whole table or just to the person sitting next to the self-promoter (such
that they were bystanders to the communication). If audience partitioning is indeed a successful
self-presentation strategy, I should find that bystanders see the self-promoter as less attention-
seeking and ultimately form a more positive impression of him/her when they are bystanders to
the communication, as opposed to members of a non-partitioned audience.
2.20.2. Method
I recruited participants through Amazon Mechanical Turk (mTurk) and applied the same
selection criteria used in previous studies. A total of 200 participants (49.0% female, M
age
=
35.3) completed the survey.
Similar to previous studies, respondents read a short scenario about sitting at a table with
a group of people they know, including Scotty and Scotty’s friend Alex. While in previous
studies respondents were the ones sharing a self-enhancing message, in this case they were asked
to evaluate Alex. I manipulated Audience Partitioned (Yes vs. No) by telling respondents that
Alex addressed the following self-enhancing message: I aced an IQ test this morning. I knew I
was a genius! either to Scotty, sitting next to Alex, or to the whole table. In other words, in the
Audience Partitioned condition they were non-addressed members (bystanders) of a partitioned
audience, while in the Audience Not Partitioned condition they were members of a non-
partitioned audience.
Respondents were asked to evaluate Alex by answering the following questions: What is
your overall impression of Alex? (1 = very unfavorable, 9 = very favorable) and To what extent
do you like Alex? (1 = not at all, 9 = a great deal). Responses to these two questions were
averaged to form a single Actual Impression measure (α = .90). Next, respondents evaluated how
42
attention-seeking they perceived Alex to be by answering: To what extent do you believe Alex is
telling the story to get everyone's attention? (1 = not at all, 9 = a great deal).
2.20.3. Results
Actual Impression. An ANOVA predicting the Actual Impression respondents formed of
Alex reveals a main effect of Audience Partitioned. Their impression of Alex was more positive
if the audience was partitioned (M
Partitioned
= 4.80, SD = 1.73, vs. M
NoPartitioned
= 4.15, SD = 1.75,
F(1, 198) = 6.95, p < .01, ωp² = .029).
Attention Seeking. The pattern of results is similar when I predict Attention Seeking; I
find the expected main effect of Audience Partitioned such that respondents perceived Alex to be
less attention-seeking when he addressed his message to Scotty, sitting next to him, as opposed
to the whole table (M
Partitioned
= 6.97, SD = 1.68 vs. M
NoPartitioned
= 7.67, SD = 1.39, F(1, 198) =
10.25, p < .01, ωp² = .044).
Mediation Analysis. Next, I turn to test the prediction that the effect of audience
partitioning on actual impressions is mediated by the perceptions that the self-promoter is
differentially attention-seeking. A model with 5,000 bootstrap samples reveals the presence of a
significant indirect effect (b = .27, SE = .08, 95% CI [.10, .50]), providing formal evidence of
mediation.
2.20.4. Discussion
Results of Study 7 indicate that audience partitioning is indeed a successful strategy to
mitigate the negative consequences of self-promotion. That is, receivers of a self-enhancing
message form a more positive impression of the self-promoter when they are bystanders in the
communication. This happens because they see the self-promoter as less attention-seeking.
Instead, they may attribute their choice of sharing the message with the addressee to some other
43
reason, other than the desire of getting everyone’s attention. I formally test this hypothesis in
study 8, by manipulating the reason why the self-promoter choses to partition the audience.
2.21. Study 8
2.21.1. Overview
Study 8 documents the effect of audience partitioning on impressions in a different
domain, social media communication, where audience partitioning is operationalized as tagging.
There is an important difference between partitioning one’s audience in face-to-face
communication and on social media. A bystander who hears a self-promoter sharing a self-
enhancing message with a specific addressee at a dinner table might not know whether the self-
promoter is even aware that others are listening to what is being said. In other words, the
bystander might believe he/she is overhearing a message that was not intended for him/her to
hear. If this is the case, the bystander might feel more inclined to cut some slack to the self-
promoter and evaluate this person less harshly. This is not the case for public social media
messages, where bystanders have no doubt that, if the public message was posted, it was meant
to be seen by others than just the tagged addressee. It is therefore important to ask whether the
positive effect of audience partitioning on impressions also holds when the bystander realizes
s/he is not overhearing the communication by accident, as in the case of social media
communication.
Moreover, study 8 was designed to show that the reason why audience partitioning works
at affecting impressions of a self-promoter is because bystanders generally believe that there is a
reason why the self-promoter is sharing the self-enhancing content specifically with the
addressee, other than the desire of getting everyone’s attention. Importantly, study 8
demonstrates that when this is not the case, and bystanders cannot think of a good reason for the
44
self-promoter to share the message with the addressee, audience partitioning no longer an
effective self-presentation strategy.
2.21.2. Method
Three hundred and eighty-four college students (50.0% female, Mage = 20.0) completed
the survey for partial course credit.
The study had a 4 (Audience Partitioned: No vs. Yes vs. Yes Justified vs. Yes Not
Justified x 2 (Topic: IQ vs. Car) between-subject design. Respondents were asked to evaluate the
Facebook profile of a user named Alex Watson, based on Alex’s latest Facebook post. Similar to
Study 2A, the post either referenced an IQ test or a car purchase (see Appendix G). These are
simply included as replicates and, a priori, I do not expect topic to matter.
In the Audience Not Partitioned condition, the Facebook post was not addressed to any
user in particular. In the Audience Partitioned condition, the post tagged/addressed a specific
Facebook user, Scotty Thompson. This set up mimics that of earlier studies. Moreover, two
additional conditions were added. In both conditions, the post tagged Scotty Thompson, as in the
Audience Partitioned condition, but some additional information was also provided. In the
Justified Audience Partitioned condition respondents read: You notice that Scotty Thomson,
whom you know, was tagged in the post. This makes perfect sense because Scotty is a car
enthusiast [enjoys this type of tests] and you recently overheard him discussing new cars [IQ
tests] with Alex. Instead, in the Not Justified Audience Partitioned condition respondents read:
You notice that Scotty Thomson, whom you know, was tagged in the post. However, you don't
see a reason for this tag.
Respondents were asked to answer the following questions about Alex Watson: What is
your overall impression of Alex? (1 = very unfavorable, 9 = very favorable) and To what extent
45
do you like Alex? (1 = not at all, 9 = a great deal). Responses to these two questions were
averaged to form a single expected Actual Impression measure (α = .91).
I expect to replicate earlier findings that impressions are more positive in the Audience
Partitioned condition compared to the Audience Not Partitioned condition. Moreover, if the
reason why partitioning works is because bystanders attribute partitioning to the relationship
between the self-promoter and the addressee, impressions in the Justified Audience Partitioned
condition should be the same as those in the Audience Partitioned condition. Further,
impressions in the Justified Audience Partitioned condition should be more positive compared to
the Audience Not Partitioned condition, but that should not be the case in the Not Justified
Audience Partitioned condition.
2.21.3. Results
As expected, the Topic of the message (IQ vs. Car) did not affect any of my results.
Therefore, I collapsed the data across conditions.
An ANOVA predicting Actual Impressions reveals a marginally significant main effect
of Audience Partitioned (F(3, 380) = 2.54, p = .056, ωp² = .012). Most important are the planned
contrast analysis reported below.
Audience Partitioned vs. Audience Not Partitioned – Replicating prior findings,
impressions of Alex were significantly more positive in the Audience Partitioned condition
(M
Partitioned
= 3.77, SD = 1.46, vs. M
NoPartitioned
= 3.29, SD = 1.56, F(1, 380) = 4.64, p = .03, ωp² =
.009).
Audience Partitioned Justified vs. Audience Not Partitioned – As expected, impressions
of Alex were also significantly more positive in the Audience Partitioned Justified condition
46
compared to the Audience Not Partitioned condition (M
Justified
= 3.81, SD = 1.75, vs. M
NoPartitioned
= 3.29, SD = 1.56, F(1, 380) = 5.63, p = .02, ωp² = .012).
Audience Partitioned Justified vs. Audience Partitioned – Impressions of Alex did not
differ between conditions (M
Justified
= 3.81, SD = 1.75, vs. M
Partitioned
= 3.77, SD = 1.46, F(1, 380)
= 0.04, p = .84).
Audience Partitioned Not Justified vs. Audience Not Partitioned – Impressions of Alex
did not differ between these conditions (M
NotJustified
= 3.47, SD = 1.37, vs. M
NoPartitioned
= 3.29, SD
= 1.56, F(1, 380) = 0.65, p = .42).
2.21.4. Discussion
Study 8 shows that audience partitioning is effective at improving impressions of
bystanders also in the context of social media communication, where bystanders should
generally realize that, if the public message was posted, it was meant to be seen by others than
just the tagged addressee. This speaks to the effectiveness of this self-presentation strategy.
Moreover, study 8 helps us better understand why audience partitioning works. The
results of this study suggest that bystanders feel there must be a reason why the self-promoter is
sharing the self-enhancing content specifically with the addressee. This is one of reasons they
tend to give the benefit of the doubt to a self-promoter who is using audience partitioning.
Importantly, this study reveals a boundary condition to the effectiveness of audience partitioning:
when bystanders cannot think of a good reason for the self-promoter to share the message with
the addressee, audience partitioning no longer affecting impressions.
2.22. Chapter Discussion
Imagine being at a dinner table with a group of friends and casually mentioning to your
spouse that you are planning on going to the mall to get new luggage to fit all the clothes you
47
need to take with you for your month-long vacation in Europe. While you might feel too
presumptuous making such a remark to everyone at the dinner table, you may still want dining
companions to hear about the vacation vis-à-vis the discussion with your spouse regarding the
luggage purchase. You might therefore feel that putting others on the sideline in the role of
bystanders is a viable strategy for sharing such self-enhancing content.
Indeed, individuals are aware that a balance must be struck when actively trying to make
a positive impression on others. They recognize that portraying themselves too favorably runs
the risk of their behavior backfiring and ultimately coming across as boastful and intentionally
attention-seeking. Consequently, they employ various strategies to facilitate self-promotion
through the delivery of self-enhancing information. I identify a novel strategy that focuses not on
varying the content of the message, but on strategically altering the structure of one’s audience. I
show that individuals sharing self-enhancing content at times intentionally address specific
recipients thereby turning everyone else into bystanders on the sideline, a behavior I refer to as
audience partitioning. In doing so, self-promoters expect non-addressed recipients to find self-
enhancing content less attention-seeking and consequently form a better impression of them in
light of the new information. Audience partitioning is thus used as an intended solution to the
self-promoter dilemma and facilitates WOM in contexts where consumers worry that talking
about their purchases and consumption experiences may make them come across as overtly
attention-seeking. The results reported in this chapter document this effect. Importantly, the last
two studies demonstrate that audience partitioning is, in most cases, a successful strategy.
Bystanders do indeed form more positive impressions of self-promoters and do so because they
see the self-promoter as less attention-seeking.
48
To the best of my knowledge, this work is the first to investigate how and why
individuals choose to partition audiences, putting a portion of their audience in the position of
perceiving themselves as bystanders exposed to communication not directed specifically at them,
as well as how partitioning affects how the audience responds to the message being shared. We
know that the presence of bystanders can have important effects on the way individuals behave,
sometimes inhibiting (Latane´ and Darley 1968; Latane´ and Nida 1981) and other times
facilitating (Zajonc 1965) certain behavioral responses. Yet previous work has largely
investigated how individuals react to the incidental presence of bystanders. I show that, under
certain conditions, individuals strategically alter the participant structure of an audience and
proactively put others in the role of bystanders. In doing so, I contribute to the impression
management literature by showing how this behavior can manifest itself in order to facilitate
self-presentation efforts, and in particular the efficacy of self-enhancing communication.
This work also adds to our understanding of how computer-mediated communication
influences individuals’ self-presentation behavior in terms of how users share information
(Barasch and Berger 2014; Schau and Gilly 2003; Schlosser 2009). Digital networks and social
media provide various tools (such as tagging) that vastly simplify people’s ability to change the
structure of their audience, thus providing great discretion in defining who are addressees and
who are bystanders. Importantly, these findings also contribute to the literature on WOM and
have important implications for practitioners by providing indications on how consumers might
be encouraged to generate word-of-mouth about purchases and consumption experiences. For
example, to the extent that a concern with coming across as a braggart may inhibit word-of-
mouth, marketers might have an easier time getting customers to discuss their purchases and
49
consumption experiences on social media if consumers are encourage to partition their
audience—perhaps through a tag a friend campaign.
An important question not addressed in this work concerns both the expected and actual
impression the message will have on the addressed recipient (as opposed to bystanders).
Naturally, the addressee is an important party in the communication whose role also deserves
further investigation. Many important questions come to mind concerning recipients addressed
by the self-promoter. First, what is the (expected and actual) effect that audience partitioning has
on the addressee? Relatedly, is there a particular type of individual who is more likely to be
singled out as an addressed recipient? We know from previous literature that concerns with
making a good impression are dampened if the opposing party is someone with whom the person
is highly familiar with (Gosnell, Britt and Mckibben 2011; Leary et al 1994). Consequently, my
expectation is that individuals might prefer addressing a close other for two main reasons: (1)
close others could be more interested in actually being informed about the self-enhancing content
shared, consequently not attributing it to the self-promoters’ desire to seek attention, and, (2)
even if they perceive the content as self-enhancing they are likely to be more forgiving and less
prone to alter their (already well established) impression of the self-promoter based on a single
act of self-promotion. It is my belief that self-promoters generally tend to address close others in
an attempt to make a good impression on more distant others while being less likely to be judged
negatively by the addressed close other.
Impression management is only one of many goals individuals try to achieve when
communicating with others. While I believe impression management is an important and
significant motivator for strategic audience partitioning, interpersonal communication, as
mentioned, serves a variety of functions for individuals in everyday life. It is therefore
50
reasonable to expect there are other contexts in which audience partitioning can be used as a
strategic communication tool. One of such context, I argue, is that of persuasive communication.
In the next chapter I therefore investigate the strategic use of audience partitioning in the context
of persuasive communication.
51
CHAPTER 3: PERSUADING THE BYSTANDER
3.1. Chapter Introduction
Imagine a consumer standing in line at an ATM machine when a volunteer from a non-
profit organization approaches trying to collect signatures for a petition. Even before learning
what the petition is about, the consumer might grow suspicious of the petitioner’s intentions. It is
clear the petitioner will try to persuade the consumer about something and this inferred intention
is enough for the consumer to raise her cognitive defenses as a way of dealing with this
imminent persuasive attempt. Similarly, a sales assistant approaching a consumer browsing a
department store asking whether she would like to try the new Marc Jacobs perfume might be
received with similar levels of skepticism. Indeed, since consumers are subject to persuasion
attempts on a daily basis, they oftentimes recognize others’ persuasive intentions when engaging
in interpersonal interactions and use their knowledge about persuasion to erect cognitive barriers
and cope with these episodes (Friestad and Wright 1994; Rule, Bisanz, and Kohn 1985; Schank
and Abelson 1977; Campbell and Kirmani 2000). In most cases the results of consumers
activating persuasion knowledge (i.e., consumers’ theories about persuasion and beliefs about
persuasion agents’ motives, strategies, and tactics) is greater skepticism towards the persuasion
agent and resistance to the persuasion attempt (see Isaac and Grayson 2017).
What would happen if, instead, a consumer is simply a bystander to a persuasive attempt
addressed to someone else? Compare, for instance, the situation when one individual is
addressed by a volunteer from a non-profit organization trying to collect signatures for a petition
and the other simply listens in to the conversation to the situation when two individuals are
addressed simultaneously. What has changed between these two scenarios is the participation
structure of the petitioner’s audience (Goffman 1959, 1981; McGregor 1986). The participation
structure of an audience describes the roles played by the various audience members. One way in
52
which a speaker can change participation roles is by dividing audience members into two
categories: addressed recipients and non-addressed recipients. I refer to this as audience
partitioning. When the audience is partitioned, those not addressed become bystanders to a
communication addressed at someone else.
In this work I argue that bystanders would be more likely to be persuaded to sign a
petition compared to when they are members of a non-partitioned audience, that is when they
themselves are addressed recipients in the communication. I propose that, when being exposed to
a persuasive attempt addressed at someone else as a bystander, consumers will not feel the same
need to be prepared to deal with a persuasion attempt that they would feel if they weren’t simple
bystanders to the communication. In other words, the persuasive message is met with lower
skepticism and its persuasive effectiveness is greater.
This work is the first to investigate how audience partitioning could be used strategically
to affect persuasion. In doing so, it contributes to the literature on persuasion and social influence
by documenting a novel persuasion strategy that can be of interest to both researchers and
practitioners alike. Moreover, the current work contributes to the literature on social cognition
and communication, which has yet to investigate empirically whether, when and how people
manage the participation structure of an audience in the context of persuasive communication.
Notably, past research has proposed that, under certain conditions an overheard communicator
can be more effective compared to a regular communicator (Walster and Festinger 1962; Brock
and Becker 1965). Walster and Festinger (1962) found that respondents who believe they are
overhearing a conversation between two individuals are more persuaded by the arguments they
heard compared to respondents who think the individuals involved in the conversation are aware
they are listening in. Importantly, in these settings, the structure of the audience is not altered
53
strategically by the speaker. Rather, listeners make different inferences about what the speaker
thinks and knows about the audience. In my work, I instead focus on how a persuasive agent
actually alters audience members’ roles.
3.2. Chapter Overview
The remainder of this chapter proceeds as follows. I begin by briefly reviewing the
relevant literature on persuasion and on factors affecting consumers’ resistance to persuasion. I
then summarize the work in marketing that has looked at targeted marketing and targeted
advertising, as this work deals with persuasion and marketer-driven participation structure of an
audience. My conceptualization, taken up next, expands upon this literature to explain how and
why audience partitioning is expected to affect how a persuasive message is received. Finally, I
present two laboratory studies in support of my conceptualization.
3.3. Literature Review
3.3.1. Persuasion, Persuasion Knowledge and Resistance to Persuasion
Persuasion involves changing individuals’ attitudes and beliefs, usually as precursor of
behavioral change (Chaiken, Wood, and Eagly 1996; O’Keefe 2002). The study of persuasion
dates back to the early days of social psychology (Petty and Cacioppo 1986). While an abundant
body of work in this literature has focused on strategies used by persuasion agents to influence
targets (Kellerman and Cole 1994), particularly relevant to the current work is prior research that
focused specifically on the target of persuasion. That research explored how consumers manage
persuasion attempts directed at them and the factors affecting their ability to resist persuasion
(Eagly and Chaiken 1995; Petty and Cacioppo 1986; Ahluwalia 2000; Karmarkar and Tormala
2009). A key finding is that consumers tend to erect cognitive defenses to resist persuasion
attempts (Russell 2002). These cognitive barriers are higher the more consumers are likely to
54
detect the persuasive intent of an agent, and the more they perceive the agent’s actions are the
result of an undue manipulative intent (Campbell 2005).
The Persuasion Knowledge Model (Friestad and Wright 1994), for instance, postulates
that beliefs and inferences about the persuader’s motives, as well as about the strategies, and
tactics used by the persuader are important precursors to consumers’ ability to resist persuasion
(see also Rule, Bisanz, and Kohn 1985; Schank and Abelson 1977; Campbell and Kirmani
2000). When consumers perceive the intention to persuade as the main motive underlying an
agent’s behavior, they are generally more likely to resist the persuasion attempt. While at times
persuasion knowledge might lead to the conclusion that the agent’s behavior is warranted and
therefore boost credibility (Isaac and Grayson 2017), in most cases the result of activating
persuasion knowledge is greater skepticism towards the intentions of the persuasion agent and
resistance to the persuasion attempt. For example, consumers use persuasion knowledge to
evaluate the extent to which a salesperson’s flattering remarks reflect the ulterior motive of
persuading the customer to buy the product (Campbell and Kirmani 2000). If flattery is attributed
to the salesperson’s desire to persuade, impressions of the salesperson’s sincerity are affected.
The activation of consumers’ persuasion knowledge has been shown to affect a set of related
marketing outcomes, including brand evaluations (Kirmani and Zhu 2007), firm evaluations
(Forehand and Grier 2003), advertising effectiveness (Xu and Wyer 2010; Phillips and
McQuarrie 2010), response to promotional activities (Nunes and Dreze 2006), and likelihood of
participating in vice behaviors (Fitzsimons, Nunes, and Williams 2007).
A distinct yet related body of work studies the phenomenon of psychological reactance
and leads to similar outcomes. Psychological reactance describes the following phenomenon.
When individuals perceive their freedom to be threatened, they attempt to restore such freedom
55
by exhibiting opposition or by resisting pressures to conform (Brehm 1966; Brehm and Brehm
2013). Psychological reactance has been studied also in the context of marketing and has been
put forward as an explanation as to why influence attempts can backfire (Clee and Wicklund
1980, p. 389). When confronted with high-pressure salespeople, in particular, consumers have
been found to react with a desire to reassert their freedom not to purchase anything, leading them
to resist the persuasion attempt (Wicklund, Slattum, and Solomon 1970; Reizenstein 1971).
Similarly, the perception that one’s behavior is externally driven rather then due to individual
agency and freedom can impair consumers’ response to promotional activities (Kivetz 2005), as
well as their response to recommendations by experts and intelligent agents (Fitzsimons and
Lehman 2004). Again, similar to what literature on persuasion knowledge would predict,
external agent seems to be trying to persuade consumers and reactance leads them to resist the
persuasion attempt.
3.3.2. Targeted Communication and Persuasion
In this work, I ask whether being a bystander to a persuasive attempt addressed to
someone else, as opposed to being a member of a non-partitioned audience targeted by the
persuasion agent, can affect how consumers evaluate the persuasion agent’s intentions, their
cognitive reactions, and ultimately the persuasive effectiveness of the message. One stream of
literature that has dealt with persuasion and the participation structure of an audience is the
literature on targeted marketing and targeted advertising.
Overall, there are many known benefits of tailoring a persuasive message to a consumer,
or a group of consumers (i.e., targeting). A clear benefit of targeting comes from the possibility
of providing to the targeted consumers only the information more relevant to their purchase
decisions. Importantly, nonetheless tailoring itself can be beneficial as it works as a signal that
56
the communication was thought with the interests of those consumers in mind. These effects
have been studied mostly in the context of targeted marketing and advertising (Whittler 1989;
Meyers-Levy 1989; Tepper 1994; Williams and Qualls 1989; Deshpande and Stayman 1994).
Benefits of targeted communication have been identified also in the context of online
advertising, where personalization can be achieved at the individual consumer level (Tucker
2014). Relatedly, Aaker, Brumbaugh, and Grier (2000) talk about non-target market effects
which refers to the negative reactions that consumers may have when they feel an advertising
communication is not customized to them, but rather was created with different consumers in
mind. Non-target consumers can view a targeted advertising as distracting and irritating (Star
1989), and they may feel neglected (Greco 1989).
Yet, there are times when targeting could actually be detrimental to advertising
effectiveness. In work investigating adolescents’ evaluations of persuasive anti-drug messages,
Crano, Siegel, Alvaro, and Patel (2007) found that messages targeted at an audience of parents
had a more powerful impact on adolescents than messages targeting adolescents directly. This
was the case in particular for heavy users of toxic substances, who had a more negative reaction
towards messages targeting adolescents directly. These findings seem to suggest that, at times,
targeting can have negative consequences. This is perhaps because, in the context of social
campaigns dealing with consumers’ health and well-being, targeted consumers employ defense
mechanism to protect their identity and self-image (see also Puntoni, Sweldens, and Tavassoli
2011).
3.4. The Current Investigation
There is a key difference between my work and the literature reviewed in the previous; in
prior work, the audience is partitioned at the group level following the principles of
57
segmentation, such that the addressed audience consists of a well-identified group of consumers
with consistent needs (different from those of the non-addressed audience). My work, I focus
instead on audience partitioning in the context of interpersonal communication when partitioning
is conducted at the individual level, and addressed and non-addressed recipients do not belong to
different segments with different needs and characteristics.
Consistent with previous work on persuasion and persuasion knowledge, I argue that
members of a non-partitioned audience who are exposed to a persuasive message, will oftentimes
recognize that they are subject to a persuasive attempt. This is particularly true in marketing
contexts such as personal selling. In these cases, consumers will question the intentions of the
persuasion agent and grow skeptical. But what if the message is instead addressed to someone
else? For instance, what if a customer in an auto showroom hears the salesperson praising a new
car while talking with a different customer? In this case the persuasion attempt is clearly directed
at another individual. I argue that the customers will be less likely to question the intentions of
the persuasion agent, and, paradoxically, end up being more apt to be persuaded.
In order to show the effects of audience partitioning on persuasion, I adopt a paradigm
whereby, across different persuasive contexts, I compare the persuasive effectiveness of a
message for members of a non-partitioned versus partitioned audiences. In particular, I compare
the reaction of members of a non-partitioned audience with that of non-addressed members of a
partitioned audience (bystanders). I expect bystanders to be more convinced by the persuasive
message compared to members of a non-partitioned audience, and hence to be more likely to
engage in the behavior desired by the persuader. Moreover, I expect this to occur because
bystanders are less likely to question the intentions of the persuasion agent, i.e., exhibiting lower
skepticism. I expect skepticism to be higher for bystanders exposed to a persuasive message
58
addressed to someone else when compared to members of a non-partitioned audience. Further, I
expect skepticism to drive how convinced they ultimately become by the message, as measured
by their attitude towards the objective of persuasion/behavior intended by the persuader. The
conceptual model tested in these studies is presented in figure 3-1.
FIGURE 3-1. Conceptual Model Tested
3.5. Study 1
3.5.1. Overview
The aim of study 1 is to test the prediction that audience partitioning affects the efficacy
of persuasion attempts. More specifically, I test whether members of a partitioned audience
(bystanders) who are exposed to a persuasive message that is addressed not to them but rather to
a different audience member form more positive attitudes towards the objective of persuasion as
compared to members of a non-partitioned audience exposed to the same persuasive message.
Hence, I expose respondents to identical persuasive messages while manipulating their role in
the communication.
59
Moreover, I set out to uncover whether the reason why bystanders form more positive
attitudes towards the objective of persuasion is that they are less likely to question the intentions
of the persuasion agent, resulting in lower skepticism. I therefore measure skepticism about the
intentions of the persuasion agent following Isaac and Greyson’s (2017) approach. These authors
propose that credibility and skepticism operate at opposite ends of the same continuum, and that
lower skepticism/higher credibility perceptions indicate lower persuasion knowledge activation. I
measure respondents’ skepticism about the intentions of the persuasion agent
3.5.2. Method
Two hundred and forty-eight college students completed the study for partial course
credit. Respondents were asked to imagine they wanted to purchase a road bike and read a
scenario describing how they had engaged in some preliminary research and were leaning toward
purchasing a Trek Novo bicycle that was priced at 250 USD. Next, they were told they went to a
bike store and began looking around and heard another customer asking the shop assistant about
road bikes. I manipulated Audience Partitioned (Yes vs. No) by telling respondents that they
simply heard the conversation between the other customer and the shop assistant (Audience
Partitioned condition), or joined the conversation by mentioning they had the same question
(Audience Not Partitioned condition). In the former case, they heard the sales assistant
addressing the other customer, while in the latter case the store assistant addressed the both of
them, simultaneously.
The sales assistant employed a typical upselling technique by recommending a more
expensive bike stating: I would definitely suggest the Trek 3100. I think it is the best performing
bike in the market! Lots of people told me they have felt the difference once they switched. It is a
very versatile bag and definitely worth the 350 USD price tag!
60
After reading the scenario, respondents answered 2 questions about their attitudes
towards the bike recommended by the sales assistant: How would you describe your attitude
toward the Trek 3100? (1= very unfavorable, 9 = very favorable), and How much do you think
you would like the Trek 3100? (1= not at all, 9 = a great deal). Responses to these two questions
were averaged to form a single Attitude measure (α = .78).
Next, I assessed respondents’ skepticism about the speaker’s intentions, by asking them
to rate the extent to which he was: helpful, credible, informative, appropriate, fair, believable,
manipulative, dishonest, fraudulent, improper, unfair, and deceptive. These items were drawn
from work by Isaac and Grayson (2017) and reflect 6 positive and 6 negative persuasion
knowledge characteristics. These authors propose that credibility and skepticism operate on
opposite ends of the same continuum; the 6 positive items measure credibility, while the 6
negative items measure skepticism. Following the authors’ approach, I reverse-coded the
positive items and found the 12 items all load on the same scale (α = .90).
3.5.3. Results
Attitude. An ANOVA predicting the Attitude of respondents towards the more expensive
bike proposed by the sales assistant reveals a main effect of Audience Partitioned. Attitude
towards the bike was significantly more positive when the audience was partitioned as opposed
to not partitioned (M
Partitioned
= 6.04, SD = 1.29, vs. M
NoPartitioned
= 5.67, SD = 1.48, F(1, 246) =
4.28, p = .04, ωp² = .013).
Skepticism. The pattern of results is similar for Skepticism. I find the expected main effect
of Audience Partitioned such that respondents were more skeptical towards the sales assistant’s
intentions (an indicator of lower persuasion knowledge) when the audience was partitioned, and
61
the focal respondent became a bystander to the communication (M
Partitioned
= 4.15, SD = 1.18 vs.
M
NoPartitioned
= 4.60, SD = 1.33, F(1, 246) = 8.02, p < .01, ωp² = .028).
Mediation Analysis. I next turn to test my prediction that the reason why bystanders end
up being more convinced compared to members of a non-partitioned audience is because they
are less skeptical about the sales assistant’s intentions. I test whether Skepticism mediates the
effect of Audience Partitioned on Attitude. The correlation between Skepticism and Audience
Partitioned is r = -.41. A model with 5,000 bootstrap samples reveals the presence of a
significant indirect effect (b = -.19, SE = .08, 95% CI [-.37, -.07]), providing formal evidence of
mediation.
3.5.4. Discussion
In summary, I find that bystanders exposed to an upsell attempt addressed to another
consumer form more positive attitudes towards the objective of persuasion as compared to
members of a non-partitioned audience. Notably, studying the effect of audience partitioning on
persuasive effectiveness in the context of an upsell attempt is a somewhat conservative test.
Respondents were told they had already set their mind on a specific bike, which should have
made them unlikely to change their opinion (Fitzsimons and Lehman 2004). Yet, I find that non-
addressed members of a partitioned audience have a more favorable attitude towards the more
expensive bike proposed by the sales assistant. This occurs because bystanders don’t question the
persuasion agent’s intentions as much, as highlighted by lower skepticism rating.
In study 2, I replicate the findings of study 1 in a different persuasive context.
62
3.6. Study 2
3.6.1. Overview
The aim of study 2 is to replicate the findings of study 1 in a different persuasion setting
in which the use of audience partitioning as a persuasion strategy could play an important role. In
this study, respondents are approached by an individual who asks them to sign a petition in the
street. Replicating my findings in this context would add to the generalizability of my effect,
since these two settings are very different: those entering a store might expect to be subject to
persuasive attempts and therefore be prone to skepticism; those approached by a petitioner in the
street might instead be less likely to question the intentions of those who approach them.
3.6.2. Method
Two hundred and six college students completed the survey for partial course credit.
Respondents read a short scenario about being in line to withdraw some cash at the campus ATM
machine with another person in line behind them. Participants were then told that a girl walked
over towards the line. I manipulated Audience Partitioned (Yes vs. No) by telling respondents
that the girl either approached the person behind them (Audience Partitioned condition) or
approached both them and the person behind them (Audience Not Partitioned condition). In both
scenarios the girl introduced herself as Alex, while the other person in line introduces himself as
Sam.
All respondents the listened to a one-minute audio file of the persuasive message the girl
used. In the message, the girl tries to persuade her audience to sign a petition to introduce exam
retakes in their school. Exam retakes were selected as a topic for the persuasive message based
on a pre-test. The pre-test asked 177 respondents from the same population to rate statements
about a variety of issues on a 9-point scale (1 = totally disagree, 9 = totally agree). I selected a
63
topic for which agreement was neither too high, nor too low (M = 5.29, SD = 2.77), in order to
ensure that respondents would not already have strong opinions on the issue.
The file was recorded by a hypothesis bling professional voice actor and was edited so
the message differed slightly between conditions to manipulate audience partitioning (see
transcript in Appendix H). In the Audience Partitioned condition the girl explicitly addressed the
person standing behind the respondent by mentioning his name (Sam) 3 times during the
recording, while in the Audience Not Partitioned condition the girl addressed both the respondent
and the person standing behind them (no name was mentioned during the recording).
After listening to the audio file, respondents answered 3 questions about their attitudes
towards exam retakes: How would you describe your attitude toward exam retakes? (1= very
unfavorable, 9 = very favorable), How much do you think you would like exam retakes at your
school? (1= not at all, 9 = a great deal), and How likely would you be to sign the petition in favor
of exam retakes at your school? (1= very unlikely, 9 = very likely). Responses to these three
questions were averaged to form a single Attitude measure (α = .93).
Next, respondents rated their skepticism about the intentions of the speaker, rating her on
the same 12 attributes drawn from by Isaac and Grayson (2017), again reflecting 6 positive and 6
negative persuasion knowledge characteristics (α = .93).
3.6.3. Results
Attitude. An ANOVA predicting the Attitude of respondents towards exam retakes
reveals a main effect of Audience Partitioned. Attitudes towards retakes were more positive
when the audience was partitioned, i.e. the petitioner addressed the person standing behind them
(M
Partitioned
= 7.15, SD = 1.82, vs. M
NoPartitioned
= 6.43, SD = 2.25, F(1, 204) = 6.43, p = .01, ωp² =
.026).
64
Skepticism. The pattern of results is similar for Skepticism about the petitioners’ intention.
I observe the expected main effect of Audience Partitioned such that respondents are less
skeptical about the petitioner’s intentions when she partitioned her audience, such that
respondents became bystanders to the communication (M
Partitioned
= 2.38, SD = 1.10 vs.
M
NoPartitioned
= 2.76, SD = 1.09, F(1, 204) = 6.20, p = .01, ωp² = .025).
Mediation Analysis. I next turn to test my prediction that the reason why bystanders end
up being more persuaded compared to members of a non-partitioned audience is because they are
more skeptical about the speaker’s intentions. I test whether Skepticism mediates the effect of
Audience Partitioned on Attitude. The correlation between Skepticism and Audience Partitioned
is r = -.54. A model with 5,000 bootstrap samples reveals the presence of a significant indirect
effect (b = -.37, SE = .25, 95% CI [-.71, -.07]), providing formal evidence of mediation.
3.6.4. Discussion
Once again, I find that bystanders exposed to a persuasive message directed to another
audience member have more positive attitudes towards the objective of persuasion compared to
members of a non-partitioned audience. This happens also when individuals are approached in a
non-sale setting, like the one used in study 2. This is important because, while individuals who
go to a store probably expect to be subject to persuasive attempts, and therefore might be more
skeptical to begin with, in the scenario of study 2 this is less likely to be the case. In other words,
partitioning seems to affect persuasion irrespective of the baseline skepticism level dictated by
the context in which the persuasion attempt takes place.
3.7. Chapter Discussion
To the best of my knowledge, this work is the first to investigate how audience
partitioning affects the effectiveness of a persuasive message. I find that audience partitioning is
65
an effective strategy to persuade non-addressed recipients (bystanders to a persuasive attempt
directed at another audience member). I show that when the persuasive attempt is directed at
someone else consumers are less skeptical about the intentions of the persuasion agent, which
translates into more positive attitudes towards the message itself.
This work contributes to the literature on persuasion and social influence, by introducing
audience partitioning as a novel element that affects the effectiveness of a persuasive message, as
well as the literature on social cognition and communication, which has yet to investigate
empirically whether, when and how people manage the participation structure of an audience in
the context of persuasive communication. Moreover, these finding have important implications
for practitioners by providing indications on how audience members react differently to a
persuasive message as a function of their role in the communication. This suggests that audience
partitioning can be used as a strategic communicative technique in the context of interpersonal
selling and interpersonal persuasive communication more broadly. Imagine a salesman sitting
around the table with a team of corporate clients – how can he minimize the cognitive barriers
the key decision-maker will erect to his sales pitch? My work suggests addressing the message to
one audience member in particular (either verbally or with non-verbal communication) could be
a successful strategy to employ.
Importantly, I show my findings hold across different types of persuasion encounters,
both in the context of personal selling and in a different persuasion setting, where the goal of the
persuasive communication is not making a sale per se, but rather obtaining a signature for a
petition. Notably my studies were conducted in controlled lab settings, in a role-playing
situation, and with student subjects. Future research should examine the generalizability of my
findings to actual rather than imagined situations and to different subject populations.
66
The paradigm chosen for my laboratory studies compares the efficacy of a persuasive
message for bystanders with the efficacy of the same message for members of a non-partitioned
audience. In the case of a partitioned audience, I do not test the persuasive effectiveness of the
message for the addressed recipient. One interesting question therefore remains about the effect
of audience partitioning on addressed members of a partitioned audience. Notably, the
psychological mechanism uncovered in this work, namely that non-addressed recipients are less
skeptical about the intentions of the persuasion agent, suggests that my findings would extent to
the comparison between addressed and non-addressed recipients in a partitioned audience. What
my findings do not speak to, however, is whether addressed recipients would receive the
message any differently compared to members of a non-partitioned audience. This is an
interesting question to be investigated in future research. It is possible that being addressed
specifically, when other audience members are instead ignored from the persuasion agent, would
lead consumers to increase skepticism even further. This speaks to the importance of considering
the effects that audience partitioning can have on both addressed and non-addressed recipients
when choosing whether it is an appropriate strategy in the context of persuasive communication.
67
CHAPTER 4: CONCLUSIONS
4.1. Chapter Introduction
Consumers are social beings who dedicate a lot of time to communicating with one
another. Interpersonal communication can at times take the form of one-to-one communication,
such that the communicator’s target audience is a single individual. In other occasions, however,
communicators engage in one-to-many communication, i.e. their target audience is greater than
one. One-to-many communication presents distinct communication challenges (e.g., the
impossibility of tailoring the message to each recipient), but also opens the door for the
communicator to use new communication strategies to achieve his or her goals.
My dissertation introduces one such strategy, audience partitioning, which I define as the
decision by the sender of one-to-many communication to change the participation structure of his
or her audience by dividing recipients into two distinct groups: addressed and non-addressed
recipients. I explored how speakers can choose to partition their audience, and do so strategically
as a way of affecting how audience members react to the content shared. In particular, I focused
on both the anticipated and actual effect that audience partitioning has on non-addressed
recipients in an audience.
4.2. Main Findings and Implications
The effectiveness of interpersonal communication is in part dependent on the inferences
the audience makes about the speaker’s intentions. In many cases the more evident the intentions
of the speaker are to the target audience, the less effective the communication is at achieving the
communicator’s goal. For instance, the inference that a speaker is sharing positive information
about himself to make a good impression on others is often times seen as attention-seeking and
braggy, which ultimately may reduce overall impressions. Similarly, in the context of persuasive
communication, the more evident the persuasive intent of a speaker is, the more likely the
68
audience is to erect cognitive barriers and to resist the persuasion attempt.
In my dissertation, I identify one way by which the communicator can minimize the
negative inferences that an audience member makes about the communicator’s intentions,
namely by addressing the message to a different audience member, turning the intended recipient
of the message into a bystander. In Essay I study audience partitioning in the context of self-
promotion. I show that self-promoters–individuals sharing self-enhancing content–strategically
choose to address specific audience members thereby turning everyone else into bystanders on
the sideline. They do so because they expect non-addressed recipients (bystanders) perceive the
self-enhancing content as less attention-seeking and consequently form a better impression of
them. Audience partitioning is thus used as an intended solution to the self-promoter dilemma
and facilitates WOM in contexts where consumers worry that talking about their purchases and
consumption experiences may make them come across as overtly attention-seeking. Moreover, I
find that audience partitioning is, in most cases, a successful strategy. Bystanders do indeed form
more positive impressions of self-promoters who partition their audience and do so because they
see the self-promoter as less attention-seeking.
These findings contribute to the impression management literature by introducing a new self-
presentation strategy that consumers use in an attempt to solve the self-promoter dilemma.
Importantly, while the use of many previously studied strategies (e.g., humblebragging) has been
shown to be ineffective and may, at times, even backfire, audience partitioning generally seems
to be a successful strategy. Relatedly, a second important contribution to the impression
management literature comes from the studies that focused on the use of audience partitioning in
the context of social media communication. These studies add to our understanding of how
computer-mediated communication influences individuals’ self-presentation behavior in terms of
69
how users share information (Barasch and Berger 2014; Schau and Gilly 2003; Schlosser 2009).
Moreover, my research contributes to the literature on WOM and my findings have important
implications for practitioners by providing indications on how consumers might be encouraged
to generate word-of-mouth about purchases and consumption experience, as discussed in chapter
2.
Essay 2 shows that the strategic use of audience partitioning is not limited to the context
of self-presentation. I find that audience partitioning can also be used in the context of persuasive
communication, since it can be an effective strategy to persuade non-addressed recipients
(bystanders to a persuasive attempt directed at another audience member). This happens because
bystanders are less skeptical about the intentions of the persuasion agent compared to addresses
recipients, which translates into more positive attitudes towards the objective of persuasion.
These findings contribute to the literature on persuasion and social influence and have important
implications for practitioners by demonstrating that audience members react differently to a
persuasive message as a function of their role in the communication.
More broadly, my dissertation contributes to the literature on social cognition and
communication. Indeed, to the best of my knowledge, this work is the first to investigate how
and why individuals choose to partition audiences, putting a portion of their audience in the
position of perceiving themselves as bystanders exposed to communication not directed
specifically at them, as well as how partitioning affects how the audience responds to the
message being shared.
4.3. Suggestions for Future Research
I believe the most important question not addressed in this work concerns both the
expected and actual effect that audience partitioning has on addressed recipients who are singled
out in front of one or more bystanders. The addressee is an important party in the communication
70
whose role deserves further investigation. For instance, in the context of self-promotion, it would
be interesting to understand both the expected and actual effect of partitioning on the impression
that the addressee forms of the self-promoter. For instance, does the self-promoter hurt his/her
image in the eyes of the addressee by coming across to them as highly attention-seeking. If that
is the case, the communicator faces a trade-off between the self-presentational benefit of
partitioning with respect to the bystanders and the negative consequences with respect to the
addressee. As I have argued in chapter 2, whether this trade-off exists probably depends on who
the addressee is. If the addressee is a close other, this person might not alter his/her (already well
established) impression of the self-promoter based on a single act of self-promotion. Generally
speaking, whether to partition one’s audience or not is a deliberate choice of the source. I would
expect individuals to partition their audience only when the expected benefits (positive effects on
bystanders) seem to outweigh the expected costs (negative effects on addressee). Future research
could further investigate the conditions under which partitioning actually conveys a cost with
respect to its effects on the addressee.
Notably, in my studies respondents were given minimal information about the
relationship between speakers, addressees and bystanders. Yet, these relationships might play an
important role in driving both the choice of partitioning one’s audience as well as the effects that
partitioning has on both bystanders and addressees. For instance, partitioning might serve
another important function, other than that of making the source’s (self-promotional or
persuasive) intentions less evident to bystanders–signaling the speaker has a special relationship
with the addressee. This could be the case when the addressee is a member of an aspirational
group, or a highly influential individual. The relationship signaling value of audience partitioning
could have important consequences both in the context of self-promotion and in the context of
71
persuasive communication. Without doubts, partitioning is a multi-faceted, multi-determined
phenomenon and future research should further investigate additional drivers and consequences
of audience partitioning.
72
References
Aaker, J. L., Brumbaugh, A. M., & Grier, S. A. (2000). Nontarget markets and viewer
distinctiveness: The impact of target marketing on advertising attitudes. Journal of
Consumer Psychology, 9(3), 127-140.
Ahluwalia, R. (2000). Examination of psychological processes underlying resistance to
persuasion. Journal of Consumer Research, 27(2), 217-232.
Argo, J. J., Dahl D. W., & V. R. Manchanda (2005). The Influence of a Mere Social Presence in
a Retail Context. Journal of Consumer Research, 32(2), 207-212.
Arkin, R. M. (1981). Self-Presentation Styles. In Impression Management Theory and Social
Psychological Research, ed. J.T. Tedeschi, New York: Academic Press, 311–334.
Barasch, A., & Berger J. (2014). Broadcasting and Narrowcasting: How Audience Size Affects
What People Share. Journal of Marketing Research, 51(3), 286-299.
Bell, A. (1984). Language Style as Audience Design. Language in Society, 13(02), 145-204.
Berger, J. (2014). Word of Mouth and Interpersonal Communication: A Review and Directions
for Future Research. Journal of Consumer Psychology, 24(4), 586-607.
Berman, J. Z., Levine E. E., Barasch A., & Small D. A. (2015). The Braggart’s Dilemma: On the
Social Rewards and Penalties of Advertising Prosocial Behavior. Journal of Marketing
Research, 52(1), 90-104.
Bochner, A. P. (1984). The functions of human communication in interpersonal bonding. In C.
C. Arnold & J. W. Bowers (Eds.), Handbook of rhetorical and communication theory
(pp. 544-621). Boston: Allyn & Bacon.
Brehm, J. W. (1966). A theory of psychological reactance.
73
Brehm, S. S., & Brehm, J. W. (2013). Psychological reactance: A theory of freedom and control.
Academic Press.
Brock, T. C., and Becker, L. A. (1965). Ineffectiveness of ‘overheard’
counterpropaganda. Journal of Personality and Social Psychology, 2(5), 654.
Brown, J. D. (2007), The Self. New York: Psychology Press.
Campbell, M. C. (1995). When Attention-Getting Advertising Tactics Elicit Consumer
Inferences of Manipulative Intent: The Importance of Balancing Benefits and
Investments. Journal of Consumer Psychology, 4(3), 225-254.
Campbell, M. C., & Kirmani, A. (2000). Consumers' use of persuasion knowledge: The effects
of accessibility and cognitive capacity on perceptions of an influence agent. Journal of
consumer research, 27(1), 69-83.
Chaiken, S., Wood, W., & Eagly, A. H. (1996). Principles of persuasion.
Chung, C. M. Y., & Darke P. R. (2006). The Consumer as Advocate: Self-relevance, Culture,
and Word-of-Mouth. Marketing Letters, 17(4), 269-279.
Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R., Freeman, S., & Sloan, L. R. (1976).
Basking in reflected glory: Three (football) field studies. Journal of personality and
social psychology, 34(3), 366.
Clark, H. H., & Carlson T. B. (1982). Hearers and Speech Acts. Language, 332-373.
Clark, H. H., & Murphy G. L. (1982). Audience Design in Meaning and Reference. Advances in
Psychology, 9, 287-299.
Clark, H. H., & Schaefer E. F. (1987). Concealing One’s Meaning from Overhearers. Journal of
Memory and Language, 26(2), 209-225.
74
Clee, M. A., & Wicklund, R. A. (1980). Consumer behavior and psychological
reactance. Journal of Consumer Research, 6(4), 389-405.
Crano, W. D., Siegel, J. T., Alvaro, E. M., & Patel, N. M. (2007). Overcoming adolescents'
resistance to anti-inhalant appeals. Psychology of Addictive Behaviors, 21(4), 516.
Dayter, D. (2014). Self-Praise in Microblogging. Journal of Pragmatics, 61, 91-102.
De Angelis, M., Bonezzi, A, Peluso, A. M., Rucker, D. D., & Costabile M. (2012). On Braggarts
and Gossips: A Self-Enhancement Account of Word-of-Mouth Generation and
Transmission. Journal of Marketing Research, 49(4), 551-563.
Deshpandé, R., & Stayman, D. M. (1994). A tale of two cities: Distinctiveness theory and
advertising effectiveness. Journal of Marketing Research, 57-64.
DePaulo, B. M., Kenny, D. A., Hoover C. W., Webb W., & Oliver P. V. (1987). Accuracy of
Person Perception: Do People Know What Kinds of Impressions They Convey? Journal
of Personality and Social Psychology, 52(2), 303-315.
Eagly, A. H., & Chaiken, S. (1995). Attitude strength, attitude structure, and resistance to
change. Attitude strength: Antecedents and consequences, 4, 413-432.
Eastman, K. K. (1994). In The Eyes of the Beholder: An Attributional Approach to Ingratiation
and Organizational Citizenship Behavior. Academy of Management Journal, 37(5),
1379–1392.
Erdem, T and Swait J. (1998). Brand Equity as a Signaling Phenomenon. Journal of Consumer
Psychology, 7(2), 131-157.
Fenigstein, A., Scheier, M. F., & Buss A. H. (1975), Public and Private Self-Consciousness:
Assessment and Theory. Journal of Consulting and Clinical Psychology, 43(4), 522-
527.
75
Fiske, S. T. (2001). Social and Societal Pragmatism: Commentary on Augustinos, Gaskell, and
Lorenzi-Cioldi. In Representations of the Social: Bridging Research Traditions, ed. Kay
Deaux and Gina Philogene, New York: Blackwell, 249–53.
Fiske, S. T. (2009). Social beings: Core motives in social psychology. John Wiley & Sons.
Fitzsimons, G. J., & Lehmann, D. R. (2004). Reactance to recommendations: When unsolicited
advice yields contrary responses. Marketing Science, 23(1), 82-94.
Fitzsimons, G. J., Nunes, J. C., & Williams, P. (2007). License to sin: The liberating role of
reporting expectations. Journal of Consumer Research, 34(1), 22-31.
Forehand, M. R., & Grier, S. (2003). When is honesty the best policy? The effect of stated
company intent on consumer skepticism. Journal of consumer psychology, 13(3), 349-
356.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with
persuasion attempts. Journal of consumer research, 21(1), 1-31.
Giacalone, R. A., Rosenfeld, P. (1986). Self-Presentation and self-Promotion in an
Organizational Setting. The Journal of Social Psychology, 126(3), 321-326.
Godfrey, D. K., Jones, E. E., & Lord C. G. (1986). Self-Promotion is Not Ingratiating, Journal of
Personality and Social Psychology. 50(1), 106-115.
Goffman, E. (1959). The Presentation of Self in Everyday Life. New York: Anchor.
Goffman, E. (1981). Forms of Talk. University of Pennsylvania Press.
Goodman, J. K., Cryder, C. E., & Cheema, A. (2013). Data Collection in a Flat World: The
Strengths and Weaknesses of Mechanical Turk Samples. Journal of Behavioral Decision
Making, 26(3), 213-224.
76
Gosnell, C. L., Britt, T. W., & Mckibben E. S. (2011). Self-Presentation in Everyday Life:
Effort, Closeness, and Satisfaction. Self and Identity, 10(1), 18-31.
Greco, A. J. (1988). Representation of the elderly in advertising: crisis or inconsequence?
Journal of Services Marketing, 2(3), 27-34.
Han, Y. J., Nunes, J. C. & Drèze X. (2010), Signaling Status with Luxury Goods: The Role of
Brand Prominence, Journal of Marketing, 74(4), 15-30.
Hareli, S., & Weiner B. (2000). Accounts for Success as Determinants of Perceived Arrogance
and Modesty. Motivation and Emotion, 24 (3), 215-236.
Hogan, B. (2010). The Presentation of Self in the Age of Social Media: Distinguishing
Performances and Exhibitions Online. Bulletin of Science, Technology & Society, 30 (6),
377-386.
Holtgraves, T., & Srull T. K. (1989). The Effects of Positive Self-Descriptions on Impressions:
General Principles and Individual Differences. Personality and Social Psychology
Bulletin, 15(3), 452-462.
Isaac, M S., & Grayson, K. (2017). Beyond Skepticism: Can Accessing Persuasion Knowledge
Bolster Credibility? Journal of Consumer Research, 43(6), 895-912.
Jones, E. E., & Pittman T. S. (1982). Toward a General Theory of Strategic Self-Presentation.
Psychological Perspectives on the Self, 1, 231-262.
Karmarkar, U. R., & Tormala, Z. L. (2009). Believe me, I have no idea what I’m talking about:
The effects of source certainty on consumer involvement and persuasion. Journal of
Consumer Research, 36(6), 1033-1049.
Kellerman, K. and Cole T. (1994). Classifying Compliance Gaining Messages: Taxonomic
Disorder and Strategic Confusion. Communication Theory, 4 (1), 3–60.
77
Kirmani, A., & Zhu, R. (2007). Vigilant against manipulation: The effect of regulatory focus on
the use of persuasion knowledge. Journal of Marketing Research, 44(4), 688-701.
Kivetz, R. (2005). Promotion reactance: The role of effort-reward congruity. Journal of
consumer research, 31(4), 725-736.
Krull, J. L., MacKinnon D. P. (2001). Multilevel Modeling of Individual and Group Level
Mediated Effects. Multivariate Behavioral Research, 36(2), 249-277.
Lampel, J., & Bhalla A. (2007). The Role of Status Seeking in Online Communities: Giving the
Gift of Experience. Journal of Computer-Mediated Communication, 12(2), 434-455.
Latané, B., & Darley J. M. (1968). Group Inhibition of Bystander Intervention in Emergencies.
Journal of Personality and Social Psychology, 10(3), 215-221.
Latané, B., & Darley J. M. (1970). Social Determinants of Bystander Intervention in
Emergencies. Altruism and Helping Behavior, 13-27.
Latané, B., & Nida S. (1981). Ten Years of Research on Group Size and Helping. Psychological
Bulletin, 89(2), 308-324.
Leary, M. R., Nezlek, J. B., Downs, D., Radford-Davenport, J., Martin, J., & McMullen, A.
(1994). Self-Presentation in Everyday Interactions: Effects of Target Familiarity and
Gender Composition. Journal of Personality and Social Psychology, 67(4), 664.
Levine, R. V., & West L. J. (1976). Attraction as a Function of Boasting, Self-Apology, and
Credibility of an Actor. Psychological Reports, 38(3), 1243-1246.
Liang, K. Y., & Zeger S. L. Longitudinal Data Analysis Using Generalized Linear
Models. Biometrika, 42(1), 13-22.
78
Little, T. D., Card, N. A., Bovaird, J. A., Preacher, K. J., & Crandall C. S. (2007). Structural
Equation Modeling of Mediation and Moderation with Contextual Factors. Modeling
Contextual Effects in Longitudinal Studies, 1, 207-230.
Lovett, M. J., Peres,R. & Shachar R. (2013). On Brands and Word of Mouth. Journal of
Marketing Research, 50(4), 427-444.
Marder, B., Joinson, A., Shankar, A., & Thirlaway, K. (2016). Strength Matters: Self-
Presentation to the Strongest Audience Rather than Lowest Common Denominator
When Faced with Multiple Audiences in Social Network Sites. Computers in Human
Behavior, 61, 56-62.
Marwick, A. E., & boyd, d. (2011). I Tweet Honestly, I Tweet Passionately: Twitter Users,
Context Collapse, and the Imagined Audience. New Media and Society, 13(1), 114–133.
McGregor, G. (1986). Language for Hearers, 8, Pergamon Press.
Meyers-Levy, J. (1989). The influence of sex roles on judgment. Journal of Consumer Research,
14, 522–530.
Miller, L. C., Cooke, L., Tsang, J., & Morgan, F. (1992). Should I Brag? Nature and Impact of
Positive and Boastful Disclosures for Women and Men. Human Communication
Research, 18(3), 364-399.
Nadkarni, A., & Hofmann, S. G. (2012). Why Do People Use Facebook? Personality and
Individual Differences, 52(3), 243-249.
Nunes, J. C., & Drèze, X. (2006). The endowed progress effect: How artificial advancement
increases effort. Journal of Consumer Research, 32(4), 504-512.
O'Keefe, D. J. (2002). Persuasion: Theory and research (Vol. 2). Sage.
79
Pennebaker, J. W., Boyd, R. L., Jordan, K., & Blackburn, K. (2015). The Development and
Psychometric Properties of LIWC 2015.
Puntoni, S., Sweldens, S., & Tavassoli, N. T. (2011). Gender identity salience and perceived
vulnerability to breast cancer. Journal of Marketing Research, 48(3), 413-424.
Petty, R. E., & Cacioppo, J. T. (1986). Message elaboration versus peripheral cues.
In Communication and persuasion(pp. 141-172). Springer, New York, NY.
Petty, R. E., Harkins, S. G., Williams, K. D., & Latané, B. (1977). The Effects of Group Size on
Cognitive Effort and Evaluation. Personality and Social Psychology Bulletin, 3(4), 579-
582.
Pfeffer, J., Fong, C. T., Cialdini, R. B., & Portnoy R. R. (2006). Overcoming the Self-Promotion
Dilemma: Interpersonal Attraction and Extra Help as a Consequence of Who Sings
One’s Praises. Personality and Social Psychology Bulletin, 32 (10), 1362-1374.
Phillips, B. J., & McQuarrie, E. F. (2010). Narrative and persuasion in fashion
advertising. Journal of Consumer Research, 37(3), 368-392.
Powers, T. A., & Zuroff D. C. (1988). Interpersonal Consequences of Overt Self-Criticism: A
Comparison with Neutral and Self-Enhancing Presentations of Self. Journal of
Personality and Social Psychology, 54(6), 1054-1062.
Preacher, K. J., Zyphur, M. J., & Zhang Z. (2010). A General Multilevel SEM Framework for
Assessing Multilevel Mediation. Psychological Methods, 15(3), 209-233.
Reeder, G. D. (2009). Mindreading: Judgments About Intentionality and Motives in
Dispositional Inference. Psychological Inquiry, 20(1), 1-18.
80
Reizenstein, R. C. (1971). A dissonance approach to measuring the effectiveness of two personal
selling techniques through decision reversal. In Proceedings (pp. 176-180). American
Marketing Association, Chicago.
Rudman, L. A. (1998). Self-Promotion as a Risk Factor for Women: The Costs and Benefits of
Counterstereotypical Impression Management. Journal of Personality and Social
Psychology, 74(3), 629-645.
Rule, B. G., Bisanz, G. L., & Kohn, M. (1985). Anatomy of a persuasion schema: Targets,
goals, and strategies. Journal of Personality and Social Psychology, 48(5), 1127.
Russell, C. A. (2002). Investigating the effectiveness of product placements in television shows:
The role of modality and plot connection congruence on brand memory and
attitude. Journal of consumer research, 29(3), 306-318.
Schank, R. C., & Abelson, R. P. (1977). Scripts. Plans, Goals and Understanding.
Schau, H. J., & Gilly M. C. (2003). We Are What We Post? Self-presentation in Personal Web
Space. Journal of Consumer Research, 30(3), 385-404.
Schlenker, B, R., & Leary, M. R. (1982), Social Anxiety and Self-Presentation: A
Conceptualization Model. Psychological Bulletin, 92(3), 641-669.
Schlosser, A. E. (2009). The Effect of Computer-Mediated Communication on Conformity vs.
Nonconformity: An Impression Management Perspective. Journal of Consumer
Psychology, 19(3), 374-388.
Schwartz, S. H. (1992). Universals in the Content and Structure of Values: Theoretical Advances
and Empirical Tests in 20 Countries. Advances in Experimental Social Psychology, 25,
1-65.
81
Scopelliti, I. , Loewenstein, G., & Vosgerau J. (2015). You Call It ‘Self-Exuberance;’ I Call It
‘Bragging’ Miscalibrated Predictions of Emotional Responses to Self-
Promotion. Psychological Science, 26(6), 903-914.
Scopelliti, I., Loewenstein, G., & Vosgerau J. (2018). Bragging through an
Intermediary. Working paper.
Sekhon, T., Bickart, B., Trudel, T., & Fournier S. (2014). Being a Likable Braggart: How
Consumers Use Brand Mentions for Self-presentation on Social Media. In Consumer
Psychology in a Social Media World, ed. Claudiu Dimofte, Curtis Haugtvedt, and
Richard Yalch., Armonk, NY: M. E. Sharpe.
Sezer, O., Gino, F., & Norton, M. I. (2018). Humblebragging: A distinct—and ineffective—self-
presentation strategy. Journal of personality and social psychology, 114(1), 52.
Star, S. H. (1989). Marketing and its discontents. Harvard Business Review, (12), 148–154.
Tal-Or, N. (2010). Bragging in the Right Context: Impressions Formed of Self-Promoters Who
Create a Context for their Boasts. Social Influence, 5(1), 23-39.
Tepper, K. (1994). The role of labeling processes in elderly consumers’ responses to age
segmentation cues. Journal of Consumer Research, 20, 503–520.
Tice, D. M., Butler, J. L., Muraven, M. B., & Stillwell A. M. (1995). When Modesty Prevails:
Differential Favorability of Self-Presentation to Friends and Strangers. Journal of
Personality and Social Psychology, 69(6), 1120-1138.
Tucker, C. E. (2014). Social networks, personalized advertising, and privacy controls. Journal of
Marketing Research, 51(5), 546-562.
Walster, E., & Festinger, L. (1962). The effectiveness of overheard persuasive
communications. The Journal of Abnormal and Social Psychology, 65(6), 395.
82
Whittler, T. E. (1989). Viewers' processing of actor's race and message claims in advertising
stimuli. Psychology & Marketing, 6(4), 287-309.
Wicklund, R. A., Slattum, V., & Solomon, E. (1970). Effects of implied pressure toward
commitment on ratings of choice alternatives. Journal of Experimental Social
Psychology, 6(4), 449-457.
Williams, J. D., & Qualls, W. J. (1989). Middle-class black consumers and intensity of ethnic
identification. Psychology & Marketing, 6(4), 263-286.
Xu, A. J., & Wyer Jr, R. S. (2010). Puffery in advertisements: The effects of media context,
communication norms, and consumer knowledge. Journal of Consumer Research, 37(2),
329-343.
Zajonc, R. B. (1965). Social Facilitation. Research Center for Group Dynamics, Institute for
Social Research, University of Michigan.
83
APPENDIX A—Pre-test S1 and S2
DV: To what extent do you see this post as self-enhancing (written with the intention of trying to
impress others)? (1 = not at all, 9 = a great deal)
N = 885
Statements pre-tested = 38
Statements selected:
Post Study Self-Enhancing Rating Difference
Went to Spago today. Not too shabby
for a $200 dinner.
S1 7.56
p < .01
Went to Shakey’s today. Not too
shabby for a $20 dinner.
S1 4.44
Driving to work in my new Mercedes
S5
S2 8.53
p < .01
Driving to work in my old Toyota
Corolla
S2 3.32
Aced an IQ test today…I knew I was a
genius!! :)
S2 7.66
p < .01
Tanked an IQ test today…I thought I
was a genius!! :(
S2 3.61
84
APPENDIX B—Stimuli S1B and S5
SELF-ENHANCMENT CONDITION
Please think about something related to yourself that you think would impress and interest others
if they knew about it.
For example, one of the following:
- an ACHIEVEMENT such as getting straight As in school or winning a marathon;
- a TRAIT or SKILL you possess such as being good-looking or being a great dancer;
- a PRODUCT you got such as a brand new car or a stylish designer handbag.
CONTROL CONDITION
Please think about something related to yourself that you think others would find interesting and
relevant to them if they knew about it.
For example, one of the following:
- a thing you LEARNED such as something you read or heard and found important to know;
- something that HAPPENED TO YOU such as being stuck in a traffic jam on your way to
work because of roadwork;
- a SHOPPING EXPERIENCE you had such as trying a virtual reality headset or finding a
bacon flavored ice-cream joint.
85
APPENDIX C—Facebook Posts Used S2
86
APPENDIX D—IQ Test S4
- Which one of the five is least like the other four?
o Dog
o Mouse
o Lion
o Snake
o Elephant
CORRECT RESPONSES: 91.0%
- Which number should come next in the series? 1-1-2-3-5-8-13
o 8
o 13
o 21
o 26
o 31
CORRECT RESPONSES: 96.0%
- Which one of the five choices makes the best comparison?
PEACH is to HCAEP as 46251 is to:
o 25641
o 26451
o 12654
o 51462
o 15264
CORRECT RESPONSES: 97.5%
Feedback: You did great! You did better than 80% of the people who took this test!
While this is not diagnostic and you should conduct a full scale test to measure your IQ, these results
suggest your IQ should be around 160.
87
APPENDIX E—Twitter Posts Used S6
88
APPENDIX F—Additional Results S6
DV = Likelihood of Sharing
Independent Variable df(error) F Sig.
Audience Partitioned 1(278) 5.41 .02
Post 2(556) 86.69 .00
Audience Partitioned*Post 2(556) 2.69 .07
By Post: M
1
= 3.25, SD = 2.46, M
2
= 1.46, SD = 1.11, M
3
= 2.55, SD = 1.96, ωp² = .234.
DV = Impression
Independent Variable df(error) F Sig.
Audience Partitioned 1(278) 6.20 .01
Post 2(556) 332.78 .00
Audience Partitioned*Post 2(556) 1.34 .26
By Post: M
1
= 4.86, SD = 1.37, M
2
= 2.17, SD = 1.35, M
3
= 3.69, SD = 1.61, ωp² = .543.
DV = Attention Seeking
Independent Variable df(error) F Sig.
Audience Partitioned 1(278) 6.15 .01
Post 2(556) 70.29 .00
Audience Partitioned*Post 2(556) 1.73 .18
By Post: M
1
= 6.63, SD = 1.90, M
2
= 8.04, SD = 1.50, M
3
= 6.90, SD = 1.82, ωp² = .199.
89
GSEM Model Results
Coeff,
Std.
Error
z p
95% Conf.
Intervals
Likelihood of Sharing
Impression 0.69 .04 17.24 .00 0.61 0.77
Attention -
Seeking
-0.07 .03 -2.22 .03 -0.14 -0.01
Tag 0.15 .13 1.10 .27 -0.11 0.40
Post
Post 2 0.17 .16 1.10 .27 -0.14 0.48
Post 3 0.12 .13 0.96 .34 -0.13 0.37
Impression
Attention -
Seeking
-0.24 .03 -8.62 .00 -0.29 -0.18
Tag 0.22 .12 1.89 .06 -0.01 0.45
Post
Post 2 -2.35 .11 -21.82 .00 -2.56 -2.14
Post 3 -1.10 .10 -10.94 .00 -1.30 -0.90
Attention Seeking
Tag -0.37 .15 -2.49 .01 -0.66 -0.08
Post
Post 2 1.42 .13 11.18 .00 1.17 1.67
Post 3 0.28 .13 2.17 .03 0.03 0.52
A random effect at the respondent level is included in the analysis
90
APPENDIX G—Stimuli S8
91
APPENDIX H—Stimuli S2
[Speaker: Hey Sam!
Addressee: Hey Alex!]
Speaker: [Sam] Do you know that many professors believe that they are building moral fiber and
preparing students for the working world by denying them the opportunity to redo assignments
and exams? And these are the same professors who set a deadline for submitting work and then
give students who do not meet the deadline a zero, thinking that the devastating score will teach
them responsibility!
What they don’t recognize is that examination retake is a learning strategy that can be used on
any exam or assignment to improve learning!
This procedure involves allowing the student to redo some or all questions missed on the graded
examination. For the effect to be most beneficial, the original exam papers should
be graded and returned as soon as possible to ensure that material is still somewhat fresh in the
student’s mind. Students look at the completed, graded exam and learn from their mistakes by
redoing all missed problems, hence the term examination retake.
Of course this should result in better learning, don’t you think, [Sam]?
Do you know that 62% of USC students freely admit that they never review returned exams?
Consequently, they do not utilize, or even recognize, the returned exam as an opportunity to
learn from their mistakes! This can result in a failure to retain much of what has been studied
previously, or to the retention of incorrect material.
Abstract (if available)
Abstract
Interpersonal communication serves a variety of functions for individuals in everyday life including helping to maintain the desired image in the eyes of others, seeking to persuade others to see one’s point of view, or attempting to convince them to take a desired action. Communicators can use a variety of strategies to increase the effectiveness of their communication. In my dissertation, across 2 essays, I introduce one such strategy, audience partitioning, and investigate its antecedents and consequences. I define strategic audience partitioning as the decision by the sender of a one‐to‐many communication to change the participation structure of his or her audience by dividing recipients into two distinct groups: addressed and non‐addressed recipients. I focus on both the anticipated and actual effects that audience partitioning has on non‐addressed recipients in an audience, whom I refer to as bystanders. In essay 1, I focus on the use of audience partitioning when the goal of the communicator is to effectively manage one’s image in the eyes of others (i.e. impression management), while in essay 2, I investigate the use of audience partitioning in the context of persuasive communication.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Disclosure of changes in taste: implications for companies and consumers
PDF
Creation and influence of visual verbal communication: antecedents and consequences of photo-text similarity in consumer-generated communication
PDF
How do consumers use brands for identity signaling: Impact of brand prominence on consumers' choice and social interaction
PDF
Competing across and within platforms: antecedents and consequences of market entries by mobile app developers
PDF
Manipulating consumer opinion on social media using trolls and influencers
PDF
Marketing strategies with superior information on consumer preferences
PDF
Essays on the unintended consequences of digital platform designs
PDF
Essays on strategic categorization
PDF
The structure of strategic communication: theory, measurement, and effects
PDF
How does status influence behavior? The impact of achieved and endowed status on consumption patterns
PDF
Savoring future experiences: antecedents and effects on evaluations of consumption experiences
PDF
Technology, behavior tracking, and the future of work
PDF
The curse of loyalty: interdependent self construal and support for corruption
PDF
Slashing liquidity through asset purchases: evidence from collective bargaining
PDF
Essays in non-market strategy and entrepreneurship
PDF
How media multitasking during narrative viewing affects persuasion: the mediating roles of transportation, engagement, identification, enjoyment, and emotion
PDF
Experiential public relations: the importance of strategic messaging, understanding target audiences, and analysis of successfully curated brand-consumer interactions
PDF
Mobile self-tracking for health: validating predictors, effects, mediator, moderator, and social influence
PDF
How product designs and presentations influence consumers’ post-acquisition decisions
PDF
An academic and professional preparatory curriculum design and supplemental academic advisement tool: self-regulation, ethics, and communication for engineering graduate students
Asset Metadata
Creator
Valsesia, Francesca
(author)
Core Title
Strategic audience partitioning: antecedents and consequences
School
Marshall School of Business
Degree
Doctor of Philosophy
Degree Program
Business Administration
Publication Date
03/09/2020
Defense Date
03/09/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
audience partitioning,Communication,OAI-PMH Harvest,persuasion,self‐presentation
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Diehl, Kristin (
committee chair
), Nunes, Joseph (
committee chair
), Mayzlin, Dina (
committee member
), Schwarz, Norbert (
committee member
), Tully, Stephanie (
committee member
)
Creator Email
francescavalsesia@gmail.com,valsesia@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-484791
Unique identifier
UC11268347
Identifier
etd-ValsesiaFr-6093.pdf (filename),usctheses-c40-484791 (legacy record id)
Legacy Identifier
etd-ValsesiaFr-6093.pdf
Dmrecord
484791
Document Type
Dissertation
Rights
Valsesia, Francesca
Type
texts
Source
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 a...
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
Tags
audience partitioning
persuasion
self‐presentation