University of Southern California Dissertations and Theses

Innovation and good intentions: evaluations of three cross-sectoral programs for at-risk populations in Southern California

(USC Thesis Other)

Innovation and good intentions: evaluations of three cross-sectoral programs for at-risk populations in Southern California

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Innovation and good intentions:
Evaluations of three cross-sectoral programs for at-risk populations in Southern
California
by
Brittany Danielle Williams
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PUBLIC POLICY AND MANAGEMENT)
May 2021
Copyright 2021 Brittany Danielle Williams
ii
Acknowledgements
This dissertation has been the work of many years and would not have been completed
without the assistance of a number of people. My foremost thanks go to the members of my
dissertation committee. Gary Painter, the chair of my committee, served as my first and last
advisor and provided much insight, patience, and helpful guidance over the years. Nicole
Esparza contributed valuable feedback particularly on the interview guide for the organizational
analysis paper during its initial stages. Christine Beckman joined my committee at a pivotal
moment, for which I will always be grateful, and offered useful advice on how to enhance my
work.
Thank you to Raphael Bostic, who was my advisor for many years during this process
before he left USC and was both supportive of my work and provided me with numerous
opportunities to grow as a scholar. Thanks also to Johanna Lacoe, who was a model co-author.
Donnajean Ward and Aubrey Hicks were always reassuring presences on campus.
My classmates, particularly Sarah Axeen, Yusun Cho, Soledad De Gregorio, Jenneille
Hsu, Jovanna Rosen, and Lee White, were a source of incalculable assistance and friendship.
Special thanks to Hui Li for answering my questions regarding organizational analysis and to
Arthur Acolin for his insights and feedback on my work generally and the POPS paper
specifically. Vincent Reina was the best doctoral sibling a person could ask for.
There were also many individuals outside of USC who kept me physically and mentally
healthy during this process or whose labor allowed me to concentrate on my studies: Diane
Black, Angela Drown, Laura Miranda, both Veronica Mirandas, Cory Seuss, Marney Stofflet,
Angie Vroom, and especially Christina Dixon and Dr. Milica Stefanovic. I am incredibly
grateful to all of you for your help.
iii
There are several people in my life who have undertaken some version of a doctoral
journey of their own and as a result always knew when to ask about the dissertation, when not to
ask, and when to ask whether they should ask: Teresa Collard, Lisa LeBleu, James Sikes,
Heather Osborne-Thompson, Jim Thompson, Harvey Stark, Jen Wagelie, and Max Wagelie-
Stark. Thank you for your sometimes vocal/sometimes silent support and your general sense of
discretion.
I consider myself exceptionally lucky for the sheer number of family and friends I have
had encouraging me over the years, and I truly believe that I would not have completed this work
without them in my life. I’d particularly like to thank Andrew Boney, Harold Davis III, Neal
Goodwin, Katie Halliday, Wolasi Konu, Christy Madden, Erin Harper Mao, Ashley Middleton,
Chita Middleton, Jerry Passannante, Chloe Reisen, Dan Simon, David Carroll Simon, Sarah
McBride Thacker, Emily Vincent Vo, Kristin Wong, and my in-laws Candace Carroll and Len
Simon. My good friend Meha Priyadarshini was a source of much needed counsel, guidance,
and humor.
Finally, I’d like to thank my mother, Donna Williams, who was my first and best teacher
and who has been my life-long personal cheering section; my husband, Matt Simon, who
graciously endured this process alongside me and whose kindness, wit, and general presence
made it bearable; and my father, Royce Williams, whose generous heart was the source of so
much love, encouragement, and laughter for me throughout my life. He did not live to see this
dissertation completed, but he always believed I had it in me, even on days when I didn’t.
iv
Table of Contents
Acknowledgements …………………………………………………………………………….....ii
List of Tables ……………………………………………………………………………………..v
List of Figures ……………………………………………………………………………………vi
Abstract ………………………………………………………………………………………….vii
Introduction ……………………………………………………………………………………….1
Chapter 1: Museums as Classrooms: The Academic and Behavioral Impacts of “School in the
Park” ……………………………………………………………………………………………...6
Abstract …..……………………………………………………………………………….7
1. Introduction …...………………………….…………………………………………….8
2. Literature Review ……………………………………………..………………………10
3. Description of the Program …………………………………………………………...12
4. Research Questions ……...…………………………………………………................15
5. Research Design …………………………………………………………....................18
6. Results …………………………………………………………...................................25
7. Discussion and Policy Implications …………………………………………………..31
8. Figures and Tables …………………………………………………………................34
9. Appendices to Chapter 1 …………….…………………………………………..........42
Chapter 2: Prison from the Outside: Evaluating POPS the Club …………...…………………...50
Abstract …..……………………………………………………………………………...50
1. Introduction …...………………………….…………………………………………...51
2. Description of the Program …………………………………..……………………….53
3. Literature Review …………………………………………………………..................53
4. Theory ……...…………………………………………………....................................56
5. Research Design …………………………………………………………....................59
6. Results …………………………………………………………...................................63
7. Discussion and Policy Implications …………………………………………………..70
8. Tables …………………………………………………………....................................73
9. Appendices to Chapter 2 ………………………………………………………….......80
Chapter 3: We All Want the Same Thing: An Organizational Analysis of a Housing Reentry
Program ………………………………………………………………………………………….83
Abstract …..……………………………………………………………………………...83
1. Introduction …...………………………….…………………………………………...84
2. Literature Review …………………………………..…………………………………86
3. Pilot Program Background and Description …..……………………...........................88
4. Collaborative Governance ..………………………………….....................................95
5. Research Questions …………………………………………………...........................97
6. Data and Methodology ……………………………………………………..................98
v
7. Findings ………………………….………………………………………………….102
8. Discussion ……………………………………………………...................................125
9. Recommendations …………………………………………………………...............128
Conclusion ……………………………………………………………………………………..135
References ……………………………………………………………………………………...137
vi
List of Tables
Chapter 1
Table 1. Descriptive statistics, means for short-term outcomes by SITP participation status
(student-year observations), 2000/01-2011/12 ………………………………………………….35
Table 2. Descriptive statistics, overall means for long-term outcomes by SITP participation status
(single observation per student), 2000/01-2011/12 ……………………………………………..36
Table 3. Short-Term academic outcomes (1996-2012, omit 2000) ……………………………..37
Table 4. Short-term outcomes, stratified by 2nd-grade test pass or fail (1996-2012, omit 2000) 38
Table 5. Cumulative week analysis, short-term outcomes (1996-2012, omit 2000) ……………39
Table 6. Long-Term outcomes (1996-2012) ………………………………………….……..40-41
Chapter 2
Table 1A. Descriptive Statistics, Overall Means for POPS Participants and Non-POPS Students
during POPS program years (2012/13 - Fall 2014), Student-Year Observations ……………….73
Table 1B. Descriptive Statistics, Overall Means for POPS Participants and Non-POPS Students
prior to the implementation of POPS (before 2012/13), Student-Year Observations ……….....74
Table 2A. Outcomes associated with POPS participation (excluding the free or reduced price
meal variable) ……………………………………………………………………………………75
Table 2B. Outcomes associated with POPS participation (including the free or reduced price
meal variable) ……………………………………………………………………………………76
Table 3. Retention outcomes associated with POPS participation using various student fixed
effects based models …………………………………………………………………………….77
Table 4A. Outcomes associated with POPS participation, excluding the first year of POPS and
the free or reduced price meal control variable …………………………………………………78
Table 4B. Outcomes associated with POPS participation excluding the first year of POPS and
including the free or reduced price meal variable ……………………………………………….79
vii
List of Figures

Figure 1: Conceptual Model of SITP Impact ……………………………………………………34

viii
Abstract
Social innovation, which consists of both the process and the solutions that result from the
process itself, has the potential to address heretofore intractable social problems and result in
more equitable outcomes and fundamental system change. This dissertation evaluates three such
programs intended to assist different at-risk populations in Southern California. While these
populations vary as do the means of the intervention as well as the form of social innovation
typified by each, all three programs use cross-sectoral partnerships and attempt to leverage
existing resources in order to address the issue at hand. These evaluations are intended not only
to ascertain the effectiveness of the programs themselves but also to contribute to the literature
on how best to utilize existing resources in novel and replicable ways in order to improve the
lives of marginalized groups.
1
Introduction
Despite over two decades worth of scholarship, social innovation remains a somewhat
ambiguous term (Aoyama & Parthasarathy, 2016; Ayob, et al., 2016; Edwards-Schachter &
Wallace, 2017; Grimm et al., 2013; Mulgan, 2006; Pol & Ville, 2008; Tracey & Stott, 2016;
Ziegler, 2017). Broadly defined by Pol and Ville (2008) as “an innovation intended to improve
the quality or quantity of life (p. 4),” more precise definitions have emphasized the complexities
of the problems addressed (Grimm et al., 2013), the degree to which the innovation impacts
existing social structures (Ayob et al., 2016; Beckers et al., 2013; Grimm et al., 2013; Voorberg
et al., 2014; Westley, 2008), and the innovation’s specific impact on marginalized populations
(Ayob et al., 2016; Aoyama & Parthsarathy, 2016).
1
It’s also often, but not always,
conceptualized as including the process by which solutions are created in addition to the
solutions themselves (Ayob et al., 2016; Edwards-Schachter & Wallace, 2017; Grimm et al.,
2013; Voorberg et al., 2014; Ziegler, 2017).
Many of these definitions also include, either implicitly or explicitly, the notion of
collaboration (Aoyama & Parthsarathy, 2016; Ayob et al., 2016; Grimm et al., 2013; Edwards-
Schachter & Wallace, 2017; Ziegler, 2017.), and cross-sectoral collaborations have emerged as
an important tool used to implement these innovations (Domanski et al., 2020). While not all
cross-sectoral collaborations are social innovations and not all social innovations feature cross-
sectoral collaborations, cross-sectoral collaborations can be a conduit for innovation as they
allow organizations in different sectors to utilize resources, methods, etc. that would typically be
unavailable to them. They can successfully facilitate outcomes that organizations would not be
able to accomplish independently of one another (Bryson et al., 2006).
1
For additional definitions see also Edwards-Schachter & Wallace, 2017 and Pol & Ville, 2008 among others.
2
This dissertation examines three programs that use cross-sectoral collaborations as a
means of facilitating socially innovative solutions intended to assist specific at-risk populations.
While these populations differ in terms of their demographics and the specific challenges they
encounter, they are connected by the persistent disadvantage each group faces. In addition to
their status as cross-sectoral collaborations, the programs themselves are further connected by
their reliance on existing resources.
However, in the context of social innovation, these programs constitute three examples of
different forms of social innovation: an intervention by a non-profit foundation; the creation of a
program by a social entrepreneur; and a government agency-initiated collaboration. This
difference in the types of social innovation embodied by each of these programs permits an
exploration of the usefulness of each under a given set of circumstances.
By studying these programs, I aim not only to evaluate the effectiveness of specific
programs in aiding their respective populations but also to contribute to the literature on how
best to creatively leverage existing resources in socially innovative ways that can be replicated
and scaled up on behalf of groups who face deeply entrenched disadvantages.
The first study examines a museum-based educational program for low-income public
school students and its impact on students’ academic and behavioral outcomes in both the short
and long-term. The program, School in the Park (SITP), was originally created to help alleviate
school overcrowding by utilizing the museums located in Balboa Park in San Diego. The
program re-purposes museum spaces as classrooms and leverages existing museum collections
and exhibits to provide experiential learning opportunities. There the elementary school students
receive instruction from both museum educators and their own teachers. However, concerns
about the program arose in recent years: namely whether the benefits of participating in the
3
program justified the amount of time spent outside of the classroom as the students who
participate spend up to 25 percent of their school instruction time at the museums (Pumpian et
al., 2005).
Using longitudinal, student level-data and a difference-in-differences framework for the
short-term outcomes and ordinary least squares and linear probability models for the long-term
outcomes, this study evaluates the academic and behavioral outcomes associated with School in
the Park attendance both in the year of participation and during the students’ high school years.
The results indicate that participation does not occur at the expense of scholastic achievement.
Students experience small short-term academic gains as well as behavioral improvements during
the years they participate in SITP, though the effects dissipate by the time the students reach high
school.
These findings are particularly useful given broader concerns that experiential learning
programs may waste precious learning time that would be better spent in a traditional classroom
following curricula specifically intended to prepare students for standardized tests. While a few
studies have found positive associations between participation in similar programs and student
outcomes, they generally lacked rigor and were limited to short-term outcomes. This paper with
its rich level of data and examination of long-term outcomes, fills this gap.
The second paper in this dissertation consists of an evaluation of the Pain of the Prison
System (POPS) the Club, an innovative program established at Venice High School to help
students facing the challenges associated with engagement with the prison system. Despite the
negative outcomes associated with parental incarceration, few programs exist that specifically
aim to assist minors who are struggling with experiences of familial incarceration and even fewer
exist that take place in a school context. POPS was formed with the intention of providing
4
support to students who have personal experiences with the criminal justice system or who have
currently or formerly incarcerated loved ones. It utilizes classroom space as a venue for the
students to safely explore and share their feelings about incarceration, and they are also
encouraged to express themselves creatively through poetry and other means. This study
contributes to the nascent evaluation literature on these programs by examining the short-term
academic and behavioral outcomes associated with students’ participation in POPS. It is one of
only two known evaluations of non-mentor, school-based programs that cater to this population
(Springer et al., 2000).
A student-level fixed effects model is used to estimate outcomes. While the results do
not support the idea that participating in POPS results in better student outcomes, there is also no
evidence that suggests that it is associated with negative outcomes. This is an important finding
given that educators may be reluctant to permit such programs due to the stigma associated with
incarceration. They may perceive students with family histories of incarceration as being at a
higher risk of behavior problems or other negative outcomes and worry that interaction with
similarly situated peers could compound these effects. As a result, school administrators may be
inclined to discourage the formation of POPS clubs or other similar school-based programs that
focus on this subset of the student population. This study provides preliminary reassurance that,
at least in certain contexts, this is not the case. Furthermore, given that POPS students are shown
to underperform on all academic measures in the years prior to their participation in POPS
compared to students who ultimately never participate in POPS, this study also provides
supporting evidence for systematic gaps in the educational outcomes between students who
experience familial incarceration and those who don’t. An outcome which is often assumed but
for which only limited evidence exists in the literature.
5
The final paper consists of an organizational analysis of a partnership between the
Housing Authority of the City of Los Angeles (HACLA) and three local nonprofits in their
implementation of a housing reentry pilot program. The reentry pilot program was created to
take advantage of existing housing by allowing recently released formerly incarcerated
individuals to reside with family members receiving Section 8 housing assistance, which they
were usually prohibited from doing. The housing authority partnered with the three local non-
profits that were responsible both for partially vetting the potential participants and for providing
supportive services for them. Given the housing challenges faced by many recently released
individuals, this program was expected to feel a pressing need. However, despite the
organizations’ continuing efforts, only seven families were reunited over the course of almost
three years. Furthermore, comparable programs instituted by other housing authorities across the
country demonstrated similar low participation rates.
This study examines to what extent the responsibility for this lack of participation lies
with the relationships between the organizations verses issues external to the organizations
themselves, such as the family dynamics amongst the potential participants, a distrust of the
housing authority, etc. Using semi-structured interviews and techniques based on the grounded
theory method, the implementation of the program is examined in the context of Emerson,
Nabatchi, and Balogh’s (2012) collaborative governance framework. While the partnerships
were in fact plagued by interorganizational issues, ultimately it seems that the avoidance of
surveilling institutions and a particular distrust of the housing authority may have been more of a
factor in the low participation numbers.
The dissertation concludes with a brief summation of these papers’ places within the
existing literature.
6
Chapter 1
Museums as Classrooms:
The Academic and Behavioral Impacts of “School in the Park”

Johanna Lacoe
University of California, Berkeley
Gary D. Painter
Sol Price School of Public Policy,
University of Southern California
Danielle Williams
Sol Price School of Public Policy,
University of Southern California

7
Abstract
Access to community cultural institutions, and theaters has the potential to improve students’
educational experiences. This article estimates the impact of School in the Park, a museum-based
educational program for low-income students that takes place within the cultural institutions and
museums of San Diego’s Balboa Park. We identify the impact of participation in the program on
short- and long-term academic and behavioral outcomes using longitudinal, student-level data
since 1996; natural variation in the timing of program implementation at two elementary schools;
and control groups of students from schools that did not receive the program. Findings indicate
that participation in the program has positive short-term impacts, but program impacts were
insignificant in the long run.

8

1. Introduction
Education policy is in the midst of an era of data-driven decision-making, accountability,
and high stakes testing. However, not all educators, policymakers, and parents agree about the
efficacy of this approach to improving student achievement. While the benefits of high stakes
testing include student achievement measures that are comparable across schools and over time,
metrics to gauge teacher effectiveness and to hold schools accountable, and indicators to identify
students who are struggling, there are also multiple unintended costs of the high stakes testing
regime (Jones et al., 2003). Concerns about “teaching to the test,” student stress, and a narrow
focus on the tested material at the expense of other subjects, such as art or music, raise the
question – does the quest to quantify achievement take away from a more holistic educational
experience for students?
Experiential education programs offer students opportunities to learn content that may
not be targeted at increasing standardized test scores in the short term. As a result, districts,
principals and teachers face increasing pressure to limit non-standard experiences in addition to
non-academic activities for students because they may take away from test preparation or do not
highlight material that will be tested come year’s end (Bassok et al., 2016; Booher-Jennings,
2005; U.S. Government Accounting Office, 2009). However, it remains an open question
whether programs not intentionally aimed at improving test scores actually have academic
benefits as measured by test scores. Experiential education programs may stimulate creativity
and curiosity, and promote longer-term positive outcomes for students such as the desire to
attend college or interest in a broader set of topics. In addition, No Child Left Behind and the
test-based education policies that have defined the recent decade and a half of education
policymaking, at least on a national level, do little to address the needs of disadvantaged students
9
and or ameliorate the achievement gap between these students and their more advantaged peers
(Ladd, 2012). Experiential education programs may provide students who struggle in the
classroom context or disadvantaged students an alternative learning environment, and a cultural
fluency often only available to middle- and upper-income students, which can translate into
opportunities to excel. Programs that occur off the school campus often leverage existing
community cultural institutions such as museums, zoos, and theaters.
2
Despite the potential of
these programs to improve the educational experience for students, particularly low-income
students, and bolster academic achievement, the lack of rigorous research about experiential
education makes it difficult for programs to withstand the pressures of the current testing-
focused environment.
Although evaluations of some experiential education programs have found positive
associations between program participation and a range of student outcomes in the short run,
there is little rigorous evidence to support the assumption that these programs impact student
achievement. Further, the previous evaluations tell us little about the long-term outcomes of
program participants. Out-of-school interventions may make school administrators wary when
schools are evaluated on test score gains, with curricula and the school day carefully structured
to meet these standards. This paper fills the gap in the literature about whether and how
experiential education programs affect academic outcomes by evaluating the impact of School in
the Park (SITP), a museum-based educational program in San Diego, on both short and long-
term student outcomes. As discussed below, it is rare to have longitudinal data on the impact of

2
While we know of no systematic overview of the costs incurred by schools for these types of programs, there is
anecdotal evidence in the literature that multi-visit programs and even single visit programs that are highly
structured tend to cover, at minimum, the cost of admission to the institution itself as well as all curriculum materials
provided (see Randi Korn and Associates, 2004; 2010). Others include transportation costs (National Gallery of Art,
2019). Crystal Bridge’s School Visit Program in addition covered the cost of substitute teachers and the students’
lunches. (Bowen, Greene, and Kisida, 2014).
10
an experiential education program. We test for the benefits of SITP using a standard difference-
in-differences framework that takes advantage of the staggered implementation of the program
across two elementary schools.
2. Literature Review
The body of research on arts and science programming that happens outside the
classroom is limited and very few studies focus specifically on museum-based education. A
meta-analysis of research on the impact of arts programs on academic achievement found only
31 studies to analyze, only one of which was published in a peer reviewed journal (Winner &
Cooper, 2000). Among the studies that have attempted to assess the impact of arts learning, there
is some indication that high quality arts enrichment programs may increase the school readiness
and educational attainment of students in these programs (Brown et al., 2010). However, these
studies suffer from methodological limitations that hamper the authors’ ability to isolate program
impacts and limit the generalizability of the results. Specifically, most programs operate on a
volunteer basis making it difficult for researchers to control for selection bias introduced when
some students select to participate and others do not. SITP presents a unique opportunity to avoid
this common problem. Within the two schools where SITP is offered, all students in the grades
served participate in the program; therefore, selection bias is less of a concern.
Recent work by Kisida et al. (2016; Bowen et al., 2014) significantly advances the
museum education literature by providing experimental evidence of the impact of single museum
visits on student critical thinking outcomes. Using stratified random assignment of school groups
to museum visits, the authors are able to convincingly isolate the causal impact of a single visit
to an art museum on short-term measures of student ability to critically examine a work of art.
The group also finds that the museum visits had larger impacts on a range of outcomes including
historical empathy, tolerance, and interest in art museums, for students from rural and high-
11
poverty schools, compared to other students (Kisida et al., 2014; Kisida et al., 2016). Although
these studies have greater internal validity than most of the other empirical work on museum
education, the treatment is only a single visit, and the control group visits the museum shortly
after the treatment group, so the authors are unable to maintain randomization to measure longer-
term outcomes. However, the findings suggest that even a single museum trip can affect how
students process information and think critically – skills that may be correlated with academic
achievement, both in the short and long term.
Given the emphasis on test outcomes, museums, and school educators have made efforts
to expand museum-based experiences beyond one-off guided tours and link these experiences
more closely to school curricula. Thus far, only a handful of studies have evaluated the impacts
of museum enrichment programs for students. These programs vary significantly in terms of
their methods and duration. While some involve multiple visits to a museum, the Hands-On
Museum’s program in Ann Arbor takes place exclusively in the students’ school, and programs
at the Museum of the City of New York and the Peabody Museum of Natural History consist
only of one-time field trips. A science program based out of a Los Angeles County natural
history museum only involves a single visit to the museum itself, though the program includes
eight additional after-school sessions led by museum educators. Similarly, although the
Guggenheim’s program in New York City includes three museum visits, it is predominantly
based in the school with twenty 90 minute sessions led by an artist in residence in the classroom
during the course of the school year. Finally, the Urban Advantage program, evaluated by
Weinstein et al. (2014), provides intensive museum-based teacher professional development for
middle school science teachers in New York City. They find that the program improves student
achievement on science exams in middle school, but the effects do not extend to high school.
12
The outcomes measured by evaluations of these programs vary substantially. Some
studies measure gains in student knowledge of specific content areas, while others focus on
improvements in student attitudes towards subjects. Museum enrichment programs, including
those that take place mainly or exclusively in classrooms via partnerships with museums, are
associated with improved student attitudes towards the subject at hand (Melber, 2003; Paris et
al., 1998; Randi Korn & Associates, Inc., 2010). They are associated with improvements in
students’ content knowledge (or their perception of their content knowledge) of the subject
matter (Melber, 2003; Paris et al., 1998; Randi Korn & Associates, 2010) though some single
museum visit programs with no follow-up instruction either in a museum or a classroom see only
modest effects (Cox-Petersen et al., 2003). Some evaluations report improved student problem
solving skills after participation (Paris et al., 1998). In most of the evaluations that utilize control
groups of similar students or randomly assign the treatment, participating students scored higher
on critical thinking evaluations than peers who did not participate (Burchenel & Grohe, 2007;
Downey et al., 2007; Bowen et al., 2014; Kisida et al., 2016). One study measures achievement
using standardized test scores and finds no statistically significant difference between
participants and nonparticipants (Adams et al., 2006). Many of these studies are cross sectional
and only report short-term outcomes. A key advantage of the current study is the ability to
provide information on longer-term outcomes.
3. Description of the Program
School in the Park is an educational program designed for third, fourth, and fifth grade
students held in the cultural institutions and museums of San Diego’s Balboa Park. Participating
institutions include the San Diego Museum of Art, the History Center, the Natural History
Museum, the Museum of Man, the Junior Theater and the Old Globe Theater, the Reuben H.
Fleet Science Center, the Air and Space Museum, the Museum of Photographic Arts, and the San
13
Diego Zoo.
3
The program’s stated goal is for students to experience “visual, auditory, and
kinesthetic information” and to provide students with a “foundation of knowledge and a context
in which to place new material” as they progress in school (SITP, n.d.). The program has
developed an explicit focus on “academic excellence” aimed at helping students excel in school
and in the future. To this end, the program’s curriculum initially combined topics appropriate for
the lessons at each particular institution with state standards and are now aligned with Common
Core and Next Generation Science Standards (Feldman et al., 2010; Higdon et al., 2017; Higdon
et al., 2018). Students visit a combination of the institutions specific to School in the Park’s
curriculum for that particular grade level. They do not visit all the institutions every year.
Students spend a week at each institution (Higdon et al., 2018), and up to 25 percent of their
instructional year away from their home school (Pumpian et al., 2005).
4
At the institution,
museum educators lead instruction with classroom teachers assisting with various learning
activities (SITP, n.d.). On a typical day, students arrive at the park at 8:40 a.m. following a 20-
minute bus ride that is used for instructional time (Pumpian et al., 2005). Museum educators then
lead the first two hours of instruction using exhibit areas and classroom space within the
museums. The classroom teacher leads the instructional periods before and after lunch, and
students return to their school by bus at 1:40 p.m. Examples of programming include: a lesson on
geometric shapes at the Museum of Modern Art, an exploration of the Renaissance period
through painting, costumes, and dramatic play at the Museum of Art, lessons on math and
science through an exploration of ancient Egypt and the human skeletal system by measuring

3
Participating institutions by grade level 3rd Grade: Historical Society, San Diego Zoo, Museum of Art. 4th Grade:
Museum of Photographic Arts, Museum of Man, Natural History Museum, Junior Theatre, Fleet Science Center. 5th
Grade: Fleet Science Center, Air & Space Museum, Hall of Champions, Natural History Museum.
4
There are a few exceptions where students may spend two weeks at one institution. Currently third graders spend
three weeks at the San Diego Zoo (Higdon, et al., 2018). See Appendix A for further details on the number of
program weeks for each grade over the years.
14
mummy bones at the Museum of Man, and a unit on air travel and aerodynamics through
studying life-size models of spacecraft at the Aerospace Museum and designing and test-flying
their own gliders (Pumpian et al., 2005). Each curricular unit is mapped to state learning
standards and goals. The program’s stated mission is to “cultivat[e] curiosity, competence,
confidence, and character” among students through authentic learning activities that align with
state standards.
During the period of study, SITP served students in two inner-city public elementary
schools, Rosa Parks and Alexander Hamilton Elementary, who participated in the activities at
Balboa Park for up to eight weeks (see Appendix A). Both schools are located in the City
Heights neighborhood of San Diego, which is a major refugee portal for families from Somalia,
Cambodia, Vietnam, Iraq, and Liberia, among other counties, and a destination for immigrants
from Latin America. In addition to serving a large refugee and immigrant student population,
many of the students who attend these two schools are low-income. In the San Diego Unified
School District, 59.4 percent of school children qualify for free or reduced price lunch (San
Diego Unified School District, n.d.), while at the two participating elementary schools over 90
percent of the students qualify (California Department of Education, 2016).
5

SITP was initially developed in response to overcrowding at Rosa Parks in the late 1990s.
The lack of space for instruction prompted administrators to think about how to utilize other
community spaces to serve all of the City Heights students. Over time, the program has evolved,
hiring full-time staff and museum educators to partner with teachers, expanding to Alexander
Hamilton Elementary, and developing curricula for the program that align with district academic
standards. Today, district administrators are questioning the ongoing utility of the program. Is it

5
As of the 2016/2017 school year, 96.4 percent of students at Rosa Parks and 96.2 percent of students at Hamilton
received either free or reduced lunch (California Department of Education, May 2017).
15
best for students to spend valuable class time outside the classroom, when they could be
practicing skills that will appear on the annual standardized exams? Does the program cause
struggling students to fall further behind? Is the district getting the most utility out of teachers
who are paid even when museum educators are providing instruction? These questions can only
be answered with a systematic evaluation of the program.
4. Research Questions
This paper evaluates the impact of SITP participation on short-term and long-term
academic and behavioral outcomes using longitudinal student-level data, the natural variation in
the timing of program implementation between the two schools, and control groups of students
from nearby schools that did not receive the program. The study addresses two primary research
questions:
1. How does participation in the School in the Park program impact academic and
behavioral outcomes for students in the years of participation?
2. Do impacts of School in the Park participation extend beyond the year of participation to
affect longer-term outcomes in high school?
Theory
Experiential education programs like SITP aim to utilize non-traditional strategies to
promote learning and to broaden students’ understanding of the world. A conceptual paper by the
American Alliance of Museums describes museum-based educational programs as providing
“vital, experiential, multi-modal and trans-disciplinary educational opportunities…more than the
ancillary field trip” (Kratz & Merritt, 2012). The creators of the School in the Park model built
directly on the main tenants of educational theory (Pumpian et al., 2005). They incorporated
elements of Constructivism, which holds that learning is a social process requiring students to
bring their own personal experience to bear as they “construct” new meaning (Dewey, 1916).
16
Key elements of this theory are inquiry-based education and educational experiences that take
place outside of the traditional classroom. A model of the impact of museum experiences on
students influenced by Dewey is structured as an educational cycle where students bring
previous knowledge from their lives to the initial museum experience, the museum triggers
reflection and inquiry, generates new problems and interests, and opens new lines of inquiry, and
these new experiences and interests are applied to the student’s life beyond the museum (Hein,
2004). As stated by Gardner (1991), museums have “the potential to engage students, to teach
them, to stimulate their understanding, and most important, to help them assume responsibility
for their own future learning” (p.202). Further, School in the Park aims to have long-lasting
effects through building students’ cultural capital (Pumpian et al., 2005). Differences in student
achievement between demographic groups may be explained in part by differences in prior
knowledge they bring to learning experiences (Marzano, 2004). By bridging formal and informal
learning (Eshach, 2007), School in the Park aimed to provide structured, scaffolded learning
experiences for students outside of the classroom environment, that nonetheless help them meet
rigorous academic standards in the short run, and have long-lasting effects over their lifetimes.
Building from these ideas, the conceptual model in Figure 1 presents the hypothesized
impact of participation in SITP on short, medium, and long-term student outcomes. Inputs into
the program include the characteristics and prior experiences of students and teachers, and the
school context in which students and teachers generally operate. Exposure to different cultures,
ideas, and fields through time in the Balboa Park institutions, and the curricular units designed to
complement these experiences, comprise the outputs of the program. These activities are
hypothesized to affect students in the short, medium, and long-term. Specifically, in the year of
participation in SITP, the process described by Dewey above may result in improved
17
engagement both during SITP and regular class instruction. Students may also attend school
more frequently because they want to participate in the program. As a result, students
participating in the program may achieve higher scores on standardized exams in that year, and
may be less likely to be retained, suspended, or expelled from school.
Alternatively, we may expect a negative or no immediate impact of participation. If the
museum education program affects cognitive and behavioral skills that develop over time, the
experience may not translate into higher test scores immediately following participation in the
program. These types of impacts may only appear over time, as the elementary school
participants grow and reflect on those experiences, and have other opportunities to apply new
skills. If this is the case, we may expect to see no impact on short-term outcomes.
At the same time, it is possible that participation in the program may have negative
effects on achievement for students who are already struggling academically. These students
may be better served by spending more time in the classroom and receiving remedial instruction,
rather than attending a multi-week unit at Balboa Park. For all of these plausible reasons, the
expected direction of the impact of the program on short-term outcomes is ambiguous.
A hope of a program such as SITP is that the benefits of exposure to new concepts and
cultures may extend beyond the year of participation if students incorporate these new
perspectives into their lives. These “broader horizons” might change students’ relationships to
museums (Randi Korn and Associates, 2004), culture (Kisida et al., 2014) as well as how they
envision their future options, including college attendance and possible careers. One way this
may be observable is if the impact of participation in SITP lasts into the high school years. In
Figure 1, immediate improvements in engagement, achievement, and behavior are hypothesized
to influence academic and behavioral outcomes in middle school and high school and college
18
preparation and enrollment. Even if the program has limited short-term outcomes, the real
benefit of participation may appear later on, as students use the new motivation and skills learned
in the program to persevere through high school and set higher post-graduation goals.
5. Research Design
To answer the question of how participation in the School in the Park program impacts
the academic and behavioral outcomes of students, a standard difference-in-differences
framework is employed (see Cannon et al., 2011 for an application of this technique) that takes
advantage of the fact that the program was implemented in the two schools at different times. In
essence, the research design compares changes in outcomes for students who participate in SITP
(before and after participation in the program), to changes in outcomes for students who do not
participate. The rich data from the San Diego Unified School District (SDUSD) allows the
development of better control groups than past studies and to measure short-term as well as long-
term outcomes like graduating from high school.
Equation 1 models short-term academic and behavioral outcomes as a function of
participation in SITP, time-varying student characteristics, student fixed effects, and school fixed
effects, using individual-level panel data.
Yisgt = α + β SITP sgt + γ STit + δ SC st + ηi + μs + ζt +wg+ εisgt [1]
In Equation 1, Yisgt represents the dependent variable for student i in grade g and school s at time
t; SITPsgt is a dichotomous variable indicating whether the child attended the School in the Park
program; STit is a vector of time-varying student characteristics; and SCst is a vector of school
characteristics for the school the student attended. In addition, the model includes a student fixed
effect (ηi), school fixed effect (μs), a time fixed effect (ζt), a grade fixed effect (wg), and an error
term (εisgt). The inclusion of the school and time fixed effects allow us to avoid confounding the
impact of SITP with other changes that may be happening in the schools at a particular point in
19
time. Further, the student fixed effect control for all time invariant characteristics of the student,
such as underlying motivation, parental support in the home, or other factors that may affect
achievement.
For the long-term outcomes, ordinary least squares and linear probability models are
estimated, as shown in Equation 2 below:
Yisg = α + β SITP sg + γ STi + δ HSs + μs + ζt + εisg [2]
The structure of the dataset used in Equation 2 is cross-sectional, with one observation per
student. In this equation, Yisg represents the dependent variable for student i in school s in grade
g; SITPsg is a dichotomous variable indicating whether the child ever attended the School in the
Park program in third, fourth, or fifth grade; STi is a vector of student characteristics (including
second grade achievement); and HSs is a vector of dummy variables for the six high school
programs which draw upon the elementary schools in our study sample.
6
The remainder of the
students are scattered in other high schools throughout the San Diego Unified School District. In
addition, the model includes a school fixed effect (μs) for the elementary school attended, a series
of dummy variables for the year the student was in third grade (ζt), and an error term (εisg). The
controls for the third grade year capture any cohort effects that may affect later outcomes in high
school, such as changes in the quality of SITP over time. Standard errors are clustered at the
elementary school level.

6
The data for this analysis include all of the students who attended elementary schools in the Crawford-Hoover
feeder system from 1996-2012. As a result, once these students get to high school, most attend the high schools in
the same system, but some move to different San Diego high schools. Because we only have information on students
who start their education in the Crawford-Hoover feeder system, the long-term cross-sectional data set has some
high schools with very few observations. Therefore including high school fixed effects for all of the high schools
attended does not result in robust estimates of the impact of the program. Instead, we maintain the fixed effects
based on the elementary school attended (effectively comparing students to peers who attended their elementary
school), and we control for the time-invariant characteristics of the high schools that serve the largest number of
students in our sample. Standard errors are clustered at the elementary school level as well, due to the same problem
with small sample sizes for some high schools.
20
Identification of the effects on student achievement and behavioral outcomes comes from
three primary sources. First, we compare students within Rosa Parks and Alexander Hamilton
Elementary schools that did not receive the program, with students from those same schools that
received the program in later years (within-treatment cohort comparison). Second, we compare
students in Rosa Parks to students in Alexander Hamilton Elementary over the period in which
students in Rosa Parks received the program and Alexander Hamilton did not (within-treatment
school comparison). This second comparison is particularly important because both schools may
have received some resources from other Price Charities investments in the community that
could also be impacting achievement independent of the School in the Park program. Finally, we
compare students in these schools to students in comparable schools in SDUSD that do not
receive this enrichment program (i.e., other schools in the Hoover High School system, and the
schools in the Crawford High School system) (control school comparison). A remaining
necessary assumption for the identification of a causal impact of SITP is that parents with
students who may be stronger do not select into Rosa Parks or Alexander Hamilton elementary
simply because the program exists at the schools.
7

7
In general, students appear to be more at-risk for academic challenges following the implementation of SITP, and
not the reverse. There are slight differences in the average student characteristics before and after the
implementation of SITP. The student population at Rosa Parks comprised more black and Asian students (9 percent
and 14 percent, respectively) and fewer Hispanics (74 percent) before SITP, compared to after (6 percent black, 10
percent Asian, 82 percent Hispanic). In the years following SITP implementation, a larger share of students qualified
for special education increasing from 6 percent to 13 percent of the student population. The exception being the
students’ second grade ELA and math test scores which were higher among students receiving SITP. However,
among Hamilton students the reverse was true: the second grade ELA and math test scores were lower among
students who received SITP. The characteristics of Hamilton students before and after the implementation of SITP
were similar to those at Rosa Parks. The Hispanic student population increased from 65 percent to 77 percent while
the number of black and Asian students decreased from 15 to 9 percent and from 16 to 10 percent respectively. The
share of students in special education increases from 9 to 13 percent, and the share of students receiving English
language support in a given year increases slightly from 55 percent to 58 percent. Similarly, when comparing the
student population in the control group schools before and after the implementation of SITP at Rosa Parks and again
before and after the implementation of SITP at Hamilton, the same changes in student characteristics occur. Before
SITP is implemented at Rosa Parks there are more black and Asian students (23 percent and 18 percent respectively)
and fewer Hispanic students (45 percent) than there after the implementation of SITP at Hamilton (14 percent black,
14 percent Asian, 64 percent Hispanic. There isn’t, however, a corresponding increase in the number of special
21
Data
We utilize administrative student records from the SDUSD, which includes student-level
demographic information, behavioral records, and measures of academic performance. We also
have school-level measures of school size and type, and student poverty.
The impact of participation in SITP on individual students is modeled over time,
controlling for student demographic characteristics as well as characteristics of their peers,
classrooms, and schools. To do so, we leverage student-level academic records from 1996 to
2012 for students in the Hoover High School feeder system (which includes the treatment group)
and the schools in the adjacent Crawford High School system (to serve as a control group). All
students who stay within the SDUSD are followed through high school, regardless of the schools
they attend following elementary school. Further, comparing students in the schools that receive
the program to their school peers enhances our ability to distinguish program impacts from
characteristics of schools. As expected, not all students that begin in our study schools will
conclude their education in SDUSD. The issue of attrition will only confound the analysis if
there are systematic differences between those that have participated in SITP and left the school
district and those that did not participate in SITP and left the school district.
The length of the program instruction in the museums and cultural institutions of Balboa
Park varies across program years, largely due to funding constraints on the SITP program.
Following the primary analysis, we estimate dosage models to understand whether the length of
time a student participates in the program moderates the effect of SITP on student outcomes.

education students among the control group students. Finally, like the Hamilton students, the control groups
students’ second grade ELA and math test scores were lower after the SITP implementation at both of the schools.
22
Measures
The student-level data sets were merged together using unique student identification
numbers, school year, and school codes. From this large, longitudinal data set, we created
multiple measures for the analysis. The multiple short-term academic outcomes include the
probability that students take the standardized exams, the actual scores achieved on those exams,
and whether the student performs at or above the standard set by the state.
In the 2001/02 school year, California changed the standardized test used throughout the
state from the SAT9 to the California Standards Test (CST). This makes it difficult to compare
results for students over time. Therefore, we have created a standardized test score measure that
compares student performance to that of his or her grade level peers each year. These z-scores
have a mean of zero and a standard deviation of one. During the 2001/02 school year, both the
SAT9 and CST exams were conducted. For the students who took both exams, we calculated z-
scores and then examined how highly correlated the scores were in that year. We find that the z-
scores for the two exams are highly correlated – the ELA z-scores are correlated at 0.894 and the
math z-scores are correlated at 0.878. Therefore, in the models that span the period of time where
both exams were given, we rely on the z-score measures of performance.
Following the examination of test scores, we consider whether SITP influences grade
retention in the third, fourth, or fifth grades and two short-term behavioral outcome measures –
the number of absences in the current year and the number of suspensions in the current year.
As mentioned earlier, one of the aspirations of the SITP program is that the impacts
extend well beyond the school year in which the student participates in the program. We
investigate longer-term outcomes for students as they reach middle school and high school.
Three measures explore middle school behavior and achievement – the number of suspensions
and expulsions in middle school, and whether the student is retained by the 8
th
grade. For a
23
subset of students who participated in SITP prior to 2005, we observe progress through high
school. For these students, we measure whether they have been retained in high school, whether
they have taken an AP course or the SAT test (markers for intended college attendance), their
SAT scores, whether they pass the California High School Exit Examination (CAHSEE), a
requirement for graduation, whether they graduate with a diploma or earn a GED, their college
enrollment, and whether they attend a two-year or four-year college.
We constructed two dropout measures to estimate the impact of SITP on the probability
of dropping out of high school (see Appendix B). These measures only include students who are
old enough to have enrolled in high school. The “least inclusive” dropout measure includes only
students that the District coded as “dropouts.” This results in a dropout rate of 2.11% (out of the
total number of students who are old enough to attend high school). The most inclusive measure
includes the students coded by the District as dropouts as well as the students who we do not
observe completing high school in the data. Students who ultimately received a GED or diploma
were coded as non-dropouts even if they had previously dropped out. This most inclusive
measure may include some students who transfer successfully to other schools but whose
transfer was not recorded by the district. The most inclusive measure results in a dropout rate of
22.05%, a much more realistic rate based on national and statewide averages.
8
While the group
we have identified as dropouts has a higher average 10th grade ELA test scores than the group
labeled by the district as dropouts (-0.210 s.d., -0.168 if the district identified dropouts are
excluded, compared to -0.600 s.d., respectively), their average score is substantially lower than
the 10
th
grade scores of high school graduates (0.342 s.d.) leading us to believe that these

8
Statewide dropout rate in 2011-12 in California: 13%. The rates are higher for English Learners (23.7), migrant
students (16.4), and socioeconomically disadvantaged students (16.4). Source: California Department of Education
http://www.cde.ca.gov/nr/ne/yr13/yr13rel42att.asp#tab1.
24
struggling students are likely not finishing high school. While we acknowledge this measure to
likely be an overestimate of the number of students who dropout, we feel it is more
representative and therefore we report results for the most inclusive dropout measure.
In addition to these outcome measures, the models include a comprehensive set of control
variables at the individual, classroom, and school levels. Individual-level controls include
measures of gender, race and ethnicity (black, Hispanic, Asian, white, or other race), English
language learner status, special education status, and whether the student moved schools in the
current year. The models also control for students’ birth country (natives, immigrants, and
refugees). Students who were born in Burma, Iran, Iraq, or Somalia are designated as refugees,
as these countries have been the largest providers of refugees to the San Diego area. Those born
outside of the United States in countries other than those four are classified as non-refugee
immigrants (referred to as simply immigrants throughout the rest of this article). Those born in
the United States are classified as natives. It is important to note that studies of immigrant and
refugee youth generally include children born in the United States to immigrant or refugee
parents in addition to those born in another country. As the parental birth countries of the
students in this sample are unknown, students who would typically be classified as immigrants or
refugees are included in the native group. Finally, at the school-level, we control for the percent
of students eligible for free or reduced price lunch (a proxy for poverty), total student enrollment,
whether or not the school is a charter school, and the years in which a school health clinic existed
at Rosa Parks in the short-term models. In the high school models, we control for whether or not
the student attended Clark Middle School, Hoover or one of the Crawford High Schools, whether
or not the high school was a charter school, and whether the student moved in elementary or
middle school.
25
Sample
The analytic sample is restricted to students in grades 2, 3, 4, or 5 at Rosa Parks,
Alexander Hamilton, or one of the other elementary schools in the Hoover High School or
Crawford High School system between the 1995/96 and 2011/12 school years. Schools are
included in the comparison group if we observe at least 10 students in the school in a given year.
The sample size changes by the outcome measure of interest because there is a smaller sample of
students for which we observe high school outcomes. In the short-term models, the first year of
implementation at Rosa Parks, the 1999/00 school year, is omitted from analysis because only
half of the students in the third grade received the program that year.
9
In the long-term models,
there is only one observation per student so all of the 1999/2000 Rosa Parks third graders are
coded as having received the program.
10

6. Results
Descriptive Statistics
Table 1 presents mean descriptive statistics of the students who attended Rosa Parks
between the 2000/01 and 2011/12 school years (and received SITP), those who attended
Hamilton between 2006/07 and 2011/12 school years (and received SITP), and those students
who attended the control schools in the Hoover and Crawford High School systems between
2000/01 and 2011/12 (who never received the program). Both the Rosa Parks and Hamilton
SITP samples are majority Hispanic. While the control group also consists of a Hispanic
majority, Hispanic students make up a smaller share of the student population than in either Rosa

9
See Appendix A for a timeline of SITP implementation across the two schools.
10
Of the 285 third graders who attended Rosa Parks during the 1999/2000 school year, 231 attend Rosa Parks in at
least one subsequent year and thus received School in the Park at some point during their education. We include all
of the third graders when constructing the long-term sample since, at most, we are only unsure about whether 54 of
the 1999/2000 third graders ever received the program. Consequently any estimate of the long-term effects is an
underestimate.
26
Parks or Hamilton (59 percent in the control schools, 80 percent at Rosa Parks, and 75 percent at
Hamilton). The control group is nearly 18 percent black and 14 percent Asian whereas Rosa
Parks and Hamilton are 7 and 10 percent black and both 10 percent Asian, respectively. Students
classified as English language learners make up a larger share of the student body at Rosa Parks
and Hamilton than the control group average. On all other demographic measures, the treatment
schools are comparable to the control schools.
The two schools have roughly the same percentage of teachers with over two years of
teaching experience as the control group. Rosa Parks and Hamilton also have student populations
that are 100 percent free and reduced lunch eligible whereas the control group population is 93
percent eligible. Rosa Parks is a larger school, serving nearly 1300 students on average,
compared to Hamilton’s 700 students and the control group’s 800 students on average.
Test-taking rates across all of the groups are high: 92 to 94 percent on average. Rosa
Parks and Hamilton SITP students’ rate of taking standardized tests are on par with that of the
control group. However, students in the control schools outperform students at Rosa Parks in
both their language arts and math test z-scores, and Rosa Parks students outperform Hamilton
students. Students in both treatment schools experience fewer absences (5.8 and 5.5 absences
versus the control group’s 6.7) and fewer suspensions (0.019 and 0.017 compared to 0.044) than
the control group.
During middle school, Rosa Parks and Hamilton SITP students have similar numbers of
suspensions and expulsions, which are higher than the control group’s averages (Table 2).
Students at Rosa Parks and in the control group have comparable average math and language arts
test z-scores in seventh grade and substantially outperform students from Hamilton.
27
For the high school outcomes, we only compare the students who participate in SITP
from Rosa Parks students to the control group, as even the oldest Hamilton students who
participated in the first year of the program are not yet old enough to have completed high school
(Table 2). Rosa Parks students have similar rates of high school suspensions, expulsions, grade
retention, and dropouts as the control group. Of students who attended SITP at Rosa Parks, 20.4
percent take an Advanced Placement course at some point in high school compared to 22.3
percent of students who attended the control schools as elementary students. The share of
students who graduate with a diploma and who take the SAT is similar across the treatment and
control groups, although in both cases Rosa Parks students have slightly higher rates of both.
Attrition
One concern in this analysis is that we can only track students within the San Diego
Unified School district. If students who were relatively disadvantaged were more likely to leave
the district, and the sample, before their middle or high school outcomes could be observed, then
the estimated impact of SITP would be biased upward. This is an especially relevant issue in a
community with a large immigrant and refugee population. The sample provides mixed evidence
for this concern. While students who participated in SITP and left the district in middle school
were significantly more likely to be Hispanic and classified as English language learners and less
likely to be white or black than those who left who had not participated in SITP, this was also
true of SITP participants who stayed in the district during middle school when compared to non-
SITP students who also remained (see Appendix D). There were no differences in the second
grade test scores of SITP and non-SITP middle school leavers, and in fact among those who
stayed, SITP participants had lower second grade language arts and math test scores. There were
no differences in the second grade math scores among high school SITP and non-SITP leavers.
28
While the second grade language arts test scores of SITP participants who left in high school
were lower than those of their non-SITP leaving peers, this was also true when comparing SITP
participants who were present in high school to non-SITP stayers. SITP participants who were
not present in high school were more likely to have experienced a suspension than their non-
SITP leavers. Both leaver and stayer SITP participants were less likely to have been retained in
middle school than their non-SITP peers. Thus, while there is evidence that SITP participants
who left the sample were more disadvantaged in terms of their nativity and behavior than their
non-SITP leaver peers, on most measures there was either no difference between SITP and non-
SITP leavers or the difference was consistent across both SITP leavers and stayers.
Short-term Models
The first set of models in Table 3 present the impact of participation in SITP on outcomes
in the year of participation. These models utilize panel data and include individual student fixed
effects. We find that participation in SITP results in significant increases in standardized test
scores in the year of SITP participation. Rosa Parks SITP students score 0.1 standard deviations
higher on the standardized math exam, compared to comparison group students in the same year.
SITP participation at Hamilton results in an increase in the likelihood of taking both of the
language arts and math exams, as well as a 0.07 standard deviation increase in language arts
scores. Further, SITP participation at both schools reduces suspensions. There is no difference in
absences between the SITP and comparison groups, but a small increase in the probability of
being retained in the third, fourth, or fifth grades for Hamilton SITP students.
One concern that has been expressed by some educators is that students who are already
struggling in school would be adversely affected by an extended period of time outside the
classroom to participate in SITP (San Diego Unified School District staff, personal
communication, 2013). To investigate this issue, in Table 4 we present the short-term results
29
stratified by whether the student passed or failed the second grade standardized language arts or
math exams. Consistent with the previous findings, we find no impact of SITP on language arts
scores for students who passed and students who failed their 2nd-grade language arts exam at
Rosa Parks. In other words, participation in the program does not appear to hinder the
achievement of students struggling in second grade compared to similarly struggling students in
comparison schools. Further, the program positively impacts math scores for Rosa Parks students
who passed the second grade standardized math test and those who failed it, compared to
students with the same second grade performance in the comparison schools. For Hamilton
students, the impact of SITP on language arts scores is concentrated among students who failed
their second grade language arts test, with no effect on students who had passed the 2nd-grade
exam. Despite the improvement in language arts scores, SITP students at Hamilton who failed
their 2nd-grade math test were slightly more likely to be retained between the third and fifth
grades, than comparison group students.
Another hypothesis is that experiential education, if conducted in a culturally relevant
way, may reach immigrant and refugee students more effectively than traditional educational
models. This hypothesis is based on the notion that students bring their prior knowledge to
museum experiences, and that the museum generates reflection and inquiry and stimulates new
interests, which are carried into students’ outside lives (Hein, 2004). Thus, while immigrant and
refugee students’ knowledge of their own cultures may be under-utilized in a typical classroom
setting, museums may offer these students the opportunity to engage with material more
effectively. In order to explore this possibility, we stratify the short-term results by whether the
student is a refugee, immigrant, or native-born (see Appendix E). While refugees at Rosa Parks
who participated in SITP were more likely to take both language arts and math exams, there
30
were no other significant impacts associated with participation for refugees at Rosa Parks and
Hamilton. However, SITP participation among immigrant students at Rosa Parks led to a 0.09
standard deviation increase in their math scores though their Hamilton counterparts experienced
a 0.10 decrease. Immigrant students at both Rosa Parks and Hamilton were slightly less likely to
be suspended and slightly more likely to be retained between the third and fifth grades. SITP
participating immigrants at Rosa Parks also experienced a reduction in their school absences.
Cumulative Weeks Analysis
Changes over the course of the program’s existence led to variation in the amount of time
students participate in the program by year and grade (see Appendix A). If participating in SITP
generates positive short-term outcomes, students who spent more time in the program may reap
larger benefits. Therefore, we analyze the impact of the total number of weeks of SITP
participation. One of the reasons that many other museum enrichment programs analyzed in the
literature have no impact may be because they simply do not provide enough exposure to the
enrichment activities. Our SITP dosage findings provide some initial evidence that this is in fact
true. In the short-term, the results depicted in Table 5 each additional week of SITP participation
for students who attended Hamilton is associated with a small (less than one percent) but
significant increase in the likelihood that students took their language arts and math standardized
tests. At both Rosa Parks and Hamilton, each additional week of SITP programming increases
student ELA test scores by 0.004 and 0.006 standard deviations, respectively. Given the average
length of participation in SITP per year is 7.68 weeks , these coefficients translate into 0.031 and
0.046 standard deviation increases in ELA scores, on average.
11
The marginal week of SITP

11
The number of weeks of participation in the program per year ranges from 5 weeks for fifth grade students in
certain years to 12 weeks for third grade students at the beginning of the program (see Appendix A), therefore the
effect of SITP on ELA scores may range from 0.02 s.d. to 0.05 s.d. at Rosa Parks, and from 0.03 s.d. to 0.07 s.d. at
Hamilton.
31
participation is also associated with a less than one percent decrease in the likelihood of being
held back at Rosa Parks, and a decrease in suspensions at Hamilton.
12

Long-term models
The long-term models utilize a dataset with one observation per student and outcomes
measured through high school. All of the long-term models include elementary school fixed
effects, and controls for the third grade year cohort. There are no significant, lasting effects of
participation in SITP through the high school years. Students who participate in SITP as
elementary school students have outcomes in high school comparable to those students who did
not participate in SITP. In Table 6, there are no significant effects of participation in SITP
overall on suspensions, expulsions, or the probability of being retained in high school, dropping
out of high school, taking the SAT, passing the high school exit exam, or ever enrolling in
college. However, SITP participation did result in a small decrease in SAT verbal scores (Panel
B). We ran an additional analysis (not included in these tables) of the effect of the cumulative
number of weeks of SITP participation on high school and college outcomes and found similarly
insignificant results. An additional week of SITP participation does not result in any significant
outcomes for Rosa Parks SITP students. These models suggest that while the length of
participation in SITP is important in the short-term, by high school these benefits fade regardless
of the duration of participation.
7. Discussion and Policy Implications
Students who participate in SITP experience small short-term gains in test scores, and
positive effects on behavioral outcomes (decreases in suspensions and retention). Furthermore,
there is evidence of a SITP dosage effect - an increase in the number of weeks of SITP

12
We also estimated models with categorical variables representing different lengths of participation in SITP, but
the results were not conclusive given the somewhat arbitrary categorization of program lengths into categories.
32
participation is associated with positive, albeit somewhat different, outcomes at both schools.
However, we find no effect of ever participating in SITP on the students’ long-term academic
and behavioral outcomes.
These findings support several important policy implications. First, there is no evidence
that exposing students to a new learning environment in lieu of traditional class time harms their
educational achievement. In fact, we observe gains for the students who are struggling prior to
the program, compared to similar students who do not receive the program. The perceived
tradeoff between out-of-classroom time and achievement is not grounded in the evidence from
this program.
Second, in the short term, we find consistent improvement in behavioral outcomes,
including a decrease in the probability of suspension. These behavioral changes might translate
into improved academic achievement in the longer term. One way to invest in student
achievement may be to indirectly build behavioral competencies – such as sitting still, paying
attention, speaking in turn – that will facilitate learning for years to come.
Third, although our analysis captured no lasting effects of the SITP program on academic
and behavioral outcomes in high school, it should be noted that given the timing of the program
and the years contained in our data, most of the long-term outcome findings are based solely on
the first four years of the program, as only those SITP participants are old enough in our data to
have completed high school. SDUSD educators indicated that the program has been redesigned
substantially over time, and it may be that an analysis incorporating subsequent years would
produce results more consistent with our short-term findings. Additional research in other
contexts is needed to better understand if long-term impacts exist for museum enrichment
programs such as SITP.
33
Fourth, the program utilizes existing community resources that many cities and states
nationwide may also be able to access. At first, the program re-purposed museum space as
classroom space, when Rosa Parks was unable to accommodate all of the students in the
neighborhood. Over time, the program grew to serve more students in more grades, utilizing
multiple cultural institutions at Balboa Park, with teachers partnering with museum educators to
provide lessons to the students. Although the program has associated costs, by leveraging
existing physical resources and partnering with museum-based educators to adapt existing
curricula to meet the needs of the district, the costs are potentially lower than other experiential
education programs.
Overall, the program demonstrates impacts in the short-term, well beyond those intended
by the program in its design. If we are concerned with affecting student achievement in the short-
run by taking time from traditional classroom activities, these concerns are not consistent with
the evidence from SITP. In fact, the most at-risk students could receive the largest benefits from
investing in experiential education programs that offer prolonged and structured experiences for
students outside the classroom. Indeed, in the data we did find a positive long term association
between SITP based on the length of exposure, but the positive impacts should not be interpreted
as causal due to the fact that students who moved the most often were also those who
experienced the least exposure. Additional research is required to determine what might allow
the short term benefits to persist over time. Finally, programs like SITP do require resources, and
benefit cost analyses or cost effectiveness analyses should be used to compare new educational
inventions that have demonstrated increased achievemen
34
Figures and Tables
Figure 1. Conceptual Model of SITP Impact
35
8. Tables
Table 1. Descriptive statistics, means for short-term outcomes by SITP participation status (student-year
observations), 2000/01-2011/12
Variable
Rosa Parks
(SITP)
2000/01-2011/12
Alexander Hamilton
(SITP)
2006/07-2011/12
Other Schools
13

(No SITP)
2000/01-2011/12
Number of Observations 6,783 1,874 61,580
Student Characteristics
Male 0.51 0.53 0.51
Black 0.06 0.09 0.16
Hispanic 0.86 0.77 0.62
Asian 0.10 0.10 0.15
White 0.01 0.01 0.06
Other Race 0.01 0.04 0.02
English Learner (ever classified) 0.81 0.77 0.67
Immigrant 0.16 0.16 0.15
Refugee 0.001 0.005 0.006
Special Education (ever classified) 0.17 0.16 0.17
3
rd
Grade 0.35 0.35 0.27
4
th
Grade 0.34 0.33 0.25
5
th
Grade 0.30 0.32 0.23
School Characteristics
% Eligible for Free/Reduced Lunch 100 100 92.1
Enrollment 1252.2 673.4 671.5
Teacher Average Years of Experience 10.5 15.4 11.3
% Teachers w 2+ Years of Experience 0.97 0.98 0.96
Charter School - - 0.04
Short-term Outcomes
ELA z-score 0.09 0.04 0.13
Math z-score 0.17 -0.03 0.13
Days absent 5.76 5.54 6.84
Number of suspensions 0.02 0.02 0.04
Retained 3
rd
-5
th
Grade 0.009 0.005 0.004
13
Other schools only include schools in the Crawford or Hoover clusters.
36
Table 2. Descriptive statistics, overall means for long-term outcomes by SITP participation status (single
observation per student), 2000/01-2011/12
Variable
Rosa Parks
(SITP)
2000/01-2011/12
Alexander Hamilton
(SITP)
2006/07-2011/12
Other Schools
(No SITP)
2000/01-2011/12
Number of Observations 3,039 1,029 21,604
Student Characteristics
Male 0.51 0.52 0.51
Black 0.07 0.10 0.18
Hispanic 0.80 0.75 0.60
Asian 0.10 0.10 0.14
White 0.01 0.01 0.06
Other Race 0.01 0.03 0.02
English Learner (ever classified) 0.80 0.77 0.63
Immigrant 0.15 0.16 0.14
Refugee 0.001 0.005 0.008
Special Education (ever classified) 0.17 0.17 0.17
Longer-term Outcomes
Retained 6
th
-8
th
Grade 0.01 0.01 0.03
Middle school suspensions 0.79 0.97 0.70
Middle school expulsions 0.03 0.03 0.02
7
th
Grade ELA z-score 0.11 -0.14 0.11
7
th
Grade Math z-score 0.09 -0.22 0.09
Retained 9
th
-12
th
Grade 0.32 - 0.33
Take AP Course 0.20 - 0.22
Take SAT Exam 0.43 - 0.41
SAT Verbal Score 423.2 - 443.7
SAT Math Score 442.7 - 454.1
HS Diploma 0.40 - 0.37
GED 0.01 - 0.01
College 0.23 - 0.22
2 year college 0.15 - 0.13
4 year college 0.08 - 0.09
37
Table 3. Short-Term academic outcomes (1996-2012, omit 2000)
(1) (2) (3) (4) (5) (6) (7)
VARIABLES Take ELA ELA Z Score Take Math Math Z Score Days Absent Suspensions Held Back 3/5
SITP*Rosa 0.010 0.040 0.006 0.096** -0.011 -0.011* 0.015
(0.005) (0.029) (0.006) (0.031) (0.088) (0.005) (0.012)
Rosa -0.051* 0.175 -0.028 0.086 -0.509 -0.041 -0.012
(0.020) (0.141) (0.024) (0.142) (0.428) (0.026) (0.016)
SITP*Hamilton 0.017** 0.066* 0.014* -0.024 0.056 -0.020* 0.009*
(0.006) (0.027) (0.007) (0.030) (0.140) (0.008) (0.004)
Hamilton -0.013 0.220** -0.002 0.143 -0.284 -0.026 -0.005
(0.012) (0.070) (0.014) (0.073) (0.330) (0.024) (0.009)
Special Education -0.136** -0.092** -0.110** -0.045 0.388* 0.054** 0.002
(0.026) (0.020) (0.021) (0.024) (0.180) (0.019) (0.004)
Moved 0.010 -0.045** 0.007 -0.051** 1.384** 0.028** -0.001
(0.006) (0.015) (0.007) (0.013) (0.165) (0.006) (0.002)
EL Status -0.070** -0.074** -0.066** -0.080** -0.087 0.013** 0.004*
(0.005) (0.019) (0.005) (0.020) (0.110) (0.004) (0.002)
School % FRL 0.00 0.003 0.000 0.002 0.005 0.000 0.000
(0.000) (0.002) (0.000) (0.002) (0.011) (0.000) (0.000)
Enrollment 0.000 0.000 0.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.001) (0.000) (0.000)
Charter 0.010 0.017 0.007 0.121 0.664 0.092* 0.004
(0.018) (0.093) (0.021) (0.103) (0.400) (0.033) (-0.018)
Health Clinic -0.004 -0.040 -0.004 -0.127** 0.405* 0.013 -0.006
(0.008) (0.047) (0.009) (0.039) (0.188) (0.012) (0.003)
Constant -0.253 2.796** -0.130 4.400** 5.348 -0.175 -2.919**
(0.182) (0.316) (0.180) (0.309) (4.172) (0.086) (0.148)
Observations 77,819 72,919 77,819 73,010 77,665 77,819 52,941
R-squared 0.522 0.845 0.514 0.823 0.720 0.550 0.718
Year FE Yes Yes Yes Yes Yes Yes Yes
School FE Yes Yes Yes Yes Yes Yes Yes
Grade FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Clusters 22 22 22 22 22 22 22
Robust standard errors in parentheses. FRL = Free or reduced price lunch. FE = fixed effect.
** p<0.01, * p<0.05
38
Table 4. Short-term outcomes, stratified by 2
nd
-grade test pass or fail (1996-2012, omit 2000)
Dep. Variable ELA Z Score Math Z Score Days Absent Suspensions Held Back 3/5
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
2
nd
Grade Test: PASS FAIL PASS FAIL PASS FAIL PASS FAIL PASS FAIL
SITP*Rosa 0.044 -0.007 0.087* 0.102** 0.023 -0.127 -0.008 -0.001 -0.013 0.037
(0.023) (0.027) (0.033) (0.033) (0.079) (0.204) (0.005) (0.007) (0.008) (0.030)
Rosa 0.138 0.153 -0.021 0.128 -0.308 -0.332 0.000 -0.0435
(0.141) (0.127) (0.202) (0.120) (0.769) (1.143) (0.029) (0.077)
SITP*Hamilton 0.005 0.107** -0.024 -0.013 0.189 -0.387 -0.009 -0.046* 0.004 0.026**
(0.020) (0.030) (0.028) (0.030) (0.126) (0.289) (0.006) (0.018) (0.003) (0.008)
Hamilton 0.276** 0.162* 0.102 0.153* -0.159 -0.102 -0.024 0.000 0.004 -0.001
(0.072) (0.063) (0.101) (0.068) (0.457) (0.727) (0.024) (0.0645) (0.005) (0.023)
Constant -2.650** -1.632** -3.341** -3.214** 5.000 9.534** -0.229 -0.268 -1.518** 3.542**
(0.448) (0.291) (0.475) (0.384) (2.428) (3.308) (0.195) (0.135) (0.177) (0.236)
Observations 38,052 20,760 38,670 14,285 40,144 16,261 40,206 16,311 26,005 10,857
R-squared 0.790 0.678 0.756 0.724 0.709 0.722 0.538 0.556 0.677 0.731
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
School FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Grade FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Clusters 22 22 22 22 22 22 22 22 22 22
Robust standard errors in parentheses. ELA Z score models (columns 1 and 2) are stratified by 2nd grade ELA test score results, while the rest of the models are stratified by
the results of the 2nd-grade math test. Models also control for current EL status, special education status, moved schools, % of student body eligible for free lunch, total
enrollment, charter school, and health clinic. FE = fixed effect.
** p<0.01, * p<0.05
39
Table 5. Cumulative week analysis, short-term outcomes (1996-2012, omit 2000)
(1) (2) (3) (4) (5) (6) (7)
VARIABLES Take ELA ELA Z Score Take Math Math Z Score Days Absent Suspensions Held back
Total SITP weeks*Rosa
0.000 0.004* 0.000 0.003 -0.006 -0.001 -0.0005*
(0.000) (0.002) (0.000) (0.002) (0.005) (0.000) (0.000)
Rosa
-0.047* 0.175 -0.027 0.103 -0.461 -0.043 0.006
(0.020) (0.132) (0.024) (0.135) (0.420) (0.026) (0.011)
Total SITP weeks*Hamilton
0.001** 0.006** 0.001** 0.003 -0.003 -0.002** 0.000
(0.000) (0.002) (0.000) (0.002) (0.010) (0.000) (0.000)
Hamilton
-0.014 0.213** -0.004 0.122 -0.246 -0.025 -0.001
(0.013) (0.068) (0.014) (0.071) (0.327) (0.024) (0.009)
Constant
-0.255 2.730** -0.134 4.351** 5.379 -0.173 -2.909**
(0.182) (0.330) (0.180) (0.323) (4.155) (0.0846) (0.142)
Observations 77,819 72,919 77,819 73,010 77,665 77,819 52,941
R-squared 0.522 0.845 0.514 0.823 0.720 0.550 0.718
Year FE Yes Yes Yes Yes Yes Yes Yes
School FE Yes Yes Yes Yes Yes Yes Yes
Grade FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Clusters 22 22 22 22 22 22 22
Robust standard errors in parentheses. Models also control for current EL status, special education status, moved schools, % of student body eligible for free lunch, total
enrollment, charter school, and health clinic. Total SITP weeks is the cumulative number of weeks of SITP programming a student has received in each year.
FE = fixed effect.
** p<0.01, * p<0.05
40
Table 6. Long-Term outcomes (1996-2012)
A. Progress/Behavior (1) (2) (3) (4)
VARIABLES ELA Z Score (10th) Suspensions HS Expulsions HS Retention 9/12
SITP Ever*Rosa -0.034 -0.120 -0.011 -0.025
(0.056) (0.082) (0.011) (0.030)
Rosa 0.012 0.222 0.019 -0.002
(0.074) (0.143) (0.016) (0.033)
Constant 0.210 -0.050 -0.022* 0.137
(0.387) (0.288) (0.010) (0.208)
Observations 7,429 10,886 10,886 8,894
R-squared 0.393 0.090 0.019 0.182
Clusters (Elem. School) 93 93 93 93
B. College Preparation (1) (2) (3) (4)
VARIABLES Take AP Course Take SAT
SAT Verbal
Score SAT Math Score
SITP Ever*Rosa -0.011 0.010 -19.30* -13.84
(0.028) (0.037) (8.429) (15.50)
Rosa -0.030 -0.023 21.63 31.65
(0.032) (0.056) (14.54) (21.16)
Constant -0.002 0.078 512.4** 469.7**
(0.065) (0.061) (15.02) (13.97)
Observations 10,886 3,592 1,481 1,481
R-squared 0.213 0.201 0.411 0.453
Clusters (Elem. School) 93 93 93 93
C. Completion (1) (2) (3) (4)
VARIABLES HS Diploma GED Pass CAHSEE Dropout (high)
SITP Ever*Rosa 0.019 0.000 0.007 -0.026
(0.027) (0.006) (0.015) (0.024)
Rosa -0.024 -0.005 -0.037 0.047
(0.034) (0.0063) (0.033) (0.028)
Constant 0.154 0.271 0.852** -0.079
(0.127) (0.215) (0.190) (0.037)
Observations 6,719 6,719 7,636 10,886
R-squared 0.221 0.041 0.117 0.221
Clusters (Elem. School) 93 93 93 93
Elem School FE Yes Yes Yes Yes
3rd Grade Cohort FE Yes Yes Yes Yes
Robust standard errors in parentheses. ** p<0.01, * p<0.05.
Models include individual student characteristics (gender, race/ethnicity, EL status, immigrant, refugee, special education,
2nd grade ELA scores, attended Clark Middle School dummy, elementary and middle school move dummies), school
characteristics (% eligible FRL, HS charter dummy), and controls for Crawford and Hoover High Schools. FE = fixed
effect.
41
Table 6 (continued). Long-term outcomes (1996-2012)
D. College Enrollment (1) (2) (3)
VARIABLES College Ever 2 year college high 4 year college high
SITP Ever*Rosa 0.006 0.004 0.002
(0.028) (0.022) (0.019)
Rosa -0.038 -0.015 -0.022
(0.028) (0.020) (0.023)
Constant 0.059 0.132* -0.068
(0.106) (0.052) (0.068)
Observations 6,629 6,629 6,629
R-squared 0.281 0.124 0.171
Clusters (Elem. School) 93 93 93
Elem School FE Yes Yes Yes
3rd Grade Cohort FE Yes Yes Yes
Robust standard errors in parentheses. ** p<0.01, * p<0.05.
Models include individual student characteristics (gender, race/ethnicity, EL status, immigrant, refugee, special
education, 2nd grade ELA scores, attended Clark Middle School dummy, elementary and middle school move
dummies), school characteristics (% eligible FRL, HS charter dummy), and controls for Crawford and Hoover
High Schools.
FE = fixed effect.
42
9. Appendix
Appendix A. Participation in SITP by school, grade, and year
School Participants Grade Participants Program Weeks Student Participants
Year Rosa Hamilton
2nd
Grade
3rd
Grade
4th
Grade
5th
Grade
2nd
Grade
3rd
Grade
4th
Grade
5th
Grade
2nd
Grade
3rd
Grade
4th
Grade
5th
Grade
Total
99/00
Yes No No Yes No No - 12 - - - 180 - - 180
00/01
Yes No No Yes Yes No - 12 9 - - 280 240 - 520
01/02
Yes No No Yes Yes Yes - 9 7 5 - 280 240 240 760
02/03
Yes No No Yes Yes Yes - 9 7 5 - 200 270 270 740
03/04
Yes No No Yes Yes Yes - 9 7 5 - 280 210 270 760
04/05
Yes No No Yes Yes Yes - 9 7 6 - 260 270 240 770
05/06
Yes No No Yes Yes Yes - 9 8 6 - 240 240 270 750
06/07
Yes Yes No Yes Yes Yes - 9 9 6 - 360 390 420 1170
07/08
Yes Yes No Yes Yes Yes - 9 8 6 - 260 300 330 890
08/09
Yes Yes No Yes Yes Yes - 9 8 5 - 260 240 270 770
09/10
Yes Yes No Yes Yes Yes - 8 8 8 - 240 250 250 740
10/11
Yes Yes No Yes Yes Yes - 8 8 8 - 220 240 240 700
11/12
Yes Yes No Yes Yes Yes - 8 8 8 - 240 240 240 720
43
Appendix B. Specifying the dropout measure
Dropout
(Least Inclusive
Measure)
Dropout (Most Inclusive Measure)
No
Dropout Dropout Missing
No Dropout 25,099 9,207 0 34,306
Dropout 0 975 0 975
Missing 0 0 11,639 11,639
Total 25,099 10,182 11,639 46,920
44
Appendix C
Table C-1. Descriptive statistics, 3rd through 5th grades, 2000/01-2005/06
Treatment Comparisons
Variable (means) Rosa Parks Alexander Hamilton Other Schools
14

Number of Observations 3,949/1,894
15
3,110/1,682 22,315/11,817
Student Characteristics
Male 0.513 0.510 0.507
Black 0.071 0.166 0.196
Hispanic 0.789 0.642 0.565
Asian 0.115 0.151 0.144
White 0.016 0.034 0.079
Other Race 0.009 0.007 0.016
English Learner 0.786 0.633 0.598
Immigrant 0.145 0.117 0.135
Refugee 0.001 0.006 0.010
Special Education 0.169 0.157 0.173
3
rd
Grade 0.369 0.360 0.371
4
th
Grade 0.351 0.331 0.331
5
th
Grade 0.279 0.310 0.298
School Characteristics
% Eligible for Free/Reduced Lunch 100 100 88.233
Enrollment 1444.06 1117.12 716.49
Teacher Average Years of Experience 8.243 9.600 10.172
% Teachers with 2+ Years of Experience 0.955 0.928 0.929
Charter School 0 0 0.069
Short-term Outcomes
ELA z-score 0.087 0.139 0.167
Math z-score 0.212 0.144 0.149
Days absent 6.036 6.907 7.083
Number of suspensions 0.014 0.033 0.046
Retained 3
rd
-5
th
Grade 0.014 0.005 0.004
Longer-term Outcomes
Retained 6
th
-8
th
Grade 0.008 0.022 0.032
Middle school suspensions 0.859 0.960 0.797
Middle school expulsions 0.034 0.037 0.028
7
th
Grade ELA z-score 0.149 0.035 0.121
7
th
Grade Math z-score 0.142 -0.006 0.096
Retained 9
th
-12
th
Grade 0.318 0.395 0.314
Take AP Course 0.212 0.192 0.222
Take SAT Exam 0.426 0.377 0.411
SAT Verbal Score 424.066 418.108 445.525
SAT Math Score 444.945 428.716 456.875
HS Diploma 0.405 0.350 0.389
GED 0.004 0.006 0.012
College 0.237 0.189 0.232
2 year college 0.157 0.122 0.135
4 year college 0.080 0.067 0.097
14
Other schools only include schools in the Crawford or Hoover clusters.
15
The first number represents the sample size for the short-term outcomes which use student-year observations,
the second number represents the sample size for the long-term outcomes, which contains a single observation per
student.
45
Appendix C
Table C-2. Descriptive statistics, 3rd through 5th grades, 2006/07-2011/12
Treatment Comparison
Variables (Means) Rosa Parks Hamilton Other Schools
16

Total Observations 2,834/1,504 1,874/1,029
20,832/10,405
Student Characteristics
Male 0.509 0.525 0.508
Black 0.061 0.101 0.156
Hispanic 0.827 0.754 0.639
Asian 0.095 0.100 0.135
White 0.008 0.013 0.043
Other Race 0.009 0.031 0.027
English Learner 0.833 0.769 0.704
Immigrant 0.172 0.160 0.155
Refugee 0.001 0.005 0.005
Special Education 0.166 0.165
0.175
3
rd
Grade
0.334 0.354 0.350
4
th
Grade
0.331 0.327 0.330
5
th
Grade 0.335 0.318 0.319
School Characteristics
% Eligible for Free/Reduced Lunch 100 100
94.671
Enrollment 984.953 673.409 552.874
Teacher Average Years of Experience 13.546 15.426 12.991
% Teachers with 2+ Years of Experience 0.995 0.983 0.994
Charter School 0 0 0.016
Short-term Outcomes
CST ELA test taken 0.895 0.918 0.894
CST math test taken 0.903 0.922 0.900
CST ELA score 334.658 332.065 338.207
CST math score 363.112 352.046 367.044
Passed CST ELA test 0.738 0.723 0.755
Passed CST math test 0.801 0.743 0.792
ELA z-score 0.091 0.040 0.140
Math z-score 0.119 -0.027 0.149
Days absent 5.377 5.540 6.351
Number of suspensions 0.026 0.017 0.058
Retained 3
rd
-5
th
Grade 0.002 0.005
0.005
Longer-term Outcomes
Retained 6
th
-8
th
Grade 0.004 0.014 0.014
Middle school suspensions 0.710 0.967 0.548
Middle school expulsions 0.026 0.027 0.014
7
th
Grade ELA z-score 0.098 -0.140 0.085
7
th
Grade Math z-score -0.102 -0.218 0.090
16
Other schools only include schools in the Crawford or Hoover clusters.
46
Appendix D.
Table D-1. Attrition and means for SITP participant middle school leavers and non-SITP leavers
Non-SITP leavers SITP leavers t score
p-value
Total Observations 3,340 347
Mean SD Mean SD
Male

0.508 0.500 0.484 0.500 0.859 0.390
Female 0.492 0.500 0.516 0.500 -0.859 0.390
Black 0.250 0.433 0.110 0.313 5.875** 0.000
Hispanic 0.521 0.500 0.749 0.434 -8.186** 0.000
Asian 0.099 0.298 0.107 0.309 -0.463 0.643
White 0.112 0.315 0.026 0.159 5.015** 0.000
Other Race 0.018 0.134 0.009 0.093 1.306 0.192
English Learner 0.443 0.497 0.657 0.475 -7.677** 0.000
Immigrant 0.175 0.380 0.208 0.407 -1.240 0.215
Refugee 0.009 0.094 0.004 0.066 0.730 0.465
Special Education 0.131 0.338 0.130 0.336 0.092 0.927
2
nd
Grade ELA z-score 0.069 1.015 0.058 0.926 0.197 0.844
2
nd
Grade Math z-score 0.054 0.987 0.090 0.952 -0.643 0.520
** p<0.01, * p<0.05
Table D-2. Means for SITP participant middle school stayers and non-SITP stayers
Non-SITP stayers SITP stayers t score
p-value
Total Observations 15,341 2,226
Mean
SD
Mean
SD
Male 0.509 0.500 0.513 0.500 -0.398 0.691
Female 0.491 0.500 0.487 0.500 0.398 0.691
Black 0.172 0.378 0.068 0.252 12.610** 0.000
Hispanic 0.600 0.490 0.807 0.395 -19.072** 0.000
Asian 0.150 0.357 0.105 0.307 5.584** 0.000
White 0.060 0.238 0.011 0.105 9.581** 0.000
Other Race 0.018 0.133 0.009 0.092 3.279** 0.001
English Learner 0.647 0.478 0.803 0.398 -14.745** 0.000
Immigrant 0.141 0.348 0.155 0.362 -1.695 0.090
Refugee 0.008 0.091 0.001 0.037 3.537** 0.000
Special Education 0.183 0.387 0.182 0.386 0.140 0.889
2
nd
Grade ELA z-score 0.066 1.010 -0.080 0.917 6.462** 0.000
2
nd
Grade Math z-score 0.075 1.000 0.021 0.937 2.407* 0.016
** p<0.01, * p<0.05
47
Table D-3. Means for SITP participant high school leavers and non-SITP leavers
Non-SITP leavers
4,064
SITP leavers
493
t score
p-value
Total Observations
Mean
SD
Mean
SD
Male 0.509 0.500 0.487 0.500 0.945 0.345
Female 0.491 0.500 0.513 0.500 -0.945 0.345
Black 0.247 0.431 0.124 0.330 6.124** 0.000
Hispanic 0.518 0.500 0.734 0.442 -9.164** 0.000
Asian 0.109 0.311 0.103 0.305 0.342 0.732
White 0.115 0.319 0.028 0.166 5.925** 0.000
Other Race 0.011 0.106 0.010 0.100 0.235 0.815
English Learner 0.450 0.498 0.686 0.465 -10.017** 0.000
Immigrant 0.143 0.350 0.196 0.397 -2.629** 0.009
Refugee 0.008 0.089 0.000 0.000 1.727 0.084
Special Education 0.149 0.356 0.154 0.361 -0.282 0.778
2
nd
Grade ELA z-score 0.074 1.020 -0.048 0.903 2.535* 0.011
2
nd
Grade Math z-score 0.053 0.991 0.009 0.926 0.922 0.357
7
th
Grade ELA z-score 0.035 0.987 0.091 1.083 -0.578 0.563
7
th
Grade Math z-score -0.068 0.965 0.148 1.068 -2.143* 0.032
Middle school suspensions 0.692 1.932 1.049 2.617 -2.587** 0.010
Middle school expulsions 0.038 0.221 0.053 0.260 -0.984 0.325
Retained 6
th
-8
th
Grade 0.075 0.263 0.023 0.149 2.563* 0.010
** p<0.01, * p<0.05
Table D-4. Means for SITP participant high school stayers and non-SITP stayers
Non-SITP stayers SITP stayers t score
p-value
Total Observations 10,116 1,510
Mean
SD
Mean
SD
Male 0.507 0.500 0.520 0.500 -0.932 0.352
Female 0.493 0.500 0.480 0.500 0.932 0.352
Black 0.176 0.381 0.060 0.237 11.574** 0.000
Hispanic 0.588 0.492 0.807 0.395 -16.552** 0.000
Asian 0.159 0.365 0.113 0.316 4.647** 0.000
White 0.061 0.240 0.011 0.106 7.990** 0.000
Other Race 0.016 0.125 0.009 0.096 1.978* 0.048
English Learner 0.640 0.480 0.817 0.387 -13.672** 0.000
Immigrant 0.130 0.336 0.136 0.343 -0.670 0.503
Refugee 0.010 0.101 0.001 0.037 3.392** 0.001
Special Education 0.178 0.382 0.179 0.384 -0.155 0.877
2
nd
Grade ELA z-score 0.059 1.012 -0.094 0.918 5.527** 0.000
2
nd
Grade Math z-score 0.076 1.005 0.028 0.931 1.767 0.077
7
th
Grade ELA z-score 0.125 0.992 0.128 0.985 -0.103 0.918
7
th
Grade Math z-score 0.102 1.014 0.115 1.018 -0.412 0.680
Middle school suspensions 0.831 2.145 0.849 2.405 -0.291 0.771
Middle school expulsions 0.027 0.174 0.031 0.184 -0.802 0.423
Retained 6
th
-8
th
Grade 0.025 0.156 0.007 0.082 4.398** 0.000
** p<0.01, * p<0.05
48
Appendix E. Short-term outcomes, stratified by immigrant and refugee status (1996-2012, omit 2000)
Dep. Variable Take ELA Test ELA Z Score Take Math Test Math Z Score
Immigrant/ (1) (2) (3) (4) (5) (6) (7) (8)
Refugee Status: Immigrant Refugee Immigrant Refugee Immigrant Refugee Immigrant Refugee
SITP*Rosa 0.014 0.223* 0.028 -0.621 0.009 0.240* 0.085* -0.350
(0.013) (0.092) (0.031) (0.358) (0.011) (0.101) (0.033) (0.199)
Rosa -0.028 1.048 0.318 -0.584 -0.027 1.326* 0.303 0.695
(0.075) (0.667) (0.199) (0.847) (0.071) (0.571) (0.196) (0.464)
SITP*Hamilton 0.004 0.131 0.043 -0.219 -0.004 0.097 -0.102** -0.192
(0.016) (0.086) (0.042) (0.446) (0.016) (0.075) (0.035) (0.143)
Hamilton 0.008 1.038 0.301* -0.433 0.016 1.213* 0.297* -0.260
(0.053) (0.537) (0.138) (0.427) (0.050) (0.478) (0.125) (0.398)
Constant 0.872* 1.031 -2.848** 3.271 0.787* 1.025 -4.418** -5.843**
(0.370) (1.028) (0.643) (2.690) (0.300) (1.047) (0.620) (1.685)
Observations 10,665 490 10,047 460 10,665 490 10,075 461
R-squared 0.498 0.576 0.837 0.825 0.476 0.592 0.824 0.828
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
School FE Yes Yes Yes Yes Yes Yes Yes Yes
Grade FE Yes Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes Yes
Clusters 22 22 22 22 22 22 22 22
Robust standard errors in parentheses. FE = Fixed Effect.
Models also control for current EL status, special education status, moved schools, % of student body eligible for free lunch, total enrollment, charter school, and
health clinic. Total SITP weeks is the cumulative number of weeks of SITP programming a student has received in each year.
** p<0.01, * p<0.05
49
Appendix E (continued). Short-term outcomes, stratified by immigrant and refugee status
(1996-2012, omit 2000)
Dep. Variable Days Absent Suspensions Held Back 3/5
Immigrant/ (9) (10) (11) (12) (13) (14)
Refugee Status: Immigrant Refugee Immigrant Refugee Immigrant Refugee
SITP*Rosa -0.326* 1.536 -0.029** -0.134 0.178* -
(0.146) (2.148) (0.007) (0.256) (0.083) -
Rosa -0.862 16.00 0.022 0.343 -0.158 -
(1.094) (13.64) (0.055) (1.110) (0.080) -
SITP*Hamilton -0.016 0.215 -0.045* -0.230 0.023** -0.068
(0.408) (1.847) (0.019) (0.186) (0.007) (0.074)
Hamilton -0.748 10.69 -0.038 -0.017 -0.011 -0.999
(0.849) (10.40) (0.044) (0.970) (0.012) (0.581)
Constant 8.352* -4.328 -0.222 -0.652 -3.962** -6.958**
(3.220) (9.796) (0.236) (0.572) (0.387) (0.889)
Observations 10,652 484 10,665 490 7,302 304
R-squared 0.667 0.673 0.500 0.629 0.711 0.830
Year FE Yes Yes Yes Yes Yes Yes
School FE Yes Yes Yes Yes Yes Yes
Grade FE Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes
Clusters 22 22 22 22 22 22
Robust standard errors in parentheses. FE = Fixed Effect.
Models also control for current EL status, special education status, moved schools, % of student body eligible for free lunch total
enrollment, charter school, and health clinic. Total SITP weeks is the cumulative number of weeks of SITP programming a
student has received in each year.
** p<0.01, * p<0.05
50
Chapter 3
Prison from the Outside:
Evaluating POPS the Club
B. Danielle Williams
University of Southern California
Please direct correspondence to:
Danielle Williams
Price School of Public Policy
University of Southern California
Ralph & Goldy Lewis Hall
Los Angeles, CA 90089
bdwillia@usc.edu
Abstract
The dramatic increase in incarceration in the United States has resulted in a growing number of
minor youth and children who experience parental incarceration and an increased need for families
and schools to deal with incarceration’s impact on children. However, few evaluations of programs
that specifically aim to assist these children have been performed. This study contributes to the
nascent literature by evaluating POPS (Pain of the Prison System) the Club, an innovative program
established at a Los Angeles high school in 2013 to help adolescents facing the challenges
associated with engagement with the prison system. POPS the Club was formed with the intention
of providing support to students who have personally dealt with the criminal justice system or who
have incarcerated family members. The club offers students a venue to safely discuss personal
issues related to incarceration and the opportunity to express themselves creatively through writing
and other means. This study uses data provided by the Los Angeles Unified School District and a
student-level fixed effects model to estimate the academic and behavioral outcomes associated
with participation in POPS.
Key words: incarceration, parental incarceration, familial incarceration, school program,
academic outcomes, behavioral outcomes, high school outcomes, fixed effects analysis
51
1. Introduction
While the total number of individuals incarcerated in the United States has declined from
the 2008 high of 2.31 million to just over 2.12 million in 2018, nearly 1 in 121 adults remain
incarcerated (Maruschak & Minton, 2020). When individuals on parole or probation are
included, nearly 1 in 40 adults are actively under some form of correctional supervision
(Maruschak & Minton, 2020). The high numbers of adults experiencing incarceration and/or
community supervision have resulted in a growing number of families who are impacted by the
incarceration of a family member or loved one. In 2008, nearly 2.7 million minor youth and
children were estimated to have an incarcerated parent in prison or jail (Pew Charitable Trusts,
2010), and as of 2011-2012, over 5 million youth and children were estimated to have
experienced the incarceration of a parent who resided in their home at some point during their
childhood (Murphey & Cooper, 2015). There is also compelling evidence that these statistics are
underestimated (Schlafer et al.,, 2019). Furthermore, the number of minor youth and children
whose lives have been touched by incarceration further increases if we include minors whose
non-residential parents have ever been incarcerated, whose parents are not incarcerated but are
under other forms of correctional supervision, or who have been impacted by non-parental
familial incarceration, but no estimates that include all of these other forms of familial carceral
experiences exist.
Unsurprisingly, research on the impact of incarceration on minor youth and children
typically focuses on the outcomes associated with parental incarceration rather than other forms
of familial incarceration.
17
Some of these outcomes are influenced by family specific
circumstances including the gender of the incarcerated parent (Cho 2009a, 2000b; Turney &
17
For an overview of existing literature see Turney & Goodsell, 2018.
52
Wildeman, 2015; Wildeman & Turney, 2014), the child’s race (Hinojosa, M.S., Hinojosa, R.,
Bright, & Nguyen, in press), the age of the child when first experiencing parental incarceration
(Cho, 2010; 2011; Turney, 2017) the absence of the parent in the home prior to incarceration
(Geller, Cooper, Garfinkel, Schwartz-Soicher, & Mincey, 2011; Turney & Wildeman, 2013),
family instability (Turney & Wildeman, 2013), and the presence of other negative parental
behaviors, including behaviors linked to a higher likelihood of incarceration (Giordano, Copp,
Manning, & Longmore, in press; Turney, 2017; Turney & Wildeman, 2015; Whitten, Burton,
Tzoumakis, & Dean, 2019) and the removal of fathers who engage in domestic violence
(Wildeman, 2010). However, there exists a growing body of evidence that supports the idea that
children who experience parental incarceration are more disadvantaged (Geller, Garfinkel, &
Western, 2011; Turney & Schneider, 2016) and experience an increased risk of a variety of
negative outcomes (Bryan, 2017; Davis & Schlafer, 2017; Geller, Garfinkel, Cooper, & Mincey,
2009; Kopak & Smith-Ruiz, 2015; Lee, Fang, & Luo, 2013; Roettger & Boardman, 2012;
Turney & Goodsell, 2018; Wakefield & Wildeman, 2013; Whitten, Burton, Tzoumakis, & Dean,
2019; Wildeman, 2010) including negative school outcomes (Cho, 2010; Cho, 2011; Hagan &
Foster, 2012; Haskins, 2014; Hinojosa, M.S., Hinojosa, R., Bright, & Nguyen, in press; Trice &
Brewster, 2004; Turney & Haskins, 2014).
Despite the negative outcomes associated with parental incarceration, a limited number of
programs exist that are specifically aimed at assisting youth and children struggling with
experiences of familial incarceration. Of those that do exist, very few take place within a school
setting. This paper is intended to be a preliminary examination of the potential for school-based
programs to assist youth impacted by incarceration through an evaluation of POPS the Club at
Venice High School in Los Angeles.

53

2. Description of the Program
POPS the Club (POPS)
18
was formed in February of 2013 with the goal of providing
support to students impacted by incarceration: either through the current or previous
incarceration of a loved one or through their own experiences with the criminal justice system.
The club meets once a week during the students’ lunch period, and lunch is provided for all
students who attend. In its earliest days, the club was akin to small group therapy with students
discussing the impact that incarceration had on their lives. However, as the number of students
attending the club increased from 18 students to 91 in the second year, the English teacher who
acts as the club’s faculty sponsor began to emphasize creative writing as a means for the students
to share their stories. The students who participated in POPS in the spring of 2013 were
encouraged to submit their work to an unaffiliated student anthology. In subsequent years, the
club produced its own anthology featuring the students’ writing and artwork and hosted an
annual reading of the students’ works at a local theater. Periodically throughout the year,
speakers who are in some way connected to the issue of incarceration and criminal justice reform
visit the club during their meetings. As of December 2020, POPS had expanded to an additional
seven high schools in the Los Angeles area as well as to eleven high schools in four other states
(POPS the Club, 2020a).
3. Literature Review
The literature on the impact of parental incarceration on student outcomes, and indeed on
child and youth outcomes in general, is plagued by the number of confounding adverse factors
experienced in conjunction with parental incarceration, e.g. poverty, lower levels of parental
education, addiction, etc. (for further discussions of the limitations of existing studies,

18
POPS stands for “Pain of the Prison System” but the official name of the organization is “POPS the Club.”

54

particularly in regards to selection bias, see Johnson & Easterling, 2012). In 2012, Murray,
Farrington, and Sekol conducted a meta-analysis of 40 studies that examine educational
performance and other child outcomes associated with parental incarceration. 28 of these studies
used samples drawn either nationally or from locations within the United States, with the
additional 12 using samples from six other Western countries. Among the studies that were most
rigorous and controlled for at least some of the aforementioned confounding issues, there was
little to no evidence of an association between parental incarceration and educational outcomes.
It should be noted, however, that the educational performance metrics used focused on
standardized test scores or tests of cognitive ability rather than other measures of successful
school performance, such as grades, graduation rates, and behavioral outcomes (Murray,
Farrington, & Sekol, 2012). Despite the meta-analysis’s findings, there is other evidence that
suggests that parental incarceration has negative impacts on schooling outcomes, though these
impacts are by no means fully understood or uniform along all dimensions of parental
incarceration.
Paternal incarceration has been found to be associated with an increased likelihood of
grade retention (Hinojosa, M.S., Hinojosa, R., Bright, & Nguyen, 2019; Turney & Haskins,
2014), and lower grades and educational achievement (Hagan & Foster, 2012). The sons of
incarcerated fathers are also more likely to be unprepared for school in terms of non-cognitive
readiness factors such as their behavior and attention spans, and as a result, are more likely to
receive a special education assignment by the age of nine (Haskins, 2014).
Studies focusing on children who experience maternal incarceration have produced more
divergent findings. Earlier work indicated an association between maternal incarceration and
increased school absences and disciplinary issues as well as an increased risk of failing among
55
adolescent children (Trice & Brewster, 2004). Cho’s more recent research (2009a, 2009b, 2010,
2011), which attempts to disentangle the effect of long-term prison incarceration on child
outcomes from that associated with arrest and short-term detainment, events that occur in tandem
with long-term incarceration, has found that maternal incarceration did not impact grades (Cho,
2009a) and actually decreased the likelihood of retention (Cho, 2009b). Maternal incarceration
did, however, increase the likelihood of dropping out (Cho, 2010; 2011).
While the vast majority of the literature focuses on the impact of parental incarceration,
there is additional research that suggests that a non-parental residential family member’s
incarceration may also have a negative impact on school absences and graduation rates (Nichols
& Loper, 2012). This is especially important to note within the context of the POPS program,
since the students who attend POPS may be dealing with the incarceration of siblings or other
non-parental family members.
Further complicating an understanding of how experiencing familial incarceration may
negatively affect students is additional research that indicates that teachers may stigmatize the
children of the incarcerated and perceive them as less competent than their peers (Dallaire,
Ciccone, & Wilson, 2010). Students in schools with a high percentage of students with
incarcerated parents may suffer from incarceration “spillovers” even if they themselves do not
have an incarcerated parent (Hagan & Foster, 2012).
Despite evidence for the negative schooling outcomes associated with familial
incarceration, few programs exist that specifically aim to assist minors who are struggling with
experiences of familial incarceration. Those that do exist generally take place outside of the
context of school and focus on improving incarcerated parents’ parenting skills, facilitating
contact while parents are incarcerated, or providing family therapy (Parke & Clarke-Stewart,
56
2003; see Hoffmann, Byrd, & Kightlinger, 2010 and Johnston, 2012 for examples and
descriptions of these types of programs). Of the few that do take place in schools, most are
programs that match children incarcerated parents with mentors who then visit them at their
schools.
19

Furthermore, few evaluations of these programs, regardless of whether they are based in
or outside of schools, have been conducted (see Conway & Keays, 2015). I know of only one
evaluation of a school-based, non-mentor program to assist children experiencing familial
incarceration (Springer et al., 2000). While the evaluation found positive effects on participants’
self-esteem, the program evaluated was short in duration, focused on elementary school aged
children, and consisted of a small sample of only ten Hispanic children. This study contributes
to the nascent evaluation literature on these programs by examining the short-term academic and
behavioral outcomes associated with students’ participation in POPS.
4. Theory
As mentioned in the literature review, childhood and adolescent experiences of familial
incarceration are usually confounded by multiple factors that also negatively impact child
outcomes. However, there are various mechanisms unique to familial incarceration that may
also negatively impact minors. These potential mechanisms include the trauma of witnessing an
arrest, being physically separated from the loved one during his or her incarceration, and
increased contact with the penal system (Braman, 2004; Comfort 2007); the emotional and
economic stress that occurs during and frequently after the incarceration (Hagan & Dinovitzer,
19
For an extensive list of programs serving the children of incarcerated parents and their descriptions see The
National Resource Center on Children and Families of the Incarcerated. The International Coalition for Children
with Incarcerated Parents also lists a few additional programs.

57

1999; Swisher & Waller, 2008); and the stigma associated with incarceration (Murray &
Farrington, 2008).
While participating in the POPS club is unlikely to directly ameliorate the trauma or
stressors that result from familial incarceration or other negative factors that may also occur
within the home, it may positively impact students’ schooling outcomes by reducing stigma if
these children are in fact stigmatized by the incarceration of a loved one. While there is some
evidence that the children of the incarcerated experience stigma (Dallaire, Ciccone, & Wilson,
2010; Phillips & Gates, 2011; Murray & Farrington, 2008), the stigmatization is generally taken
for granted rather than explicitly documented (Phillips & Gates, 2011). Furthermore, a loved
one’s incarceration is a concealable stigma, which means that students with incarcerated loved
ones may endure different types of stigma: anticipated/perceived stigma - “the degree to which
individuals expect that others will stigmatize them if they know about the concealable
stigmatized identity” (Quinn & Chaudoir, 2009, p. 636); self-stigma – the internalization of
negative attitudes associated with the stigmatized group (Phillips & Gates, 2011); and
experienced/discriminatory stigma – the treatment that occurs when others are aware of the
stigma (Phillips & Gates, 2011). The types of stigma experienced depend on whether the
students have revealed their relationship to the incarcerated (or have had it revealed for them),
and it is possible for students to experience all three types of stigma simultaneously as they may
have partially revealed their relationship to certain individuals while concealing it from others.
Each type of stigma may potentially impact student outcomes differently.
POPS participation may mediate stigma’s effect on student outcomes in two ways.
While the impact of the stigma associated with the incarceration of a loved one is not fully
known, stigma in general is associated with negative academic and health outcomes (Major &
58
O’Brien, 2004). Moreover the presence of other types of concealable stigmas, such as having an
eating disorder or mental illness, familial poverty status, or sexual orientation, have been
associated with increased depression and illness (Quinn & Chaudoir, 2009) and lower self-
esteem (Frable, Platt, & Hoey, 1998). When compared to those with conspicuous stigmas, those
with concealable stigmas tended to spend more time alone and feel more comfortable in social
settings when individuals who share their stigma are present (Frable, Platt, & Hoey, 1998).
Thus, by providing a supportive space filled with other students who have also experienced
familial incarceration, participation in POPS may mediate students’ experiences of stigma,
particularly self-stigma and the associated feelings of shame and isolation. This in turn may
increase student engagement in school, resulting in better behavioral or academic outcomes. The
very act of choosing to participate in POPS seems to lend credence to this idea. If adolescents
with a family history of incarceration do experience stigma, whether self-inflicted or imposed on
them, why would they choose to reveal a concealable stigma, especially when they are of an age
when they are best poised to control their personal narratives both among their peers and their
teachers? The perceived benefits of affiliation and support must be expected to outweigh the
social cost of revealing oneself to be a member of a stigmatized group. Secondly, the existence
of POPS and its ability to put a positive and engaged “face” on students who are dealing with
familial incarceration may reduce students’ experienced stigma by changing the attitudes of
faculty and peers and reducing the discrimination experienced due to their disapprobation. This
may in turn also improve student outcomes.
Finally, POPS may improve academic achievement through a much more direct means.
The club emphasizes creative writing, and the students frequently work on writing assignments
59
during meetings and outside of school.
20
This additional work may effectively amount to extra
class time. The extra work may improve the students’ technical writing skills and provide a
means for them to better engage with their schooling.
Of course, as POPS is an extracurricular club that meets only once a week, it is quite
conceivable that it may have no measurable impact on these very concrete measures of academic
achievement. In this case, however, a null finding should also be seen as a positive, if not as
exciting, outcome. Given that POPS students are especially vulnerable due to their familial
circumstances, school administrators may have concerns regarding putting similarly situated
students together socially. In fact, there is evidence that adolescent children who have a recently
incarcerated father are more socially isolated than their peers and also have friends who are less
academically successful and more likely to evince delinquent behaviors, including substance use,
fighting, and skipping school (Bryan, 2017). Concentrating these students socially through their
participation in the club could serve to exacerbate these issues through negative peer effects. A
null finding would provide a counterpoint to this fear.
5. Research Design
I use a student fixed effects model to examine the question of whether participation in
POPS is associated with positive academic and behavioral outcomes of students. The research
design compares the change in the students’ outcomes before and after POPS participation to the
change in the non-participating Venice students’ outcomes over the same period.
20
Creative writing and, to a lesser extent, visual arts were emphasized during the years covered in the analysis. In
2017, POPS updated their curriculum to focus on three areas: self-expression (through writing, visual arts, and
performance), self-empowerment (including techniques to reduce stress with a special emphasis on mindfulness),
and community engagement (creating videos of the guest speakers who visit the club to build a collection of
resources for the community at large) (POPS the Club, 2019).
60
Short-term academic and behavioral outcomes are modeled as a function of participation
in POPS, time-varying student characteristics, a student fixed effect, and a middle school fixed
effect using individual-level panel data as shown in the following equation:
Yist = α + β POPS t + γ STit + ηi + μs + εist
Yist = the dependent variable for student i who attended middle school s at time t
POPSt = dichotomous variable indicating whether the student participated in POPS in a given
year
STit = vector of time-varying student characteristics
ηi = student fixed effect
μs = middle school fixed effect
εist = error term
The inclusion of the student fixed effect is intended to control for the students’ time invariant
personal characteristics that may ultimately impact academic achievement or behavior. The
inclusion of the middle school fixed effect is intended to control for differences in the quality of
education the students received immediately prior to high school.
Data
The analysis is conducted using student-level administrative data provided by the Los
Angeles Unified School District (LAUSD). The data contains demographic, academic, and
behavioral information on all students who attended Venice High School at some point between
the 2011-2012 school year (the year prior to the start of POPS) and January 2015. The data also
contains eighth grade records for these students. While the data contain school codes for the
middle school attended by the students during their 8
th
grade year, allowing me to include a
middle school fixed effect, they do not contain school codes any other high schools attended by
the students either before or after their enrollment at Venice High. There is also no information
that identifies which students have experienced familial incarceration aside from participation in
POPS.

61

As mentioned above, POPS activities were substantially different during the club’s initial
school year, 2012/2013, than in those that followed. It was more akin to a support group and
only met for approximately four months that year. The emphasis on creative writing wasn’t as
marked until the next year, which was also the first year that the POPS student anthology was
published. Thus, after conducting an initial analysis and estimating an overall treatment effect,
an additional model that excludes the first year of the program is included.
Measures
The data was merged using unique student identification code and school years. From
this data, seven measures were used to examine outcomes associated with participation in the
club. Academic outcomes include language arts and math standardized test scores, fall and
spring gpa, and the likelihood of being retained. The test scores have been standardized into z-
scores with a mean of zero and a standard deviation of one. This was done so that students’ test
performance could be compared to their peers’ test outcomes across years, regardless of what
grade they were in or which standardized test they took.
21
The two behavioral outcome measures
are the number of absences and the number of suspensions during the year of participation.
In addition to controlling for student-level fixed effects, the models include controls for
student characteristics that may change over-time: English language learner status and special
education status. Another time-variant variable, free or reduced price meal status, exists in the
data as well. While free or reduced price meal status is often used as a proxy for students’
household incomes and, particularly in education studies, socioeconomic status and has been
found to be more strongly associated with actual income relative to other proxy variables

21
Beginning in 2007, California has permitted students with an individualized education program (IEP) to take the
California Modified Assessment in lieu of the California Standards Test required for most students (California
Department of Education, 2015).

62

attempting to measure it (Day et al., 2016), there is additional evidence that the measure is
limited in its usefulness as a measure of poverty (Snyder and Muso-Gillette, 2015) and especially
of socioeconomic status (Harwell & Lebeau, 2010). There are also the inherent limitations in
attempting to use it as a poverty measure, namely its dichotomous nature and the fact that there
are multiple pathways by which a student may qualify for the program, not all of which share the
same guidelines. Furthermore, in the specific instance of the POPS data, there is a large
disparity in the number of students who have ever received free or reduced price meals at some
point during or after the eighth grade and the percentage who receive them in any given year.
Approximately 60 percent of students on average received free or reduced price meals in any
given year but of those same students over 80 percent received free or reduced price meals at
some point during these years. This seemingly indicates that a substantial percentage of students
are cycling on and off of the program. One possible explanation for this is that statistics are in
fact accurate representations of a high degree of program churning, which may be due either to
substantial year to year fluctuations in family income or participation in programs that
automatically qualify students to receive free or reduced price meals. However, alternate
explanations unrelated to actual changes in family, such as an unwillingness by older students to
participate in the program potentially due to stigma, cannot be ruled out. There is also a
substantial amount of missing data with this particular variable in the dataset. In some of the
earlier years in the data, over half of the students are missing information regarding their
participation. This combined with the differences in the percentages of students who receive
meals and the general concerns about the measure’s validity in the literature, raises questions
about the measure’s accuracy. For these reasons, versions of the models detailed above that both
include and exclude the free or reduced price meal variable are included in the analysis.

63

A middle school fixed effect, based on the school the student attended in the eighth
grade, is also included as a means of further controlling for students’ pre-high school academic
preparation.
Sample
The sample consists of the eighth grade and high school records of all students who
attended Venice High School at any point during the 2011-2012 school year up until January
2015. If the students attended another high school prior to their enrollment at Venice High or
left Venice High to attend another LAUSD high school, their subsequent records were also
provided up until January 2015. However, there is no information on which schools they
attended aside from Venice High nor is there any means of identifying which students moved.
The sample size for the outcome measures vary due to the reduced number of observations
depending on the year (e.g. only test scores for the 2012/2013 school year and earlier were
available).
6. Results
Descriptive Statistics
Table 1A presents the mean descriptive statistics POPS and non-POPS participants
during the years that POPS existed at Venice High (2012-2013 – Fall 2014). Table 1B presents
student high school outcomes for POPS and non-POPS students in the years before POPS
existed. All of these statistics are calculated using student-year observations. As a result,
students may be present in both the POPS and non-POPS samples of Table 1A according the
years they participated.
As shown in Table 1A, during the period when POPS was available, the demographics of
the student-year observations for both POPS and non-POPS population were majority Hispanic,

64

62 and 67 percent respectively.
22
There were also roughly the same percentage of non-Hispanic
white students in both groups, 10 percent of POPS students verses 13 percent. While POPS and
non-POPS participating students were about equally likely to be Hispanic or non-Hispanic
whites, the POPS student population consisted, at a significant level, of more female (53 percent
verses 43 percent) and black students (26 percent verses 11 percent) compared to their non-POPS
peers. The latter is unsurprising given the higher rates of incarceration experienced by the black
community. There were also fewer Asian POPS students, with just 1 percent of the POPS
student population identified as Asian verses 8 percent of the non-POPS group.
In terms of time-varying student demographic characteristics, only 3 percent of POPS
students were classified as limited English proficient (LEP), while 11 percent of non-POPS
students were LEP. 47.1 percent of POPS students received free or reduced lunch during the
2012-2013 school year and the fall of 2014 while 55.1 percent of non-POPS students received it
during the same time period. Both of these differences were statistically significant, but the free
and reduced lunch rate was only significant at the 0.05 level. POPS and non-POPS students
were similar in terms of the percentage of students with special education classifications during
this period.
Students who participated in POPS had both lower test scores and lower GPAs during
their eighth-grade year than their non-POPS peers. Their math scores were 0.17 standard
deviations lower and their language arts scores were 0.21 standard deviations lower, the latter of

22
In this discussion of the descriptive statistics, the term “student population” is used for simplicity’s sake.
However, as noted above, the descriptive statistics are based on student-year observations rather than a single
observation per student. Thus students who appear more frequently during the sample years are included at a greater
rate than those with fewer records and impact the measurement of the demographic percentages accordingly. For
Table 1A this also means that students are only counted as POPS students for the number of years they participated
and thus may also appear in the non-POPS group during the years they didn’t participate if there are records for
them in those years.

65

which was statistically significant at the 0.05 level. Both their fall and spring GPAs during the
eighth grade were statistically significantly lower than their peers by approximately two-tenths of
a point, though the spring GPA was only significant at the 0.05 level.
The underperformance of POPS students persisted in high school in the years prior to the
implementation of POPS (Table 1B) with the disparity in achievement between POPS and non-
POPS students growing, particularly in terms of their math test scores substantially. Those who
eventually participated in POPS scored 0.44 and 0.26 standard deviations lower than their non-
POPS peers on their math and language arts tests respectively. The fall and spring GPAs in high
school dropped to almost four-tenths of a point lower than their peers, and their retention rate
was 16 percentage points higher, nearly twice that of students who never participated in POPS.
However, the number of absences and the rate of suspensions between the two groups were
essentially the same.
POPS students’ academic outcomes during the years they participated in POPS (Table
1A) were again lower than their peers on average but the difference in their test scores were no
longer meaningful, and the gap in their GPAs was somewhat reduced. Their retention rate was
substantially reduced and there was no difference between it and that of their peers. Their
suspension rates were also the same, but they had a greater number of absences. On average
POPS students were absent 10.41 days in the years they participated in POPS while their
counterparts were only absent 7.82 days.
Student Outcome Models
Table 2A depicts the impact of participating in POPS in the year of participation,
controlling for students’ special education status and English proficiency and includes both
student and middle school fixed effects. Table 2B’s specification is identical to that of Table 2A,

66

except that the additional control variable of receiving free and reduced price meals is also
included in the model. In both of these models, there are no significant effects on any of the
outcomes examined. In the years students participate in POPS, there is no association between
their participation and their test scores, GPAs, retention rates, absences, or suspensions. In fact,
when examining versions of the model with fewer controls, the only outcome associated with
POPS participation at a statistically significant level is the rate of retention (Table 3).
Table 3 looks at the rate of retention across these permutations of the student fixed effects
model. When middle school fixed effects are excluded and only POPS participation is
considered, participation is associated with a 13.6 percent reduction in the likelihood of being
retained as compared to their non-POPS peers. This drops to 8 percent when all of the time-
varying controls are included but remains significant. When middle school fixed effects are
added, the only permutation of the model that shows an association between participation and a
reduction in the likelihood of being retained is the one without time-varying controls.
Finally, a model excluding the first year of POPS participation is presented in Table 4.
As mentioned above, POPS’s format in its first year was markedly different from the subsequent
years and participation was also for a shorter period of time. Table 4A presents the results
controlling for students’ special education status and English proficiency and includes both
student and middle school fixed effects. Table 4B is the same but also includes the free or
reduced price lunch control variable. There is no evidence in either model that associates
participation in POPS with any significant effects on GPA, retention likelihood, absences, or
suspensions.
It should be noted that, in regards to the test scores, no estimates are available for the
second and third years of POPS. The most recent test scores available in the data were for the

67

2012/2013 school year, the first year POPS existed, so the estimates in Tables 2A and 2B solely
represent the impact associated with having participated in POPS that year. Consequently, given
that only 18 students participated during the club’s inaugural year and that the California
Standards Test/California Modified Assessment isn’t administered to students past the eleventh
grade, the estimates are based on very few observations: 7 language arts scores and 8 math
scores.
Dosage model
The length of time a student participates in POPS may magnify the outcomes associated
with joining. However, while at the beginning of the analysis I intended to estimate dosage
models in order to determine if the duration of participation resulted in different outcomes, upon
further inspection, there were too few students who had participated in multiple years. Only 13
students participated in POPS for two years and only 5 participating in all three years, which is
perhaps unsurprising given the newness of the club (See Appendix A).
Additional Limitations
Internal Validity Concerns
The most significant limitation of this data is its inability to distinguish which students
within the control group have been impacted by familial incarceration and which haven’t.
Without the ability to identify them and construct a true comparison group, the estimates
produced by the models are biased. If we assume that students who do not face the challenges
unique to the experience of familial incarceration are better off than those who do, then the
effects associated with participation in POPS have likely been underestimated. However, as

68

mentioned in the literature review, the evidence as to whether or not the experience of familial
incarceration leads to negative outcomes for impacted children and adolescents is mixed.
There is an additional, more minor, sample cross-contamination issue due to the
possibility that some students who join POPS may not have experienced familial incarceration
but attend anyway in order to support their friends or significant others who are affected. These
students differ both from their fellow POPS members and their peers who are not impacted by
familial incarceration. Even if there were data on students who participated but had made it
known that they were not affected first-hand, the construct of the club as it currently exists would
make it impossible to identify all of them. For while the club is clearly aimed at the students who
have been impacted by incarceration, students are not required to volunteer any information
regarding their reason for attending or the nature of their relationship to incarceration.
Selection Bias
As with most of the existing literature on the impact of the incarceration of family
members or other familial involvement or contact with the criminal justice system on child
outcomes, selection bias is a major impediment to rigorous analysis in this paper. Students who
have been impacted by familial incarceration but choose not to participate in POPS
fundamentally differ from those who do. There is an additional dimension to the selection bias
issue in that students who were already at Venice when POPS began but who chose to participate
in later years may differ from those who joined earlier or upon attending the high school. As the
club became more established at the high school and more students joined, students may have
perceived that the stigma associated with “outing” themselves as being impacted by familial
incarceration was reduced. To the extent that this stigma impacts students’ decisions to
69
participate, some students who didn’t participate in the first year or two may have decided to join
POPS in subsequent years as a result of the reduced stigma.
Diffusion
Diffusion is also a concern due again to the element of stigma and its potential to isolate
and negatively impact students’ mental health. Students who are impacted by familial
incarceration but chose not to participate may have benefited from POPS simply by learning of
its existence. Knowing that other students are dealing with the issue of familial incarceration and
are willing to publicly identify themselves as such may make students feel less alone in the
circumstances of their own life even if they are not willing to do so themselves. They may feel
less stigmatized and be more willing to speak to friends or teachers about their situation.
Furthermore, if the presence of the club at the school does reduce stigma among teachers and
peers towards students who have dealt with familial incarceration, then students benefit if others
are aware of their situation regardless of their POPS participation.
External Validity
Given the limitations of the data, particularly the fact that it only contains information
regarding the first three years of POPS participation, a period during which the form and mission
of the club were still evolving, the results should not be generalized to other POPS clubs. Those
clubs were founded in years subsequent to this analysis at which point a concerted effort had
been made to develop a specific structure for the club that could be replicated at other schools.
Furthermore, there are likely significant differences in the characteristics of these schools. In
terms of geography alone they differ greatly; many of the clubs are located in Southern
California but there are also clubs located as far away as Georgia and Alaska.
70
In fact, the timing of the data may also render these results non-representative of even the
POPS club at Venice High School itself. Aside from the differences between the first and
second years of implementation which are taken into consideration in the Table 4 model, the club
may have undergone additional changes in the early years before fully establishing a consistent
format. These changes may have been less pronounced than the difference between year one and
two but may nonetheless result in estimates that don’t accurately reflect the impact associated
with participating in the current incarnation of Venice High’s POPS club.
7. Discussion and Policy Implications
The results of these models provide no evidence to support the hypothesis that students
who joined the Pain of the Prison System club experienced improvement in academic or
behavioral outcomes compared to their non-POPS peers.
Despite the null findings and the significant limitations of the analyses themselves, there
are still encouraging takeaways from the analyses. Foremost is that while the preliminary
evidence does not support the idea that POPS participation results in better outcomes, there is
also no evidence that indicates that it is associated with negative student outcomes. This is
particularly noteworthy as educators may be leery of clubs or programs that encourage students
with familial histories of incarceration to socialize with each other. Due to the stigma associated
with incarceration, they may perceive students from these families as being at a higher risk of
behavior problems or other negative outcomes and worry that interaction with similarly situated
peers could compound these deleterious effects. As a result, school administrators may be
inclined to discourage the formation of POPS clubs or other similar school-based programs that
focus on this subset of the student population. This study provides preliminary reassurance that,
at least in certain contexts, this is not the case. Though it falls short of justifying the existence of
71
clubs and programs of this nature, it does offer some tenable support for their creation, which
students and teachers who endeavor to form them at their schools may need in order to offset the
concerns of school administrators.
Additionally, the comparison of the high school outcomes of POPS and non-POPS
students in the years prior to the existence of POPS indicates that there are substantial
achievement differences between the two groups. Students who subsequently joined POPS
underperformed on all academic measures: test scores, GPA, and retention rates. In fact, in the
latter case, the likelihood of eventual POPS students being retained in the years prior to the
implementation of POPS was nearly double that of students who would ultimately never
participate in POPS, 34 percent verses 18 percent. This provides supporting evidence that there
are systematic gaps in the educational outcomes between students who experience familial
incarceration and those who don’t. From a policy perspective, these differences suggest that
school districts and educators should endeavor to identify these at-risk students and design and
implement interventions (whether in the form of extracurricular activities or other school-based
outreach programs) that provide support and additional resources to them. Furthermore, it also
indicates that outside groups that offer programs designed to assist children and adolescents
dealing with familial incarceration should consider investing in programmatic elements that
specifically encourage better school performance among these children or even liaising directly
with schools.
Additional research should consider alternative measures of impact as school outcomes
may not be the most appropriate measures to gauge the impact of joining POPS. While school
outcomes are more readily measured and perhaps more useful when it comes to advocating for
the clubs’ establishment within schools, given the inherent social aspect of joining the club,
72
participation may have more of an impact on interpersonal development, internalizing behaviors,
mental health (especially feelings of isolation), and friendship networks. Future research should
incorporate measures that examine these aspects. Finally, while it would be difficult to conduct
research that substantially addresses the internal validity issues, incorporating data from the other
POPS clubs would at least strengthen the external validity of the results.
73
8. Tables
Table 1A. Descriptive Statistics, Overall Means for POPS Participants and Non-POPS Students
during POPS program years (2012/13 - Fall 2014), Student-Year Observations
Variable POPS Students Non-POPS Students
Number of Student-Year Observations 189 6,927
Student Characteristics

Male 0.43** 0.53
Female 0.57** 0.47
Black 0.26** 0.11
Hispanic 0.62 0.67
Asian 0.01** 0.08
White 0.10 0.13
Other Race 0.00 0.01
Limited English Proficient 0.03** 0.11
Special Education 0.13 0.13
% Eligible for Free/Reduced Lunch 47.1%* 55.1%
8
th
Grade Achievement Characteristics
8
th
Grade ELA z-score -0.18* 0.03
8
th
Grade Math z-score -0.13 0.04
8
th
Grade Fall GPA 2.49** 2.69
8
th
Grade Spring GPA 2.49* 2.67
High School Academic Outcomes
ELA z-score -0.57 0.03
Math z-score -0.23 0.02
Fall GPA 2.12** 2.36
Spring GPA 2.01** 2.32
Retention 0.10 0.13
High School Behavioral Outcomes
Days Absent 10.41** 7.82
Suspensions 0.005 0.006
** p<0.01, * p<0.05
74
Table 1B. Descriptive Statistics, Overall Means for POPS Participants and Non-POPS
Students prior to the implementation of POPS (before 2012/13), Student-Year
Observations
Variable POPS Students Non-POPS Students
Number of Student-Year Observations 188 5,745
Student Characteristics

Male 0.52 0.53
Female 0.48 0.47
Black 0.21** 0.01
Hispanic 0.67 0.68
Asian 0.02** 0.08
White 0.10 0.13
Other Race 0.00 0.01
Limited English Proficient 0.09 0.14
Special Education 0.07 0.13
% Eligible for Free/Reduced Lunch 60.8% 59.0%
Pre-POPS High School Academic
Outcomes
ELA z-score -0.20** 0.06
Math z-score -0.33** 0.11
Fall GPA 2.05** 2.42
Spring GPA 2.02** 2.39
Retention 0.34** 0.18
Pre-POPS High School Behavioral
Outcomes
Days Absent 8.78 9.30
Suspensions 0.04 0.04
** p<0.01, * p<0.05
75
Table 2A. Outcomes associated with POPS participation (excluding the free or reduced price meal variable)
(1) (2) (3) (4) (5) (6) (7)
VARIABLES ELA Z score Math Z score Fall GPA Spring GPA Retention Days Absent Suspensions
POPS -0.196 0.121 -0.055 -0.100 -0.0497 1.396 -0.0332
(0.324) (0.440) (0.047) (0.062) (0.0422) (0.855) (0.0196)
Special Education 0.165 0.196 0.0578 0.159 -0.0958 1.494 -0.135**
(0.141) (0.188) (0.0722) (0.0816) (0.0552) (1.436) (0.0288)
Limited English -0.0634 0.114 0.106** 0.111** -0.0522* -2.677** 0.0255
Proficiency (0.0692) (0.0856) (0.0364) (0.0382) (0.0259) (0.765) (0.0151)
Constant 0.0532** 0.0196 2.456** 2.434** 0.110** 7.775** 0.0488**
(0.0200) (0.0260) (0.0111) (0.0124) (0.00844) (0.221) (0.00454)
Observations 5,730 5,610 10,614 9,054 9,162 9,663 10,908
R-squared 0.001 0.001 0.002 0.003 0.002 0.003 0.004
Middle School FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Standard errors in parentheses
** p<0.01, * p<0.05
76
Table 2B. Outcomes associated with POPS participation (including the free or reduced price meal variable)
(1) (2) (3) (4) (5) (6) (7)
VARIABLES ELA Z score Math Z score Fall GPA Spring GPA Retention Days Absent Suspensions
POPS -0.190 0.121 -0.0177 -0.0874 -0.0534 1.491 -0.0303
(0.324) (0.440) (0.0463) (0.0614) (0.0421) (0.856) (0.0196)
Special Education 0.162 0.195 0.0704 0.169* -0.0997 1.511 -0.134**
(0.141) (0.188) (0.0711) (0.0808) (0.0551) (1.436) (0.0288)
Limited English -0.0546 0.123 0.0725* 0.0900* -0.0455 -2.776** 0.0228
Proficiency (0.0694) (0.0859) (0.0359) (0.0379) (0.0258) (0.767) (0.0151)
Free or Reduced -0.0463 -0.0455 0.203** 0.184** -0.0569** 0.450 0.0163**
Price Lunch (0.0295) (0.0368) (0.0136) (0.0174) (0.0119) (0.253) (0.00569)
Constant 0.0845** 0.0501 2.328** 2.310** 0.149** 7.497** 0.0385**
(0.0283) (0.0358) (0.0139) (0.0170) (0.0116) (0.271) (0.00578)
Observations 5,730 5,610 10,614 9,054 9,162 9,663 10,908
R-squared 0.002 0.002 0.032 0.023 0.006 0.003 0.005
Middle School FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Standard errors in parentheses
** p<0.01, * p<0.05
77
Table 3. Retention outcomes associated with POPS participation using various student fixed effects based models
(1) (2) (3) (4) (5) (6)
VARIABLES Retention Retention Retention Retention Retention Retention
POPS -0.136** -0.0754* -0.0801* -0.123** -0.0497 -0.0534
(0.0397) (0.0376) (0.0376) (0.0450) (0.0422) (0.0421)
Special Education -0.0644 -0.0684 -0.0958 -0.0997
(0.0546) (0.0545) (0.0552) (0.0551)
Limited English -0.0219 -0.0162 -0.0522* -0.0455
Proficiency (0.0249) (0.0249) (0.0259) (0.0258)
Free or Reduced -0.0537** -0.0569**
Price Lunch (0.0117) (0.0119)
Constant 0.123** 0.107** 0.142** 0.120** 0.110** 0.149**
(0.00279) (0.00815) (0.0112) (0.00293) (0.00844) (0.0116)
Observations 14,411 10,322 10,322 12,718 9,162 9,162
R-squared 0.001 0.001 0.005 0.001 0.002 0.006
Middle School FE No No No Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes
Standard errors in parentheses
** p<0.01, * p<0.05
78
Table 4A. Outcomes associated with POPS participation, excluding the first year of POPS and the free or reduced price
meal control variable.
(1) (2) (3) (4) (5) (6) (7)
VARIABLES ELA Z score Math Z score Fall GPA Spring GPA Retention Days Absent Suspensions
POPS - - -0.0858 -0.127 -0.0369 1.300 -0.0349
(0.0537) (0.0784) (0.0452) (0.925) (0.0225)
Special Education 0.232 0.282 0.108 0.104 -0.0225 -2.894 -0.137**
(0.168) (0.223) (0.0853) (0.105) (0.0600) (1.717) (0.0348)
Limited English -0.0978 -0.0553 0.129** 0.135** -0.0469 -3.250** 0.0225
Proficiency (0.0889) (0.109) (0.0424) (0.0465) (0.0268) (0.868) (0.0177)
Constant 0.0424 0.0278 2.461** 2.463** 0.0973** 8.101** 0.0525**
(0.0249) (0.0320) (0.0132) (0.0158) (0.00915) (0.263) (0.00549)
Observations 3,867 3,821 8,356 6,760 6,836 7,403 8,582
R-squared 0.909 0.874 0.865 0.879 0.571 0.747 0.533
Middle School FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Standard errors in parentheses
** p<0.01, * p<0.05
79
Table 4B. Outcomes associated with POPS participation excluding the first year of POPS and including the free or
reduced price meal variable
(1) (2) (3) (4) (5) (6) (7)
VARIABLES ELA Z score Math Z score Fall GPA Spring GPA Retention Days Absent Suspensions
POPS - - -0.0415 -0.104 -0.0429 1.504 -0.0318
(0.0529) (0.0774) (0.0450) (0.927) (0.0225)
Special 0.230 0.282 0.116 0.113 -0.0269 -2.906 -0.136**
Education (0.168) (0.223) (0.0838) (0.103) (0.0598) (1.715) (0.0348)
Limited English -0.0914 -0.0397 0.0887* 0.103* -0.0371 -3.463** 0.0196
Proficiency (0.0894) (0.109) (0.0418) (0.0460) (0.0268) (0.870) (0.0178)
Free or Reduced -0.0292 -0.0748 0.211** 0.222** -0.0675** 0.832** 0.0156*
Price Meals (0.0419) (0.0517) (0.0163) (0.0234) (0.0136) (0.294) (0.00685)
Constant 0.0620 0.0780 2.330** 2.313** 0.143** 7.602** 0.0428**
(0.0377) (0.0471) (0.0164) (0.0222) (0.0130) (0.316) (0.00694)
Observations 3,867 3,821 8,356 6,760 6,836 7,403 8,582
R-squared 0.909 0.874 0.870 0.883 0.574 0.747 0.533
Middle School FE Yes Yes Yes Yes Yes Yes Yes
Student FE Yes Yes Yes Yes Yes Yes Yes
Standard errors in parentheses
** p<0.01, * p<0.05
80
9. Appendices
Appendix A. POPS Student Participation by Year
School
Year
Grade
9
th
10
th
11
th
12
th
TOTAL
2012/13 5 4 3 6 18
2013/14 10 16 23 42 91
*

Fall 2014 11 10 23 36 80
**
* 8 of the 91 students who participated in the 2013/14 school year had participated the year prior.
** 15 of the 80 students who participated during the fall semester of 2014 participated in the year prior (2013/14).
Of those 15 students, 5 participated in all three years (2012/13 – Fall 2014).
81
Appendix B. Overall Means for POPS Participants by year, Student Observations
Variable 2012-2013 2013-2014 2014-2015
Number of Observations 18 91 80
Student Characteristics

Male 0.56 0.49 0.34
Female 0.44 0.51 0.66
Black 0.39 0.26 0.24
Hispanic 0.44 0.60 0.69
Asian 0.00 0.02 0.00
White 0.17 0.11 0.08
Other Race 0.00 0.00 0.00
Limited English Proficient 0.00 0.04 0.01
Special Education 0.06 0.11 0.16
% Eligible for Free/Reduced Lunch 61.1% 52.7% 37.5%
Pre-POPS Achievement
Characteristics
8
th
Grade ELA z-score -0.08 -0.15 -0.25
8
th
Grade Math z-scores -0.07 -0.08 -0.21
8
th
Grade Fall GPA 2.38 2.39 2.63
8
th
Grade Spring GPA 2.42 2.37 2.62
High School Academic Outcomes
ELA z-score -0.57 - -
Math z-score -0.23 - -
Fall GPA 1.96 2.00 2.30
Spring GPA 1.97 2.01 -
Retention 0.11 0.10 -
Behavioral Outcomes
Days absent 18.28 14.28 4.86
+

Number of suspensions 0.06 0.00 0.00
+

+ Note: The number of absences and suspensions in 2014/15 school year are just for the fall semester of the school year.
82
Appendix C. Descriptive Statistics, Overall Means for all Students During POPS Years
(2012/13 - Fall 2014) and Before POPS Years (pre-2012/13), Student-Year Observations
Variable
Post-POPS
Students Pre-POPS Students
Number of Student-Year Observations 7,116 5,933
Student Characteristics

Male 0.53 0.53
Female 0.47 0.47
Black 0.12* 0.10
Hispanic 0.67 0.68
Asian 0.07 0.08
White 0.13 0.13
Other Race 0.01 0.01
Limited English Proficient 0.10** 0.14
Special Education 0.13 0.12
% Eligible for Free/Reduced Lunch 54.9%** 59.5%
8
th
Grade Achievement Characteristics
8
th
Grade ELA z-score 0.03* 0.07
8
th
Grade Math z-score 0.04 0.07
8
th
Grade Fall GPA 2.68 2.70
8
th
Grade Spring GPA 2.67 2.70
High School Academic Outcomes
ELA z-score 0.02 0.05
Math z-score 0.02* 0.09
Fall GPA 2.35** 2.41
Spring GPA 2.32** 2.38
Retention 0.13** 0.19
High School Behavioral Outcomes
Days Absent 7.90** 9.28
Suspensions 0.01** 0.04
** p<0.01, * p<0.05

83
Chapter 4

We All Want the Same Thing:
An Organizational Analysis of a Housing Reentry Program

B. Danielle Williams
University of Southern California

Raphael Bostic
Federal Reserve Bank of Atlanta

Please direct correspondence to:
Danielle Williams
Price School of Public Policy
University of Southern California
Ralph & Goldy Lewis Hall
Los Angeles, CA 90089
bdwillia@usc.edu

Abstract

This study of a partnership between the Housing Authority of the City of Los Angeles (HACLA)
and three local nonprofits in their implementation of a housing reentry pilot program uses a
collaborative governance framework to examine the program’s low participation rate. Prior to
implementation, the program, with its goal of reuniting recently released formerly incarcerated
individuals with family members who receive Housing Choice Voucher assistance and would thus
typically be prohibited from housing them, was expected to be successful by all of the
organizations involved. However, nearly three years after it began, only seven families had
enrolled. Using techniques based on the grounded theory method, this paper first explores non-
interorganizational reasons behind the mismatch in the perceived demand for the reentry pilot
program and the actual take-up rate. It then examines the partnership between the organizations
in the context of the collaborative governance framework articulated by Emerson, Nabatchi, and
Balogh (2012) to ascertain if the collaborative dynamics between the four organizations played
any role in the program’s limited participation. We find that while the dynamics of the
collaboration were such that a successful collaboration would have been difficult without an
immense overhaul of existing norms and protocols, it may be that several external factors,
particularly tenants’ distrust of the housing authority, may have made success difficult regardless.

84
1. Introduction
Among the many challenges faced by formerly incarcerated individuals upon reentry into
the community, securing housing is a particularly difficult one (Geller & Curtis, 2011; Keene,
Rosenberg, et al., 2018). The inability to do so can imperil the post-release progress of recently
released individuals as housing is thought to anchor other aspects of the reentry process,
including preventing recidivism (Clark, 2016; Hiller et al., 1999; Lattimore et al., 2010; Lutze, et
al., 2014; Matheson et al., 2011; Metraux & Culhane, 2004; Western et al., 2015; Zhang, et al.,
2006). Those who do find accommodation typically rely on family to provide housing during the
first few months after their release (Clark, 2015; La Vigne, et al., 2004; Solomon, et al., 2006;
Western et al., 2015). However, for low-income families who receive public housing assistance
this is usually not an option as families receiving rental assistance are not permitted to allow a
formerly incarcerated person to live in the same unit and risk losing the assistance if the non-
permitted family member is discovered living with them (Curtis et al. 2013).
Recognizing that this restriction can be a substantial hardship for these families as well as
detrimental to the community, the Housing Authority of the City of Los Angeles (HACLA)
addressed the issue through the creation of a pilot reentry program in April of 2013. It was one
of the first and largest housing authorities in the nation to do so. The intention was to permit up
to 200 formerly incarcerated individuals who were less than a year into their post-release lives to
reunite with family members receiving rental assistance via the Housing Choice Voucher
program (Section 8).
23
In implementing the program, the housing authority partnered with three
local nonprofit organizations. These nonprofits were tasked with recommending participants to
23
Throughout this paper the Housing Choice Voucher program will be referred to as “Section 8.” While the
program was officially renamed, it is almost always referred to as Section 8, even by HACLA employees themselves
and is still used in their job titles.

85
the housing authority and providing supportive service to smooth participants’ reintegration both
into the family home and society at large. All four organizations believed that the pilot program
would fill a very urgent need for a particularly at-risk population and that many families were
already housing formerly incarcerated family members despite the risk of losing their housing
assistance. It was assumed that these families would choose to opt-in in order to “come out of
the shadows” as one housing authority employee put it. Thus, the organizations expected that
the number of families who would want to participate would be greater than the number of spots
available. So much so, in fact, that HACLA chose not to advertise the existence of the program
at first while one of the nonprofits created substantial new content for their website detailing the
particulars of the program and providing an initial “application” that could be submitted through
their online system. However, the pilot program was significantly undersubscribed. Only seven
families were reunited between May of 2014, when the nonprofits began referring families, and
February 2017.
There are numerous potential explanations for why the pilot program wasn’t as successful
as anticipated, including non-housing challenges that the formerly incarcerated face upon leaving
custody, family dynamics, the relationship between the families and the organizations involved,
the application process itself, as well as other barriers. This paper uses a “critical” and
“revelatory” case study design to generally explore the issues related to the low participation
from the perspective of these organizations’ staff (Yin, 2009). Using these issues, which are
derived from interviews with staff, as the foundation of the analysis, the relationships between
the organizations are then examined in the context of Emerson et al.’s (2012) collaborative
governance framework to determine if any elements of these relationships precluded a successful
implementation of the pilot program.

86
2. Literature Review
Though there is evidence that stable housing is a crucial part of preventing recidivism and
promoting a successful reintegration into society (Clark, 2016; Hiller et al., 1999; Lattimore et
al., 2010; Lutze et al., 2014; Matheson et al., 2011; Metraux & Culhane, 2004; Western et al.,
2015; Zhang et al., 2006), the challenges of finding housing for the formerly incarcerated remain
formidable (Geller & Curtis, 2011; Keene, Rosenberg, et al., 2018). Indeed, many formerly
incarcerated individuals face not only the barriers that low-income individuals generally face in
finding housing including unaffordable rents (Aurand et al., 2017), the lack of financial resources
needed for deposits, fees, etc. (Bergman et al., 2020), racial discrimination (Dawkins, 2003;
Korver-Glenn, 2018; Ondrich et al., 2000; Rosen, 2014; Yinger, 1986), and the limited supply of
subsidized housing and resulting lengthy waitlists (Aurand et al., 2016; Fischer & Sard, 2017;
Sally et al., 2018), etc. but are also plagued by challenges unique to their situation.
Among these unique challenges, finding landlords who are willing to rent to those with
criminal convictions is a considerable hurdle to obtaining housing (Evans, et al., 2019; Evans &
Porter, 2015; Keene, Smoyer, & Blankenship, 2018). In a study that used matched pairs of
testers, one of whom claimed to have a criminal conviction during an initial contact call while
the other made no mention of a prior conviction, landlords refused to show the property in
question to more than half of the individuals who disclosed a criminal conviction with just 43%
being invited to visit the property whereas they offered to show the property to almost all
individuals (96%) who did not disclose a conviction, though there was substantial variation in
willingness based on the type of conviction reported (Evans & Porter, 2015).
Landlords’ reluctance to rent to individuals with a criminal conviction gives formerly
incarcerated individuals few options for securing housing autonomously, and many end up

87
relying on family for housing (La Vigne et al., 2004; Nelson et al., 1999; Roman & Travis, 2004;
Solomon et al., 2006; Western et al., 2015). Those who cannot must seek transitional housing,
which is in short supply and by definition a temporary solution, or risk homelessness (Roman &
Travis, 2004). However, either option can result in precarious housing situations (Western et al.
2015), and in fact, Geller and Curtis (2011) using data from the Fragile Families and Child
Wellbeing dataset found that the median number of moves per year for formerly incarcerated
men during their first five post-release years was 2.6 whereas with the population at large more
than one move per year is considered a high degree of housing instability. These men were also
more than twice as likely to experience homelessness than their counterparts and were more
likely to be evicted. Interestingly, the higher rate of eviction was specifically mediated by
whether or not the men themselves or their partners had lived in public housing prior to their
incarceration (Geller & Curtis, 2011). This finding touches on two other related challenges
formerly incarcerated individuals face when seeking housing: the prohibitions that prevent them
from accessing subsidized housing and a lack of accurate information regarding their eligibility
for such housing.
Barring the few housing authority programs scattered across the country that are intended
specifically to assist the formerly incarcerated in obtaining their own residences, typically the
quickest way for these individuals to reside in subsidized housing would be through moving in
with a family member who already receives housing assistance. However, despite evidence that
familial support, particularly material support in the form of housing, transportation, financial
resources, etc. is an important element in preventing recidivism and promoting reintegration
(Clark, 2016; Herbert et al., 2015; La Vigne et al., 2004; Mowen et al., 2019; Naser & La Vigne,
2006; Western et al., 2015) a variety of prohibitions often prevent this sort of reunification.

88
Formerly incarcerated individuals are also often unclear on their eligibility for receiving
assistance and frequently assume that all formerly incarcerated individuals are barred from
receiving assistance (Bradley et al., 2001; Keene, Rosenberg, et al., 2018). Complicating the
issue of awareness of eligibility is the fact that regulations can and do vary greatly across
housing authorities, even those in close physical proximity to one another
24
(Curtis et al., 2013).
Furthermore, aside from a few specific federal-level bans,
25
there can be a great deal of
discretion in determining who is eligible for housing assistance (Curtis et al., 2013; Tran-Leung,
2015). Case workers may favor individuals who they perceive as being more likely to succeed
which adds an additional element of uncertainty to the process (Dickson-Gomez et al., 2007).
This lack of clarity results in formerly incarcerated individuals not taking full advantage of those
housing resources that are available to them.

3. Pilot Program Background and Description
Pilot Program Background
Acknowledging both the challenges that formerly incarcerated individuals face when they
seek housing as well as the potential benefits that stable housing offers to these individuals as
well as society broadly, then Secretary of the Department of Housing and Urban Development,
Shaun Donovan, sent a letter in June 2011 to all public housing authority executive directors.
The letter emphasized that there were only two categories of offenses that explicitly barred those
convicted of them from public housing for life: producing methamphetamine in federally assisted
housing and sex offenses that required lifetime sex offender registration by the state (Donovan &

24
A good example of this is the difference in the look-back period between the Los Angeles city housing
authority and the county housing authority which differed for many years.
25
For example, individuals who must register as sex offenders and those convicted of methamphetamine
production are permanently barred from receiving housing assistance.

89
Galante, 2011). The letter encouraged housing authorities to use the discretion available to them
to provide housing for the formerly incarcerated, specifically with the goal of reuniting families.
HACLA’s Director of Section 8 proposed a program that would permit formerly
incarcerated individuals who had been released in the past year to reunite with family members
who received Section 8 assistance. Prior to the implementation of the pilot program, individuals
who had been convicted of any offense were generally prohibited from residing in homes
receiving housing assistance for at least three years. HACLA’s board approved the program in
April 2013. Due to safety concerns expressed by those responsible for the public housing
division, individuals who wanted to reside with family living in public housing were not eligible
to participate.

Other Housing Authorities’ Reentry Programs
In the wake of the HUD letter, other housing authorities also implemented similar
programs to accommodate formerly incarcerated individuals seeking housing assistance. These
programs varied substantially in their structure, diverging on a variety of factors including
eligibility, program requirements, and even working definitions of what constituted a family.
26

Some programs also distinguished between public housing and Section 8 vouchers, allowing one
group to participate but not the other. For instance, New York City Housing Authority’s
program focused on reuniting formerly incarcerated individuals with family members who lived
in public housing but excluded those families who received Section 8 assistance while the

26
Some programs narrowly defined family, including only parents, adult children, and spouses as the leaseholders
who were eligible to take in a formerly incarcerated family member, and at least one program only permitted the
reunification of parents with their minor child/children. Other programs included grandparents, siblings, a partner
with whom the formerly incarcerated individual shared a child, and even step-relationship ties. (Bae et al., 2017;
Hamlin, 2020)

90
Chicago Housing Authority’s program offered reunification with families residing in public
housing or receiving housing assistance and also allowed individuals who had successfully
completed a reentry program to apply for the housing waitlist independently (Bae et al., 2016;
Hamlin, 2020). However, despite the differences in these programs’ structures, they were all
generally plagued by lower than expected participation rates.
While the Chicago Housing Authority eventually filled most of the fifty slots it had set
aside for its pilot program, during the first two years of the program’s existence only one family
participated (Hamlin, 2020). The New York Housing Authority’s program was more successful;
two years into the implementation they had over fifty families participating but that was out of a
total of 150 places available. By the time the evaluation took place three years after the launch
of the program, only 85 families were participating with six of them having “graduated” from the
program after the two-year probationary period (Bae et al., 2016). However, other programs
were unable to gain momentum at all. For example, the San Antonio Housing Authority had
only four participants out of fifty by the second year and by the third had decided to end the
program all together (San Antonio Housing Authority, 2019; San Antonio Housing Authority,
2020).
These programs and others like them differed in terms of the organizations they partnered
with. Some, like HACLA, partnered exclusively with a number of nonprofits, others partnered
solely or in some combination with nonprofits, correctional agencies, landlords, etc. The number
of housing authorities engaged in implementing reentry programs and the diversity of the types
of both the programs and collaborations involved lends evidence to the idea that the
undersubscription to these programs cannot entirely be attributed to a failure of
interorganizational dynamics.

91

HACLA Pilot Program Description
In order to be eligible to participate in the pilot program, formerly incarcerated
individuals:
• Had to have been released from jail or prison within the last twelve months (eventually
this timeframe was extended to 24 months).
• Could not have been convicted of producing methamphetamine in federally assisted
housing.
• Could not have been convicted of an offense that requires lifetime sex offender
registration by the state.
• Could not have been convicted of domestic violence.
• Could not have been evicted from public housing or Section 8 housing for a drug related
activity in the three years prior to applying, though exceptions to this rule were possible if
the individual had completed a supervised drug rehabilitation program or if the
circumstances leading to the conviction no longer existed.
• Had to have a close family member who was a Section 8 leaseholder. Close family was
broadly defined to include: parents, grandparents, adult children, and siblings). Family
members of public housing residents were not permitted to participate.

Of the three nonprofits that collaborated on the program, one was involved in the
program since the very beginning and considered its involvement to have been integral in the
development of the pilot program. Two more nonprofits eventually joined the initiative. Each of
the three nonprofits signed Memorandums of Understanding with HACLA and each were

92
allocated 25 placements for families with the understanding that if a nonprofit filled all of its
slots, then it would potentially be given additional slots depending on whether or not the
placements had been allocated to any additional organizations that might start participating in the
program.
Potentially eligible families were identified either by HACLA or the three nonprofits.
They were most frequently found by the nonprofits among their general clientele. After vetting
them, the nonprofit would then refer them to HACLA. In other cases where the potential
participants first contacted HACLA, rather than one of the nonprofits, they were referred to the
organizations before their eligibility status was confirmed. After the program was initially
implemented, HACLA began using an additional identifying mechanism. If an existing
leaseholder attempted to add an adult family member to the household via HACLA’s standard
protocols and the family member was found during HACLA’s background check to be ineligible
under normal regulations, they were then given the option of participating in the reentry pilot
program and were referred to the organizations. The nonprofit then determined whether or not
the family was a good fit for the program. If the family was eligible and perceived to be a good
fit, the family was referred to HACLA for additional vetting. This vetting consisted of a
background check of the individual and a private interview between a HACLA ombudsperson (a
tenant advocate) and the leaseholder to ensure that the leaseholder did in fact want to reunite
with the formerly incarcerated family member. If the background check didn’t reveal any
disqualifying offense, HACLA sent the leaseholder and formerly incarcerated individual letters
telling them that they were approved to participate. However, this did not automatically permit
the formerly incarcerated family member to live in the home as landlords, who were free to run
additional background checks, had to agree to the addition of the formerly incarcerated

93
individual on the lease. If the landlord did not agree, the family was given the option of moving
to a different residence. Once the formerly incarcerated member was added to the lease, they
were free to stay in the home as long as there were no new offenses. If a new offense was
committed by the formerly incarcerated individual, HACLA had determined that its first course
of action would be to terminate the individual’s permission to live in the household, rather than
revoke the voucher from the lease holding family member, which was HACLA’s standard
response prior to the pilot program.

Description of the Nonprofits
While all three of the participating nonprofits already worked with formerly incarcerated
individuals, the size of the organizations, the resources available to each, and the scope of their
work varied.

Organization A
Organization A was the first nonprofit involved with the reentry pilot program, and the
founder had been very active in advocating for the program to the Director of Section 8.
Typically, the nonprofit’s work is focused on assisting formerly incarcerated women. Its
programs range from support services during the earliest stage of release, which includes group
housing, transportation, and assistance in acquiring government documents, to 12-step programs
and career guidance. They also provide legal assistance for family reunification (typically
regaining custody of children) and record expungement. Through a grant provided by the Los
Angeles County Sherriff’s Office for roughly the first two years of the pilot program’s existence,
they were able to hire a full time staff member whose primary responsibility was coordinating

94
their pilot program efforts, including outreach efforts to identify eligible participants, guiding
them through the application process, as well as addressing participants’ other reentry needs and
providing case management services.

Organization B
Organization B was the second nonprofit to agree to participate in the pilot program.
This organization offers a broad set of programs to South Los Angeles residents with a variety of
needs, including but not limited to, formerly incarcerated individuals. Their services for
formerly incarcerated individuals include pre-release services, counseling, vocational and
educational opportunities, and case management services for probationers. They were also under
an active contract with HACLA to provide services coinciding with the redevelopment of a large
public housing complex. There was no additional funding or staff allocated specifically to
HACLA’s reentry program.

Organization C
Organization C provides services to homeless individuals as well as those at risk of
becoming so. Their slate of community programs focuses on education, job preparedness,
employment assistance, and vocational training as well as health, including physical and mental
health services as well as substance abuse treatment. Organization C also has specific residential
programs that provide supportive services as well as short-term and transitional housing for the
formerly incarcerated, veterans, homeless individuals recovering from serious injuries and
illnesses, and those with mental health issues. There was no additional funding or staff allocated
specifically to HACLA’s reentry program.

95

Program Participation
Given the prior partnerships between HACLA and one of the organizations as well as
their respective resources, the partnership between HACLA and the nonprofits seemed well-
suited to creating “collaborative advantage” with HACLA providing access to much needed
housing for the nonprofits’ clientele and the nonprofits providing services to ensure a successful
transition, services that were far outside of the scope of both HACLA’s mission and practice
(Huxham & Vangen, 2005). However, during the first three years of the program, only seven
families were successfully reunited. Given this seeming “collaborative inertia” as well as the
emphasis placed on the importance of collaboration in the implementation of other housing
authorities’ reentry programs (Bae et al., 2016; Bae et al. 2017), this study pays special attention
to the collaborative governance elements that may have impacted the participation rate while not
excluding factors external to the organizational interactions which may also have contributed to
the low number of participants (Huxham & Vangen, 2005).

4. Collaborative Governance
Collaborative governance and its potential to solve complex issues has gained increasing
traction in the public administration literature in the past two decades (Agranoff & McGuire,
2003; Bingham et al., 2008; Andrews & Entwhistle, 2010; Ansell & Gash, 2008). The broad
collection of partnerships between public, private, and nonprofit organizations that fall under the
term collaborative governance seemingly promise more successful alternatives to traditional
managerial governance structures and top down implementation (Bingham et al., 2008; Ansell &
Gash, 2008) and may be needed to address problems that require a high degree of specialized

96
knowledge or take place within complex institutional structures (Ansell & Gash, 2008). They
may also have a greater capacity for adaptation than traditional governance structures (Emerson
& Gerlak, 2013).
In this context, HACLA’s reentry pilot program seemed well situated to utilize the
advantages offered by collaborative governance. HACLA would provide housing, a much
needed resource for the nonprofits’ clientele, while the nonprofits would utilize their expertise to
ensure the success of the reintegration process both within the home and within the community
at large, something which was outside of HACLA’s purview. By working together, it was
assumed that they would be able to this specific aspect of reentry housing both more efficiently
and successfully than they could individually. As already mentioned, the resulting low
participation rate was unexpected. In order to better understand the mismatch in the demand for
the pilot reentry program with the actual outcomes despite the perceived advantages of
collaborative governance, this study uses Emerson et al.’s (2012) “integrative framework” for
collaborative governance to structure the interorganizational analysis.
Emerson et al.’s (2012) framework is useful for the purposes of our analysis as it takes
into account not only the dynamics between the organizations and the resulting collaborative
actions, which collectively they define as a “collaborative governance regime,
27
” but also
includes the additional dimensions of system context, the drivers of collaborative dynamics, and
collaborative outcomes, a dimension which encompasses both impacts and adaptation.
28

Furthermore, the framework is not conceived of in a linear fashion but rather as an iterative

27
Emerson et al. (2012) define their concept of a collaborative governance regime as “encompass[ing] the particular
mode of, or system for, public decision making in which cross-boundary collaboration represents the prevailing
pattern of behavior and activity” (p. 6).
28
The definitions for each of the dimensions, as well as an explanation as to how they function in relation to each
other, will be elaborated on below in the findings section.

97
process with changes in the dimensions impacting the other dimensions which may in turn
impact the originally altered dimension. This iterative approach is well suited to examining the
partnership between HACLA and the nonprofits given its generally unsettled nature. Similarly,
its specific inclusion of adaptation as a component of the integrative framework is useful in the
context of this study due to the inability of the organizations to successfully adjust their
interactions and processes in order to accomplish the initial goals of the collaboration.

5. Research Questions
Evaluations of the various housing authority reentry programs are, unsurprisingly
considering their relative newness, sparse. Recommendations from one of the few programs that
have been evaluated focused on broadening the eligibility requirements, engaging a diverse set of
stakeholders to collaborate, and educating tenants on housing authority policies (Bae et al.,
2016). An overview of existing housing authority programs yielded additional approaches
including specifically engaging local police departments as well as housing authority staff and
again emphasized collaboration across a wide array of stakeholders (Bae et al., 2017). The
breadth and variety of suggested best practices and recommendations indicate potentially a host
of programmatic pitfalls. In order to better understand what elements may have played a role in
the pilot program’s low enrollment rate, we look at the following questions:
1.) What factors contributed to the pilot program’s low participation rate?
2.) What evidence is there that the difficulties with the pilot program implementation
stemmed from interorganizational dynamics?

98
3.) If there is evidence for the collaborative dynamics, as defined by Emerson et al.’s
framework (2012), being insufficient, why were the organizations unable to adapt and
address these challenges?

6. Data and Methodology
The data consist of semi-structured interviews conducted with 17 staff members from the
housing authority and two of the nonprofits. To our knowledge, the third nonprofit had only one
staff member involved in a meaningful way with the pilot program and that individual did not
respond to interview requests. However, we did have extensive notes from two phone calls with
the individual that we were able to use in lieu of an interview. In addition, the first author had
maintained a running log of program details, meetings, various program developments, and other
pertinent information. This was due to our involvement with the pilot program as the
prospective program evaluator for several years prior to the conception of this study. It was only
upon realizing that participation was going to remain low for the foreseeable future, thus
precluding the intended quantitative evaluation of the impact on the formerly incarcerated
individuals and their families, that this study was undertaken to better understand the reasons for
the lack of enrollment.
While our own immersion in the program enabled us to identify the majority of the
organizational staff who had participated (even minorly) in the execution of the pilot program,
each individual was asked at the end of the interview to share the names of anyone they thought
should be interviewed about the program. Several additional interviewees were identified this
way. Eleven former and current HACLA staff members were interviewed along with a total of
six former and current staff members from two of the nonprofits. The interview guide focused

99
on the program’s creation and implementation, the relationships with the other organizations, and
differences between the organizations themselves. Some but not all of the questions on the guide
were created with the collaborative governance regime model in mind, though these questions
were not structured around the concepts verbatim (Emerson et al., 2012; Emerson & Nabatchi,
2015). While the questions were designed to cover all elements of the collaborative governance
regime model, an emphasis was placed on addressing the elements that make up “capacity for
joint action,” as our observations led us to believe that it in particular, or rather the lack thereof,
might be a substantial barrier to the growth of the pilot program.
To this end, we used grounded theory techniques and concepts to explore potential
explanations for the low participation rate. Please note that we use the terms “technique” and
“concepts” rather than “method” as we used a modified version of the “classical” grounded
theory method. Our approach mostly focused on using the coding schemes and concepts such as
“theoretical sampling” as described by Glaser and Strauss (1967) and Corbin and Strauss (2015)
in our analysis. We avoid referring to it as using proper grounded theory method for two
reasons. Foremost is grounded theory’s emphasis on inductive methods. Glaser and Strauss’s
(1967) original text called for researchers “literally to ignore the literature of theory and fact on
the area under study” (p. 37). While Corbin and Strauss’s (2015) work later acknowledged the
helpfulness of prior knowledge, specifically in the form of literature reviews, in creating
questions for interviews, making comparisons, and forming analytical questions, Glaser
continued to staunchly reject this approach (Glaser, 2013). Given that we wanted to look at the
program’s implementation specifically in terms of its interorganizational relationships, it would
have been both impractical to approach this work from a “naïve” perspective and disingenuous to
claim that we were doing so considering the extent to which we were already involved with the

100
program prior to the beginning of this study. We already suspected that certain aspects of the
program impacted the participation rate. Ignoring these insights would not only be impossible
but would result in rejecting potentially valuable observations. Thus, our approach to examining
the literature was more in line with the Constructivist grounded theory method which rejects the
possibility of a “tabula rasa” (Thornburg & Dunne, 2019). However, as we wanted to adhere to
the spirit if not the law of the grounded theory method, specifically in being open to concepts
that emerge from the data itself, we undertook a two-step literature review (Corbin & Strauss,
2015; Thornberg & Dunne, 2009). Based on our own observations, we suspected that the cross-
sectoral partnership between the organizations had played a role in the program’s low enrollment
and wanted to pay special attention to any relevant collaborative governance elements. Thus, the
initial stage of our literature review was limited to what was necessary to formulate questions
related to collaborative governance. The second stage of the literature review occurred after the
interviews had already been collected and involved an examination of similar programs
undertaken by other housing authorities and their outcomes in order to better contextualize the
elements that emerged from our study.
The other significant way in which this study deviates from grounded theory method is
that it is used, in part, to examine a theoretical framework. Corbin & Strauss (2015) explicitly
discourage the application of theoretical frameworks, arguing that “it would contradict the
purpose of the [grounded theory] method. (p. 52).” Despite this, I chose to use grounded theory
techniques due to their inherent flexibility and out of a desire to remain open to both
organizational-based and other concepts that might emerge from the data independently of the
collaborative governance framework. It especially seemed appropriate as this was intended as a
“revelatory” case study (Yin, 2009). Furthermore, it should be noted that while the interviews

101
included questions intended to better elucidate the nature of the cross-sectoral collaboration, our
initial observations and program log notes were agnostic in this regard. Though the program log
tended to focus on more practical aspects of the implementation and are not as extensive as they
would have been had our initial aim been a case study, they do have the advantage of having
been written without a particular theoretical standpoint in mind. Rather it was our observations
and notes which led us to consider the low enrollment rate as a potential collaborative
governance issue and consequently choose to use the collaborative governance framework to
explore it.
The coding was done using NVivo software. The initial round consisted of listening to
the interviews and doing basic in vivo coding. We also in vivo coded the notes from the phone
calls with the individual who was not interviewed. After the initial coding was done for an
interview, the codes for that particular interview were then compiled into subcategories.
29
These
subcategories were then compared to existing subcategories taken from previously coded surveys
and further refined. This process was then repeated by reading the transcripts in order to identify
any additional relevant subcategories and to continue to compare and hone the subcategories
across interviews. The relevant pieces of the transcript that were attached to subcategories were
then reread, and the subcategories consolidated into categories. These categories were then
examined in the context of the collaborative governance framework. This analysis was then used
to derive a “core” category from the data.
30

Thus, while it wouldn’t be appropriate to describe this study as having used the grounded
theory method, it is most certainly indebted to the method both in terms of the concepts and

29
Subcategories are also frequently referred to as “concepts” (Corbin & Strauss, 2015).
30
For further discussion of the differences between categories, subcategories, and core categories see Corbin &
Strauss, 2015, pp. 216-217, 220.

102
techniques used in the analysis. While Robrecht (1995) critiqued Corbin and Strauss’s emphasis
on procedure, arguing that it resulted in researchers looking “for data” rather than “at data,” the
strategies they describe provided useful guidance in how to approach this study analysis. For
instance, it was their “theoretical sampling” concept which emphasizes following concepts that
emerge from the data rather than just focusing on people, that led us to further pursue questions
regarding HACLA’s “street-level” bureaucratic processes and the tenants’ relationship to the
housing authority and its employees. As a result, HACLA advisors (case managers) who had
dealt firsthand with clients seeking to participate in the pilot program were interviewed (Corbin
& Strauss, 2015; Glaser & Strauss, 1967). The advisors were the ones who had direct
interactions with the families who sought to participate in the program. These individuals were
not involved in the creation or higher-level administration of the pilot program, and as a result,
their interview guide focused more on the ground-level processes of the program, their own
introduction to the program, and their opinions of it and its successfulness. Like the interview
guides for the staff, the guides also included questions related to collaborative governance
concepts.

7. Findings
Categories that emerged from the data as important potential factors in explaining the
reasons behind the low program participation rate are discussed in detail below. They are
grouped according to the dimension of the collaborative governance framework they relate to.
This allows us to examine the factors that were specifically classified as relating to the
interorganizational relationships and to assess the extent to which they were likely to have

103
impacted the participation rate. Brief summaries of the dimensions are provided before each
group in order to give context to the way in which these factors fit within the framework.

System context
In Emerson et al.’s (2012) integrative framework for collaborative governance, system
context consists of a variety of external influences, including the political atmosphere, existing
laws, levels of conflict and trust, resource conditions, etc. that shape the collaborative
governance regime. However, the system context should not be interpreted merely as the
conditions that exist at the beginning of the collaboration but rather as external forces that impact
the collaborative governance regime at any point. Below are the external factors that affected the
collaborative governance regime at various points during the collaboration between HACLA and
the nonprofits and help explain the low program participation rate.

Tenant Distrust of the Housing Authority
Tenants’ distrust of the housing authority was one of the most frequently cited factors
influencing the low participation rate of the pilot program. Nearly every nonprofit and housing
authority staff member interviewed mentioned the distrust and fear that HACLA tenants have of
HACLA itself, and while frequency should not be equated with importance, the consistent
discussion of the issue as well as numerous examples of tenant interactions and the proffering of
various solutions specifically related to it indicates how much it plagued the pilot program. It
became clear that many people believed that the program was a ruse by HACLA intended to
trick individuals into revealing that they had a formerly incarcerated family member living with
them and revoke their housing assistance as a result. Housing authority employees reported that

104
individuals who called to ask about the program were often reluctant to give their names out of
fear that it was actually a trap to catch those who were already violating the policy. As one
HACLA employee recalled of an interaction with a client: “the [tenant] was telling me that a lot
of her friends were saying don’t do it because they're just going to trap you up.” Their distrust of
the housing authority persisted even once they had been assured that the program was real as
many refused to even reveal where they had heard about the program or who they had heard
about it from.
This lack of trust was further exacerbated by communication issues between the tenants
and HACLA. Early on in the implementation of the program, over 6,000 HACLA tenants
31
were
selected at random to receive a letter from HACLA alerting them of the reentry pilot program in
order to better publicize the program. HACLA had originally been reluctant to advertise the
program due to the mistaken belief, shared by the nonprofits, that demand for the program would
quickly outstrip the allocated number of spots. When the participants did not materialize,
HACLA sent the letter in the hopes that officially informing some of the tenants would increase
interest in the program. However, the majority of the responses to the letter were phone calls
expressing concern and fear that the individuals themselves were being accused of either
allowing an unauthorized individual to live with them or of having a criminal record that would
disqualify them from receiving their housing assistance. One HACLA employee recalled:
We were inundated with calls and questions like hundreds of calls … like an old, old
man: ‘I committed that crime in 1964; don't throw me out.’ … And I would say, ‘sir,
read the letter, take a breath.’ He's so panicked, ‘take a breath, read the letter.’ ‘I have

31
Though there are approximately 48,000 Section 8 voucher holders, letters were only sent to a random selection of
6,000 due to the cost associated with a large-scale mailing.

105
read the letter. It's says, incarceration, that was 1960…’ -- so, I had to read it to him. So,
and so he was not the only one, we had many other calls for it.
Those who did express interest in participating followed the pattern mentioned of expressing a
general leeriness of HACLA and similarly did not want to give their names or provide other
contact information.
Efforts were made to address the tenants’ suspicions and to further publicize the
existence of the program with one of the nonprofits periodically setting up a table in the lobby of
the main HACLA office in order to provide information. HACLA and the same nonprofit spoke
about the pilot program as one of their offerings at various reentry fairs with seemingly little
success in terms of increasing the participation rate. One potential solution suggested by both
nonprofit and HACLA staff members was to engage local churches in order to spread the word
about the program and improve its legitimacy in the eyes of the community. However, while
there were piecemeal efforts made by individual staff members to speak to their own churches,
no coordinated effort was ever undertaken. Part of this was due to HACLA’s concern of their
outreach being misconstrued – “often times too when we put out a broader message to the
community, we get so much interest [in] other things too. People say, are you open for Section 8
now, or not? … So we have to be mindful of that and be very specific.”

Missing Stakeholders
In addition to the lack of church involvement and broader community organization
involvement, key stakeholders in the criminal justice arena, namely the Sheriff’s and Probation
Departments, did not participate in the program despite the fact that in the earliest stages of
developing the reentry pilot program, HACLA sought to partner with both agencies. Section 8

106
management at HACLA had viewed the partnership as a natural one given the need that the
program would fill for the Sheriff’s and Probation Departments’ clients. However, concerns
arose among higher management at the two agencies regarding the political perception of the
formerly incarcerated receiving special treatment or vouchers being “given” to the population
while there were operational issues among the rank-and-file staff at the agencies. After nearly
two years of attempting to negotiate parallel memorandums of understanding with both agencies,
Section 8 management at HACLA decided to abandon these efforts and pivot towards engaging
community organizations.
Within HACLA itself, there was a missed opportunity to engage the advisors and other
ground-level workers in implementing the program. Each year every Section 8 leaseholder along
with all of the adult members of the household must visit a HACLA office and meet with an
advisor, which means that if advisors had been directed to tell every tenant they met with about
the program then within a year every single HACLA tenant would have been aware of the
existence of the program. While it’s doubtful that this would have immediately alleviated
tenants’ fears regarding the program being a trap, it would have laid the groundwork towards a
broader awareness of the existence of the program as well as a better understanding of what the
program itself was.
One of the issues that contributed to the disconnect between the advisors and reentry pilot
program was that the program was very much seen as a top-down initiative with Section Eight
advisors learning of the program via memo. Some staff were uncertain as to how the program
was even permissible under HUD regulations, an indication that the impetus for the program
wasn’t conveyed to advisors. While there was no monolithic response from these staff members
- some supported the program while others expressed concern – one of the goals of the program,

107
the elimination of fraud, was contradictory to values expressed by staff in terms of the way in
which it was executed, namely simply now permitting what formerly had not been allowed.
Asking individuals whose primary duties include rooting out fraud to now facilitate behavior that
had heretofore been prohibited represented a significant shift in mindset for staff, one that was
not without tension. In fact, there was at least one report of a tenant being discouraged from
participating in the program by staff.
The disconnect was further exacerbated by the fact that in an attempt to streamline the
process, all potential cases were given directly to the ombudspeople. This was done in order to
better facilitate the application process for families, since the ombudspeople were familiar with
the program and interacted directly with the nonprofits. However, as a result, the approval
process was completely opaque to the advisors with the advisors only participating in the process
once the family member had been approved to join the household. While the protocol for adding
the formerly incarcerated family member to the lease was already familiar to the advisors as it
was identical to any other instance of adding an adult family member to the household, it was
also lengthy. The amount of time it took to process the application was relevant because it was
subsequently viewed as a waste of time if the formerly incarcerated family member reoffended
or never rejoined the household for some other reason. One advisor expressing his skepticism of
the program as a function of perceived time wasted spoke of one specific case where
[T]hey got out of jail, they went back, they got out, they went back there. So if that’s
going to be the case, and we'd bring them in, and then they just go right back to jail. We
feel like we just wasted all the time that we -- it's a long process, the paperwork to try get
them to fill it out, to try to go to the checks and checking the income, checking the
background, all of that takes so much time, almost an entire day… I mean each of us

108
have about 500 to 600 cases already to begin with. So something like that in addition to
the normal annual review is a really time-consuming thing for us. But if it's going to work
fine, good. But if it's going to be like that client that I had, then I feel like we just wasted
time. So that’s just my honest feeling…
Even an advisor who was initially very enthusiastic about the program became disillusioned
upon working with a family due to the problems that arose with the dynamics at home that
eventually led to the formerly incarcerated family member leaving the household. She recalled:
Like I said, I was so excited, and I was like, oh my God, yay, you know, I'm doing this.
I'm actually walking through it, and when, you know, all the paperwork was done, and I
told her, hey, you got approved, whatever. And you get that phone call, and you hear the
pain, there’s like a sense of guilt, and like, okay, I know [the program’s] good, but I also
have to help her get out of it.”
Based on their experience, the advisor felt that there needed to be more emphasis placed on
providing therapy and supportive services, seemingly unaware that these services were provided
in some measure by the nonprofits. Unsurprisingly, when asked about the nonprofits, the advisor
had little knowledge of their role in the pilot program.

Shifting Family Dynamics
As the HACLA advisor encountered, the relationship between the leaseholder and the
family member could undergo profound changes within a short period of time - changes that
ended their participation all together. Ombudspeople at HACLA reported similar experiences
with the participants they assisted, including at least one instance of the formerly incarcerated
family member being withdrawn from the program due to a domestic violence incident. It was

109
these changing dynamics that resulted in more withdrawn applications and terminated
reunifications than recidivism.
32
One especially telling example came to our attention when a
colleague reached out to a woman in order to ask her to take part in the second round of
interviews. When we had last spoke to her, she was seeking permission to allow her recently
released mother to live with her and seemed excited by the prospect of her mother joining her in
her home. However, when contacted six months later, not only was she no longer participating
in the reentry pilot program, she informed us that she was not speaking to her mother and had
nothing to do with her anymore. There were several other instances of families whose
relationships deteriorated to the point that they dropped out of the program. These developments
are somewhat unsurprising given the evidence that even when families want to assist their
formerly incarcerated family member, they incur a profound amount of stress in doing so (Grieb
et al., 2014). Furthermore, there is some evidence for the existence of a mismatch between the
pre-release expectations between those being released and their family members found in the
literature. One study reported that women who were awaiting the return of a male significant
other were over twice as likely overall to predict that their relationship would be “pretty” or
“very hard” (Bir, et al., 2015). Similarly, within the couples themselves, men consistently
reported more positive outlooks on the likelihood of having a good relationship post-release than
their partners (Bir, et al., 2015). The disparity in these outlooks may itself increase stress in the
relationships.
Another factor that may play a role in the deterioration of these relationships is that the
leaseholding family member may actually not truly want to invite the formerly incarcerated
family member into the home in the first place. Staff members at two of the nonprofits

32
To our knowledge during the time period of the study, only one family was disqualified from participating in the
program due to the formerly incarcerated family member recidivating.

110
acknowledged that family members may experience conflicting emotions: “Often times, the
family members don’t want to say no, and they hope that someone would say no for them. … So
the mama wants him. He wants to go there, but it's not a good fit because deep down in her heart,
she don’t want him there.” This was echoed by some of the HACLA staff who also reported
instances of interactions, some subtle, some overt, with leaseholders that made them doubt
whether or not the family member truly wanted to reunify. One recalled, in the context of the
leaseholder deciding between adding the formerly incarcerated individual as an adult family
member only or as a co-tenant who would have the same rights as the leaseholder: “…when we
advise them [of] the difference [between] the two and what can lead to what and whatever,
they're right away, ‘no, I'm protecting my household and my children and I'm not going to allow
this individual to take that away from us.’” One of the nonprofit staff members also echoed the
idea that the leaseholder’s fear that the formerly incarcerated family member could potentially
threaten their housing may have discouraged some eligible families from participating:
Let's give it about a three to six-month trial run, and I think that would make family
members feel a little bit more comfortable as well because they felt uncomfortable with
that. They were, okay, it's one thing to let them live here, but to add them to my lease,
now, I feel like my housing is going to be jeopardized if they do something.

Drivers
Emerson et al. (2012) specifically separate the concept of “drivers” of the collaborative
governance regime from the system context because these drivers are the instigating factors that
exist at the outset and initiate the collaboration; they do not play an iterative role like the other
dimensions of the framework. Drivers consist of four elements: leadership, consequential

111
incentives, interdependence, and uncertainty, at least one of which has to be present in order for
a collaborative governance regime to take root. All four of these elements were present at the
beginning of the reentry pilot program collaboration. Consequential incentives existed due to the
opportunity presented by the letter from the Secretary of HUD to address the issue at hand. In
terms of interdependence, while the nonprofits certainly would have been unable to undertake
such a project without the cooperation of the housing authority, HACLA arguably could have
implemented a version of the program, although one without supportive services, that did not
rely on the nonprofits. However, this was not the type of program that they had envisioned, and
the housing authority’s desire to have these services provided led them to seek out partners for
the endeavor. This desire to partner with other organizations was further driven by uncertainty
as to how best to proceed in terms of recruiting participants. This uncertainty existed despite the
fact that both HACLA and the nonprofits were initially confident of the outcomes of their efforts.
Their collaborative efforts also had a leader in the form of Section 8 leadership, which was able
to overcome the intraorganizational barriers to establishing the pilot program and who worked to
cultivate relationships with other agencies and organization.

The Collaborative Governance Regime
“Collaborative governance regime” is the term coined by Emerson et al. (2012) to
describe systems of public decision making that are dictated primarily by the behavior and
actions of cross-boundary partnerships. The collaborative dynamics and collaborative actions
that result from the circumstances set in motion by the aforementioned drivers are what makes
up a collaborative governance regime. It’s the interaction of these two dimensions that

112
determine the impacts of the regime and thus the potential opportunity to change the system
context.

Collaborative Dynamics
Collaborative dynamics consist of three interactive components which operate in an
iterative rather than linear fashion: principled engagement, shared motivation, and the capacity
for joint action, each of which are comprised of four elements.

Principled engagement
Principled engagement refers to the shared purpose and agreements that emerge over time
in order to accomplish the goals of the cross-boundary partnership. It consists of the elements of
discovery, definition, deliberation, and determination. Discovery, which is related to identifying
shared interests and pertinent information, and deliberation, which concerns the analysis and
communication of difficult issues among the organizations, were both present but insufficiently
so in certain aspects (Emerson et al., 2012). Without discovery, the organizations would not
have partnered with each other to begin with. However, there was also a failure on the part of
the organizations to identify a structural issue with the program. A limited form of deliberation
also manifested itself early on when issues emerged with the initial attempts at reunification.
These problems resulted in candid conversations between HACLA and the involved nonprofit
about what went wrong and how best to move forward. However, the solution, which was
decided by HACLA underscores the lack of joint determination the partnership entailed.
Ultimately it was HACLA who wielded authority over how precisely the program would
proceed. Furthermore, there was an overall lack of communication, which is discussed in more

113
detail below as it emerged as a factor that impacted the participation rate, about the program that
limited both deliberation and determination. There were also issues in defining specific
terminology which led to friction between HACLA and the one of the nonprofits in particular.
For instance, the nonprofit staff used the phrase “formerly incarcerated” to indicate those
individuals who had been released from custody whereas many members of HACLA staff
referred to these same individuals as “ex-offenders,” a term which the nonprofit staff found
problematic. While some HACLA staff did adopt “formerly incarcerated,” others continued to
use “ex-offender” despite the nonprofit explicitly mentioning it as an issue during at least two
meetings.

Lack of communication
The distinct lack of communication between HACLA and the nonprofits as well as
between the nonprofits themselves severely hampered the efforts of the pilot program. Face-to-
face meetings between all of the organizations were sporadic, even at the very beginning of the
partnership. As the collaboration progressed there were meetings where only one organization
showed up or that were cancelled at the last minute due to organizations pulling out. There were
also discussions of the nonprofits sharing best practices with each other to determine what was
working and what wasn’t in terms of recruitment, but despite the persistent low levels of
participation, there were never any meetings or phone calls actually scheduled. Later on, once
HACLA had appointed a staff member to act as a liaison, HACLA sent to the nonprofits on a
monthly basis requesting referrals, but they were often met with silence from some of the
nonprofits. As one staff member recalled: “[T]hey just don’t respond. There’s just no response
[to] the emails.” For their part, the nonprofits also felt like the communication with HACLA was

114
less than ideal. When asked how they would characterize the communication and relationship
with HACLA, one nonprofit staff member replied:
I am still waiting for this meeting that's supposed to happen. … I know every now and
then I'd get [an] email from HACLA asking how things are going, what are we doing and
there is an extension of help. And I just…want to say yes, send us some more folks.
Communication broke down to such an extent that not only were some of the nonprofits unaware
of how many families were actively participating in the program as well as other details but one
staff member even viewed it as information that HACLA was potentially unwilling to share with
them. They put it succinctly: “I am not clear on how many people have actually taken advantage
of the program itself. It's still a well-guarded secret in my opinion.” It also led to missed
opportunities due to a lack of information - information that interviewees sometimes seemed to
learn about for the first time during the interviews themselves:
Interviewer: Why does he no longer qualify [for the pilot reentry program]? Because he
is over the time limit?

Interviewee: Yeah, he went over the time limit.

Interviewer: Oh that’s so interesting. So how long was his parole then because now they
have upped it for two years?

Interviewee: Okay, so they are on supervision for 13 months.

Interviewer: For 13 months, okay. So you know now they have upped the program;
they've upped it for two years. Did [HACLA] tell you that?

Interviewee: No.

Interviewer: So now you can be up to 24 months released and go on the housing
authority’s program.

Temporary vs permanent reunification

115
A structural issue that may have negatively impacted the participation rate and one that
the organizations failed to identify was that the program made no distinction between permanent
and temporary participation. While it was seemingly conceived of by both the nonprofits and
HACLA as a long-term program, this ignored the reality that some of these reunifications were
likely never intended to be permanent by the family members. A long-term reunification model
made sense in the context of reunifying spouses as well as partners who share children, but it
didn’t necessarily fit the needs of other familial relationships. There were several instances of
adult children either moving in with their parent and then leaving the household because their
ultimate goal was to find their own housing or dropping out of the program during the
application process because they were able to secure other accommodations while they were
waiting to be approved. The various layers of approval required may have also dissuaded certain
families from participating as the burden may have appeared to be too onerous to leaseholders if
they only intended to allow the formerly incarcerated family member to stay with them for a few
months or even just a few weeks. It also ignored the immediacy of the need for housing among
formerly incarcerated individuals upon exiting custody. This may account for why, to our
knowledge, no sibling pairs, who presumably would not intend to reside in the same home
permanently, attempted to participate in the program during the time period covered. In fact, the
only reunified family that we were able to interview all three times over the course of a year as
planned was a married couple with two young children in the home.

Shared Motivation
The second component of collaborative dynamics is “shared motivation,” which
emphasizes the interpersonal relationships that are necessary to sustain collaborative governance

116
arrangements. It consists of mutual trust, mutual understanding, internal legitimacy, and shared
commitment (Emerson et al., 2012). While there was no explicit distrust among the nonprofits
and HACLA, there simply was not enough interactions to establish the sort of trust that would
sustain mutual understanding, with mutual understanding in turn underpinning both legitimacy
and shared commitment. Though not obvious on its face, the differing client populations that
were prioritized by the respective organizations also undermined a shared commitment to the
program. While all involved wanted the program to succeed in housing formerly incarcerated
individuals with family members, their existing modes of operation and the associated
perspectives prevented them from establishing agreed upon priorities and procedures.

Different client populations
Ostensibly the nonprofits and HACLA’s shared goal of providing family-based housing
to recently released formerly incarcerated individuals should have helped facilitate the shared
commitment that would allow them to accomplish said goals. However, the nonprofits and
HACLA were used to catering to two separate, if sometimes overlapping, populations: formerly
incarcerated individuals and low-income tenants, respectively. Their emphasis on the needs of
their particular clientele resulted in differing priorities across the organizations which
undermined the formation of shared commitment as well as mutual trust. In fact, the very first
attempt at reunification was terminated because of problems that arose due to this issue. The
nonprofit recommended a family to HACLA for participation in the pilot reentry program.
Everything seemed to be progressing well until a phone call between the ombudsperson assigned
to the case and the leaseholder during which it was revealed that the individual did not in fact
want their spouse to rejoin them in the home:

117
And I said … this is Section 8. ‘Yeah, I know y’all trying to make that man move in with
me’ … and I said no, no, no, no. I'm calling to find out if you want to participate in this
and have your kid's dad move back in with you and she was relieved and said ‘oh, you’re
on my side’… and she is so happy that she doesn’t have to do this.
Due to this and other comments made by the leaseholder which indicated that she had felt
pressured by the nonprofit to participate, HACLA asked the individual to meet with them and the
nonprofit so that the nonprofit could gain insight into why she felt pressured. As a result, both
the nonprofit and HACLA instituted their own solutions to prevent the issue from occurring
again, but these decisions were portrayed in the interviews as being made independently of each
other. HACLA revised its protocols to include a one-on-one interview between the leaseholder
and an ombudsperson to ensure that the individual was actually a willing participant in the
program and specifically cast this change as a “safeguard” that they undertook themselves in the
context of acknowledging that the nonprofits were approaching the program from a different
client perspective:
Because then your perspective … for the point of entry is from … the perspective of the
potential household member not the family …. and we said okay, we will figure out how
to institute safeguards ourselves which include independent interviews with the family …
building in a little bit of a delay in the process …
For their part, the nonprofit addressed the issue by adding a meeting with their in-house mental
health provider as a requirement of their application process.
The delay that was subsequently built into the application process by HACLA in addition
to their already existent protocols, including the background check and the family appointment
with the advisor, was of concern to the nonprofits. Recognizing that formerly incarcerated

118
individuals’ housing challenges are most acute immediately upon release, they emphasized the
speed at which the formerly incarcerated family member could potentially enter the home even
after the problems that arose with the first referral family. One nonprofit suggested essentially
“pre-approving” families to participate in the program so that there would be fewer additional
steps for families to complete upon the formerly incarcerated individual’s release and thus would
allow the individual to reside in the home more quickly. The housing authority, on the other
hand, remained most concerned with ensuring that the leaseholders not be adversely affected by
participating in the program, specifically in terms of their safety and the risk of potential eviction
due to the actions of the returning family member. In their opinion, the process on their end did
not take that long and to the extent that it did, the slower speed was considered a positive as it
gave the leaseholder more time to reconsider the arrangement and pull-out of the program prior
to the formerly incarcerated family member entering the home if necessary.

Capacity for Joint Action
The final component of the collaborative dynamics concept is “capacity for joint action”
which refers to the ability of the collaborative governance regime to create and sustain joint
actions that did not exist prior to the collaboration between the organizations (Emerson et al.,
2012). We paid particular attention to this component and its elements (procedural and
institutional arrangements, leadership, knowledge, and resources) when we designed our
interview guides due to our suspicion, based on our existing knowledge of the situation, that
these elements in particular may have impacted the collaboration between HACLA and the
nonprofits. Among them, procedural and institutional arrangements and resources stood out
given the lack of clarity surrounding program procedures as well as the lack of allocated
119
resources, with one notable exception, dedicated to the program. Leadership - in this case
defined as more than just a driver of the collaboration – as an important element of the
collaboration itself, seemed less problematic. While knowledge, specifically the “generation of
new, shared knowledge,” seemed a likely source of dysfunction given the communication issues
mentioned above (Emerson et al., 2012, p. 16).
Unclear Processes and Missing Information
Procedural and institutional arrangements must be sufficiently articulated both at the
intraorganizational and interorganizational levels in order to successfully manage the interactions
necessary for a successful collaboration (Emerson et al., 2012). However, in this case, the
communication issues mentioned earlier in the context of principled engagement had
ramifications for other aspects of the partnership, namely a lack of clarity surrounding the
admission process. In fact, one of the early attempts at reunifying a family failed because of
differing standards between HACLA and one of the nonprofits of what constituted a good fit for
the program. The individual the nonprofit referred had a domestic violence conviction but
intended to live with a parent while the offense in question involved a romantic partner. To the
nonprofit the fact that the conviction involved a different type of relationship was sufficient for
them to feel that the individual was eligible for the program. The housing authority did not share
this opinion and not only overruled the nonprofit in this specific case but also decided to
completely bar individuals with domestic violence convictions from participating in the program
outright.
There were also discrepancies between the referral processes across the nonprofits. Of
the two nonprofits that ever referred families, one undertook a significantly more thorough
120
vetting process which included a home visit and a family meeting with a mental health
professional as well as other requirements. The other eventually viewed the process as more
bureaucratic, and in fact characterized the notion that the vetting process had to be very involved
as a “miscommunication” that had initially prevented the nonprofit from referring more
individuals:
[I]t is that type of miscommunication that is out there. It's like you have to do this, and
you have to do that, and you have to do this, and you have to do that, but when I actually
just went through the process with [participant’s name] it was a very simple process.
The other nonprofit only learned of the differences in the referral process via a program
participant. The staff was also not aware of whether the other organizations had successfully
referred anyone to the program. During an interview with one staff member, they even asked the
interviewer for information regarding it: “That’s why I asked have any of the other organizations
submitted any referrals because it was brought to our attention that there isn’t any kind of
interview process. It’s just ‘give me your name and number, and I can submit the referral.’”
It’s unclear to what extent HACLA was aware of the discrepancies between the two
differing referral processes and to what degree it mattered to them. While HACLA expressed an
expectation that the participants would be involved with the nonprofits, at least initially, there
was no discussion between HACLA and the nonprofits as to precisely what this involvement
should entail or what it’s duration should be. Similarly, there was confusion as to who was to be
responsible for recruitment of participants. While many staff members at all the agencies
believed that HACLA and the nonprofits were equally responsible for referring participants, a
not insignificant number, both at HACLA and at the nonprofits, believed it was the primary duty
of the opposite organization to refer individuals to the program.
121
Limited Leadership
The lack of input from HACLA regarding the specifics of the referral process and
program expectations stemmed at least in part from the view of Section 8 leadership that there
were few demands that could be placed on the nonprofits given the completely voluntary and
unremunerated nature of their participation -- “We were trying to keep a kind of a very minimal
set of expectations on folks because you know we are not funding the [program] -- you know
there is no money in it.” This lack of incentives for the nonprofits aside from the potential
housing for their clients prevented Section 8 leadership from providing the degree of leadership
the reentry pilot program seemed to require given the unexpected and substantial setbacks in
terms of the number of participants. Interestingly, HACLA had worked with homelessness
advocacy organizations on another program that was similarly unfunded by HACLA but which
provided Section 8 vouchers directly to participants. It met with far greater success, implying
that the opportunity to live with family is considered a significantly weaker inducement to
participate by at least the intended population if not also the organizations themselves. Indeed,
staff at both of the nonprofits that made referrals expressed a desire for a program that would set
aside vouchers specifically for recently released formerly incarcerated individuals.
Another issue that arose in terms of leadership was simply that the Director of Section 8
who had been the driving force behind the program left the housing authority six months after
the first referrals were received from the nonprofits.
Lack of resources
122
The lack of additional incentives mentioned above underscores a greater dearth of
resources for the program in general. From HACLA’s perspective, the program was conceived
of as one that could simply be merged into the nonprofits’ typical activities. However, given the
issues that arose during the initial stages of implementation, it became clear that the program
would require more resources, particularly in terms of staff time and effort, than initially
assumed. However, out of the housing authority and the three nonprofits, only one of the
nonprofits was able to allocate specific resources, including hiring a dedicated staff member,
towards the implementation of the program. Unsurprisingly it was this nonprofit that was the
most active participant and sent the most referrals to the housing authority. The other two
nonprofits, as well as HACLA itself, had to roll the demands of the program into the duties of
existing staff.
Collaborative Actions
The final component of the collaborative governance regime are collaborative actions.
Not to be confused with impacts, collaborative actions can take a number of forms and may be
performed by all of the organizations involved or by individual organizations undertaking
specific, pre-agreed upon actions (Emerson et al., 2012). In the case of the pilot reentry
program, few of the actions undertaken by the participant organizations were truly collaborative,
particularly across all of the organizations. For instance, the church outreach plan mentioned by
HACLA and the nonprofits had the potential to be a collaborative action even if it was
undertaken only by one of the organizations, but no cohesive, unified determination was ever
made regarding implementing it. One exception was an outreach strategy undertaken by one of
the nonprofits with the cooperation of HACLA which involved staff from that nonprofit
123
periodically setting up and staffing a table in HACLA’s lobby in order to inform tenants about
the program.
Impacts
Impacts in the context of the collaborative governance framework include both intended
and unintended changes to undesirable conditions, either already in existence or those
anticipated. Impacts emerge from the collaborative actions of the collaborative governance
regime but in turn also affect the collaborative dynamics (Emerson et al., 2012). For the pilot
reentry program, impacts included not only the number of families that participated but also
whether those families kept or lost their Section 8 assistance, any potential improvements or
declines in the family’s overall economic wellbeing, and whether the recently released formerly
incarcerated family member was rearrested or reincarcerated post entering the home.
Adaptation
The final component of Emerson et al.’s (2012) integrative framework for collaborative
governance is adaptation, both that which results in external transformative change and potential
new challenges and that which results in changes to the collaborative governance regime due to
alterations in the system context or as a response to the existing impacts or lack thereof. In fact,
there were a number of adjustments made in order to adapt to challenges that arose during the
implementation, some specifically targeted at increasing the number of participants with others
aimed to improve various aspects of the program or interorganizational dynamics. These
adaptations had the potential to address not just the issues at hand but also problems associated

124
with the various collaborative dynamic elements which in turn could have strengthened the
partnership.
As mentioned earlier, in the wake of the failed first reunification attempt both HACLA
and the nonprofit involved altered their protocols, aiming to ensure a safer process for the
leaseholder as well as a smoother process for everyone, including the organizations themselves.
However, while these changes presented the organizations with the opportunity to not only
reinforce their shared commitment to the program but also address issues surrounding other
elements, particularly deliberation and determination, these changes were decided upon and
enacted essentially independently of each other. Similarly, later on when the same nonprofit
pared down some of their internal paperwork and procedures in order to simplify the referral
process in the hopes that this would both make the process less daunting for potential
participants as well as somewhat faster, they did so under their own auspices whereas a
discussion with the other organizations may have offered the nonprofits a chance to better codify
their protocols collectively.
Other adaptations included the efforts made by HACLA to increase awareness of the
program by publishing it on their website and in their tenant newsletters in the hopes that this
would increase participation. HACLA was also able to identify a point in the general application
process where adults seeking to be added to an existing household would normally have been
rejected due to failing a background check. Instead these individuals were notified by HACLA
of the existence of the reentry pilot program, provided contact information for all three of the
nonprofits, and told that they could potentially live with their family if they were subsequently
referred back to HACLA by one of the nonprofits and then successfully completed the rest of the
application process.

125
HACLA also attempted to increase the interorganizational communication as well as
overall engagement by appointing a specific staff member to act as a program liaison to the
nonprofits, though it should be added that this role was in addition to the staff member’s existing
duties and not one exclusive to the reentry pilot program. However, despite HACLA’s efforts to
be more proactive in reaching out to the nonprofits in the hopes that these overtures would in
turn spur more referrals from the nonprofits, the number of participants did not increase in any
significant way.

8. Discussion
Despite these adaptative efforts and their shared goal, the organizations were unable to
improve program participation. As illustrated above, much of this failure can potentially be
attributed to the inability of the organizations to form a productive collaborative governance
regime due to a lack of truly collaborative actions. The absence of joint action capacity seemed
to particularly test the cohesiveness of the partnership. As Emerson et al., (2012) point out
If initial joint action does not occur or the impacts are not close to the identified targets,
[collaborative governance regimes] will be pressured by their partners to make necessary
adjustments in their targets, their theory of action, their collaborative dynamics, or their
investments in building more capacity for joint action. If needed changes are not made,
the low costs of exiting the regime will lead to low participation or departure (p. 19).
This was certainly true in the case of two of the nonprofits. Although there was some indication
in the interviews that one nonprofit was on a trajectory towards more engagement with the
program, they had only submitted a single referral to the housing authority by the conclusion of

126
this study. Another nonprofit was even less engaged and had essentially fallen away from the
program without ever formally withdrawing their participation.
The general success of the collaborative governance regime, particularly the component
of shared motivation, was further undermined by the fact that the primary missions of the
housing authority and the nonprofits involved two different populations. This meant that not
only did the organizations approach the issue from different perspectives but that it ultimately
resulted in goals that were not necessarily prioritized to the same degree or even shared at all.
That this hindered the collaboration is unsurprising given the limited evidence that one of the
most important factors in keeping collaborations intact is the organizations sharing a
commitment to the same population (Thomson, 1999 as cited in Thomson et al., 2009)
All of this was exacerbated by the continued informal nature of the collaboration.
Despite the existence of memorandums of understanding between HACLA and each of the
nonprofits there were no functional measures in place to hold the organizations accountable to
one another. Furthermore, there were no agreements in place amongst the nonprofits themselves
or a collective one with HACLA that established institutional arrangements and procedural
norms. This contributed to the difficulties in establishing a healthy collaborative governance
regime as these collaborations are likely most successful when there is an accountability system
in place that documents inputs, processes, and outcomes and how those change over time
(Bryson et al., 2006; Page, 2004).
While the collaborative dynamics dimension of the collaborative governance regime
would have benefitted from a clearer articulation of expectations surrounding the partnership’s
goals and protocols as well as a general effort to develop closer ties between the organizations,
it’s unclear that this would have substantially increased the participation rate given HACLA

127
tenants’ general wariness of the program. Fear and distrust of housing authorities were
encountered by other housing authorities that implemented similar programs to accommodate
formerly incarcerated individuals seeking housing assistance (Bae et al., 2016; Bae et al., 2017).
This reluctance to engage may be rooted, at least in part, in their interactions with the
legal system via the mechanism of “system avoidance” (Brayne, 2014) and the related concept of
“opting out” (Lageson, 2016). “System avoidance” refers to the idea that individuals who have
been involved with the legal system avoid governmental entities and other institutions that keep
formal records, even those that are not explicitly related to the legal system such as hospitals,
banks, and schools (Brayne, 2014), whereas “opting out” refers to the impact that online
information related to criminal justice records have on familial relationships (Lageson, 2016).
As a result, these individuals “opt out” of situations involving familial duties, such as those that
relate to child wellbeing, particularly schooling, out of fear that their history may be discovered
(Haskins & Jacobsen, 2017; Lageson, 2016). System avoidance can even lead to reduced
political participation not only for those who have had contact with the criminal justice system
(Weaver & Lerman, 2010) but among their partners as well (Sugie, 2015).
Such concern about the potentially negative consequences of interacting with
bureaucratic institutions is not unfounded. Between 1997 and 2006, 10,980 people were arrested
in “Operation Talon” which lured individuals with arrest warrants into visiting food stamp
offices by contacting them and telling them either that supposedly there was something wrong
with their benefits or that they were entitled to a “bonus” and needed to visit the offices as a
result. When they arrived they were arrested (Gustafason, 2011).
In this light, it seems likely that the low participation rate associated with HACLA’s
reentry pilot program as well as others may be yet another collateral consequence of

128
incarceration and other types of involvement with the criminal justice system. In terms of
HACLA’s program, this issue may have been compounded by the additional requirements, such
as home visits, placed on prospective participants by the nonprofit – the nonprofit that also
happened to be the one that was most engaged - during the referral process. These requirements
were intended to ensure a healthy and successful transition into the family home, but as an
additional level of surveillance they may have perceived as more of a potential threat to the
family’s wellbeing (Hamlin, 2020). A staff member of one of the nonprofits echoed this
sentiment: “Because it's hard, when you have people like [those released due to] AB 109 for
instance, a lot of them are here, not because they can't go home, but because they don’t want the
probation officers and the LAPD to invade their homes.”

9. Recommendations
The significance of this reluctance to engage with surveilling institutions and agencies,
including housing authorities, is what leads us to our first and most salient recommendation.
While we believe that it is this factor, which is external to the collaborative governance regime,
that is the most culpable for the low participation rate, there were organizational dynamics at
play that also undermined the effectiveness of the program. Thus, we include specific
recommendations to address interorganizational issues as well.
Recommendation 1: Any housing authority wishing to implement a similar program
intended to assist formerly incarcerated individuals should first, in conjunction with all
partner organizations and agencies, develop a plan to specifically address the issue of
these individuals and their families’ general mistrust of housing authorities.

129
Given that housing authorities found it extremely difficult to be successful in terms of
participation numbers, the housing authorities and the organizations they collaborate with must
develop a thorough plan clearly articulating the specific actions to address public mistrust that
will be undertaken both as a collaborative body and as individual organizations as well as a
timeline for when these actions should be accomplished. Examples of such actions include but
are not limited to: making information more readily available through flyers and other means of
advertisement with content that has been created with the input of all organization involved in
order to ensure a unified message; increasing the housing authority’s transparency through a
more open acknowledgement of the program on their website, in tenant newsletters, etc.;
increasing coordinated outreach to other service providers who may not be part of the program
but who interact with the target population on a regular basis, as well as neutral third parties, like
churches and other religious or community organizations, that are not affiliated with the program
nor specifically work with the formerly incarcerated but are trusted sources of information within
the community and are well situated to advertise the program and provide accurate information
regarding it. Committing to this course of action should also have the effect of improving, per
the collaborative governance framework, the collaborative dynamics processes of deliberation
and determination, which in turn should improve the overall level of engagement with the
program on the part of the organizations.
Recommendation 2: Directly confront the issue that the housing authority and many of
the involved organizations typically cater to different client populations. Specific
concerns about how the needs of the two client populations may conflict with each other
should be articulated and joint decisions made as to how best to prioritize these needs.

130
Many of the organizations housing authorities liaise with for the reentry pilot programs, whether
nonprofits or correctional agencies, will be accustomed to viewing programs and services from
the perspective of formerly incarcerated individuals. (The two meaningful exceptions to this
would be any landlord or tenant organizations that the housing authorities might engage with.)
These differing perspectives must be addressed in order to jointly set goals and an appropriate
course of action. Discussing these different perspectives will provide the organizations with a
chance to improve not only principled engagement, specifically the elements of definition,
deliberation, and determination, via jointly made decisions based on open discussion, but more
importantly will also provide an opportunity to improve mutual understanding, an element of
shared motivation that Emerson et al., (2012) define as “refer[ing] to the ability to understand
and respect others’ positions and interests even when one might not agree” (p. 14), and as a
result also impact the strength of the shared commitment and mutual trust.
Recommendation 3: Establish protocols and procedures based on discussions with all
involved parties rather than individual agreements between the housing authority and
each organization.
While individual agreements between the housing authority and the partner organizations are
useful, and this recommendation should in no way be interpreted as an argument against their
use, there also needs to be a collective agreement that establishes procedural protocols and
institutional arrangements among the participating organizations. This does not mean that all of
the organizations will undertake the same actions but rather that there will be a coordinated
agreement delineating each organization’s role and establishing how best to accomplish the
group’s goals. This would help address deficiencies in the collaborative governance regime’s

131
capacity of joint action component, particularly facilitating group interactions, the subject of our
next recommendation, over time (Emerson et al., 2012).
Recommendation 4: The organizations should commit to frequent meetings as well as
other forms of communication, particularly at the beginning of the collaboration when the
partnership is most tenuous and ill-defined.
It is highly likely that given the distrust directed at housing authorities, and by extension their
reentry programs, that the success of any program will be gradual rather than immediate. In fact,
it may be that the lack of success of reentry programs is partially explained by the fact that these
programs may need many years to succeed and that the two or three years that many were
evaluated at was simply not enough time for them to gain significant momentum. Furthermore,
addressing the distrust of the housing authorities will require a significant amount of planning. If
the time it takes to actively create and implement the program as well as the time between these
actions, i.e. “lapsed time,” is not accounted for or takes longer than anticipated, it can act as a
demotivator (Huxham, 1996). Even once the program has been rolled out, if participation
continues to occur at a slow rate, it could discourage staff who may feel that it is a waste of their
time to participate in meetings when the program is not being subscribed to and it seems like
there’s nothing to report. However, maintaining engagement among the partner organizations
must remain a priority since such engagement encourages the sharing of information, allows the
organizations to quickly and collectively troubleshoot any problems, and provides more
opportunities to legitimize the entire endeavour, which would improve both the capacity for joint
action and shared motivation, respectively.
Recommendation 5: Efforts must be made to share resources among the organizations
when possible.

132
The sharing of resources can be crucial to the success of a collaboration (Thomson & Perry,
2006) and can be used to address power imbalances and resolve conflicts within the partnership
(Bryson et al., 2006). In order to leverage resources to best reinforce the legitimacy of the
partnership, it would be advantageous for housing authorities, when possible, to provide some
form of resources to its partners. In fact, this was mentioned by a HACLA staff member while
being interviewed: “I think maybe if we, you know, throw the nonprofits a bone or some kind of
funding maybe they'd have more of an involvement in it because right now just there is no
money tied to it. So, we just -- we did a memorandum of understanding, paper agreement, but
there is no money in the process. So, let[’s] say we gave each one of them $5,000 to implement
this, anyway that they wanted to do it.” Providing resources, whether in terms of financial
support, staff, supplies, or other items, would also help address the existing leadership vacuum as
accountability is inherently built into the transaction, which would allow HACLA to more
actively articulate expectations. The exchange of resources amongst the various partner
organizations should also be encouraged. This could range from actions as simple as including
all of the organizations’ information on flyers, websites, etc. even when they’re produced by a
single organization, to sharing the cost of mailings, to including other organizations in grant
applications, etc. Furthermore in situation where resources for reentry programs aren’t abundant,
sharing resources could help mitigate the issue and improve the capacity for joint action
(Emerson et al., 2012).
Recommendation 6: The format of the reentry program should address the fact that not all
of the families that participate intend for the formerly incarcerated family members’
residency within the home to be permanent.

133
The housing authority and its partner agencies must determine how best to programmatically
account for the differences in the types of family arrangements. The speediness of being granted
permission to join the household is important for all participants, but especially for those who are
only looking to live with a family member temporarily and for whom the program holds most
value immediately upon release. While it may be easiest for those reentry programs that are
based in public housing and thus more directly controlled by the housing authorities to account
for this when designing their programs, it is also possible, depending on each housing
authorities’ own regulations, for reentry programs that involve Section 8 to account for the
differences in these arrangements as well. For instance, one potential mechanism that could
address this would be to initially place automatic time limits on the length of residency rather
than allowing the formerly incarcerated individuals to be added to the household indefinitely
initially. Permission to live in the home would expire automatically at the end of the pre-
determined time period, perhaps three or six months or a year, unless the formerly incarcerated
individual had fulfilled all program requirements and remained in good standing. The
leaseholder would also still have to want the formerly incarcerated individual to remain in the
home. In order to guarantee that the leaseholder was still participating of their own freewill, they
could then be interviewed independently again in a process similar to the one HACLA enacted.
Eventually after a certain number of reviews, the formerly incarcerated family member could be
granted permanent permission to join the household. Building this caveat into the application
process would require more staff resources as the families would have to be re-evaluated on a
regular basis but could justify a quicker turnaround time as the long-term wellbeing of the family
would be less at risk. It would also provide a means for leaseholders who don’t necessarily want
to have to engage with the re-entry program long-term but who would like to provide temporary

134
shelter for a family member to do so. It might also help assuage concerns they themselves have
about adding the family member to the lease.

The recommendations above are intended as a starting off point to guide the creation of
new housing authority reentry programs and rework existing ones rather than as an ultimate
solution. It is almost certain that even if these particular issues are satisfactorily addressed,
others will arise, as befits the iterative nature of a collaborative governance regime. However, if
the actions suggested do in fact help improve the collaborative dynamics in addition to
addressing the specific concerns themselves, future problems should be more easily resolved in
the presence of a robust collaborative governance regime. Finally, while these programs have
heretofore had relatively low participation rates, concerted, sustained efforts should first be made
to address the host of factors contributing to the poor numbers before scrapping the programs all
together. These programs present a unique opportunity to utilize existing resources to fulfill a
significant need - an opportunity that should not be abandoned until additional attempts at
modifying and adapting them have first been made.

135
Conclusion

While these studies are centered around evaluations of three very different programs,
both in terms of their missions and the populations they serve, they are thematically linked by
their examination of innovative efforts using existing resources to address challenges faced by
severely disadvantaged populations. Moreover, they all utilize cross-sectoral partnerships to do
so and in doing so demonstrate the breadth and possibilities, as well as the potential challenges,
that these collaborations are capable of. Additionally, all three papers go beyond evaluations of
their respective programs and contribute to their respective wider literatures.
The School in the Park evaluation, for instance, is the first study that we are aware of that
exams the long-term outcomes of an arts-oriented and/or museum-based school program. It is
also one of the most rigorous evaluations that currently exists on the topic in its assessment of
short-term, as well as long-term, outcomes. Furthermore, its findings that students are not
negatively impacted by participating provides important evidence, at a time when the value of
such programming continues to be questioned, that participation need not be a choice between
allowing students to partake in experiential learning programs outside of a traditional classroom
setting and maximizing learning opportunities.
Similarly, the evaluation of the POPS program offers evidence, albeit more limited, that
students who participate in the program are not negatively impacted by their participation,
implying that students with familial histories of incarceration are not necessarily adversely
impacted when they socialize with peers who share those experiences. The study is also one of
only two known evaluations that look at school-based programs involving student interaction
rather than programs centered around mentoring activities and is the only one that examines
academic and behavioral outcomes among the students who participate. The POPS evaluation

136
further contributes to the literature by providing preliminary evidence that there is, in fact, a
systematic achievement gap between students who experience familial incarceration and those
who don’t; a gap which is often assumed to exist but for which evidence remains limited. POPS
students consistently underperformed academically in the years before they participated in POPS
when compared to their peers who never participated in POPS.
Finally, the study of HACLA’s reentry pilot program adds not only to the incipient
literature on these types of reentry housing programs which are being implemented by a variety
of housing authorities nationwide but uniquely does so via a collaborative governance context.
Despite the potential for collaborative governance arrangements to successfully engage in
flexible service provision, most collaborative governance studies have focused on conservation
or environmental partnerships and service provision. Few examine the potential for
collaborative governance in the fields of housing or reentry despite the fact that the public actors
in both of these of policy areas increasingly rely upon partnerships with the private and non-
profit sectors to implement their programs. This paper’s exploration of potential barriers to
successful collaborations has the potential to help undergird future endeavors in these specific
areas. Furthermore, the findings underscore the need for a greater emphasis on thoughtfully
engaging the community and enabling its participation throughout the entire process, from
problem identification through implementation and scaling, when attempting to implement
socially innovative programs. The findings also suggest that using community organizations as
proxies for the community itself in determining programmatic needs and goals as well as
identifying the means to achieve those goals is an insufficient substitute for true citizen co-
production.

137
References
Adams, M., Foutz, S., Luke, J., & Stein, J. (2006). Thinking Through Art Isabella Stewart
Gardner Museum School Partnership Program Year 3 Preliminary Research Results.
Edgewater, MD: Institute for Learning Innovation. https://vtshome.org/wp-
content/uploads/2016/08/3Gardner-Museum-Thinking-Through-Art.pdf
Agranoff, R. (2006). Inside Collaborative Networks: Ten Lessons for Public Managers. Public
Administration Review, 66(s1), 56–65. https://doi.org/10.1111/j.1540-6210.2006.00666.x
Agranoff, R., & McGuire, M. (2003). Inside the matrix: Integrating the paradigms of
intergovernmental and network management. International Journal of Public
Administration, 26(12), 1401-1422. https://doi.org/10.1081/PAD-120024403
Anderson, D., Piscitelli, B., Weier, K., Everett, M, & Taylor, C. (2002). Children's museum
experiences: Identifying powerful mediators of learning. Curator: The Museum Journal
45(3), 213-231. https://doi.org/10.1111/j.2151-6952.2002.tb00057.x
Andrews, R., & Entwistle, T. (2010). Does cross-sectoral partnership deliver? An empirical
exploration of public service effectiveness, efficiency, and equity. Journal of Public
Administration Research and Theory, 20(3), 679-701.
https://doi.org/10.1093/jopart/mup045
Aoyama, Y., & Parthasarathy, B. (2016). The rise of the hybrid domain: Collaborative
governance for social innovation. Edward Elgar Publishing.
Aurand, A., Emanuel, D., Yentel, D., Errico, E., Chapin, Z., Leong, G.M., & Rodriguez, K.
(2016). The Long Wait for Home. Washington, DC: National Low Income Housing
Coalition. https://nlihc.org/sites/default/files/HousingSpotlight_6-1_int.pdf

138

Ayob, N., Teasdale, S., & Fagan, K. (2016). How social innovation ‘came to be’:
Tracing the evolution of a contested concept. Journal of Social Policy, 45(4), 635-653.
https://doi.org/10.1017/S004727941600009X
Bae, J., diZerega, M., Kang-Brown, J., Shanahan, R., & Subramanian, R. (2016). (rep.) Coming
home: An evaluation of the New York City Housing Authority’s Family Reentry Pilot
Program. New York, NY: Vera Institute of Justice.
Bae, J., Finley, K., diZerega, M., & Kim, S. (2017). (rep.). Opening Doors: How to develop
reentry programs using examples from public housing authorities. New York, NY: Vera
Institute of Justice.
Bassok, D., Latham, S., & Rorem, A. (2016). Is kindergarten the new first grade?. AERA
Open, 2(1), 1-31. https://doi.org/10.1177/2332858415616358
Bekkers, V.J.J.M., Tummers, L.G., & Voorberg, W.H. (2013). From public innovation to social
innovation in the public sector: A literature review of relevant drivers and barriers.
Rotterdam: Erasmus University Rotterdam
Bergman, P., Chetty, R., Deluca, S., Hendren, N., Katz, L. F., & Palmer, C. (2020). Creating
Moves to Opportunity: Experimental evidence on barriers to neighborhood choice. The
National Bureau of Economic Research. https://doi.org/10.3386/w26164
Bingham, L. B., O’Leary, R., & Carlson, C. (2014). Frameshifting: Lateral thinking for
collaborative public management. In L. B. Bingham & R. O’Leary (Eds.) Big ideas in
collaborative public management. Routledge.

139
Bir, A., McKay, T., Lindquist, C., & Steffey, D. (2015). The Experiences of Families During a
Father’s Incarceration: Summary of Baseline Descriptive Findings. Washington, DC:
Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health
and Human Services. Retrieved from https://aspe.hhs.gov/system/files/pdf/137556/MFS-
IP%20BaselineReport.pdf
Booher-Jennings, J. (2005). Below the bubble: “Educational triage” and the Texas accountability
system. American Educational Research Journal, 42(2), 231-268.
https://doi.org/10.3102/00028312042002231
Bowen, D. H., Greene, J. P., & Kisida, B. (2014). Learning to think critically: A visual art
experiment. Educational Researcher, 43(1), 37-44,
http://doi.org/10.3102/0013189X1351267
Bradley, K. H., Oliver, R. B. M., Richardson, N. C., & Slayter, E. M. (2001). No place like
home: Housing and the ex-prisoner. Boston, MA: Community Resources for Justice.
Retrieved from https://b.3cdn.net/crjustice/a5b5d8fa98ed957505_hqm6b5qp2.pdf
Braman, D. (2004). Doing time on the outside: Incarceration and family life in urban America.
University of Michigan Press.
Brayne, S. (2014). Surveillance and system avoidance: Criminal justice contact and institutional
attachment. American Sociological Review, 79(3), 367-391.
https://doi.org/10.1177/0003122414530398
Brown, E. D., Benedett, B., & Armistead, M. E. (2010). Arts enrichment and school readiness
for children at risk. Early Childhood Research Quarterly, 25(1), 112-124,
http://doi.org/10.1016/j.ecresq.2009.07.008

140
Bryan, B. (2017). Paternal Incarceration and Adolescent Social Network
Disadvantage. Demography, 54(4), 1477-1501.
https://doi.org/10.1007/s13524-017-0589-8
Bryson, J. M., Crosby, B. C., & Stone, M. M. (2006). The design and implementation of Cross‐
Sector collaborations: Propositions from the literature. Public Administration
Review, 66(s1), 44-55. https://doi.org/10.1111/j.1540-6210.2006.00665.x
Burchenel, M., & Grohe, M. (2007). Thinking through art: Transforming museum curriculum.
The Journal of Museum Education 32(2), 111-122.
https://doi.org/10.1080/10598650.2007.11510563
California Department of Education. (2016, May 17). Unduplicated student poverty – Free or
reduced price meals data 2015–16 [Data file]. http://www.cde.ca.gov/ds/sd/sd/filessp.asp
California Department of Education. (2015, February 19). California Modified Assessment
(CMA) pilot test. Retrieved from http://www.cde.ca.gov/ta/tg/ sr/cmastar.asp
Cannon, J. S., Jacknowitz, A., & Painter, G. (2011). The effect of attending full‐day kindergarten
on English learner students. Journal of Policy Analysis and Management, 30(2), 287-309.
https://doi.org/10.1002/pam.20560
Cho, R. M. (2009a). The impact of maternal imprisonment on children’s educational
achievement results from children in Chicago public schools. Journal of Human
Resources, 44(3), 772-797. https://doi.org/10.3368/jhr.44.3.772
Cho, R. M. (2009b). Impact of maternal imprisonment on children's probability of grade
retention. Journal of Urban Economics, 65(1), 11-23.
https://doi.org/10.1016/j.jue.2008.09.004

141
Cho, R. M. (2010). Maternal incarceration and children’s adolescent outcomes: Timing and
dosage. Social Service Review, 84(2), 257-282.
Cho, R. M. (2011). Understanding the mechanism behind maternal imprisonment and adolescent
school dropout. Family Relations, 60(3), 272-289. https://doi.org/10.1111/j.1741-
3729.2011.00649.x
Clark, V. A. (2016). Predicting two types of recidivism among newly released prisoners. Crime
& Delinquency, 62(10), 1364–1400. https://doi.org/10.1177/0011128714555760
Comfort, M. (2007). Punishment beyond the legal offender. Annual Review of Law and Social
Science, 3, 271-296. https://doi.org/10.1146/annurev.lawsocsci.3.081806.112829
Conway, J. M., & Keays, A. (2015). Children with incarcerated parents: A quantitative
evaluation of mentoring and home-based counseling and case management services. New
Britain, CT: The Institute for Municipal and Regional Policy.
Cox-Petersen, A. M., Marsh, D. D., Kisiel, J., & Melber, L. M. (2003). Investigation of guided
school tours, student learning, and science reform recommendations at a museum of
natural history. Journal of Research in Science Teaching 40(2), 200-218.
https://doi.org/10.1002/tea.10072
Curtis, M. A., Garlington, S., & Schottenfeld, L. S. (2013). Alcohol, drug, and criminal history
restrictions in public housing. Cityscape, 15(3), 37-52.
https://www.jstor.org/stable/26326830
Dallaire, D. H., Ciccone, A., & Wilson, L. C. (2010). Teachers’ experiences with and
expectations of children with incarcerated parents. Journal of Applied Developmental
Psychology, 31, 281–290.

142
Davis, L., & Shlafer, R. J. (2017). Mental health of adolescents with currently and formerly
incarcerated parents. Journal of Adolescence, 54, 120-134.
https://doi.org/10.1016/j.adolescence.2016.10.006
Dawkins, C. J. (2004). Recent Evidence on the Continuing Causes of Black-White Residential
Segregation. Journal of Urban Affairs, 26(3), 379–400.
https://doi.org/10.1111/j.0735-2166.2004.00205.x
Day, S. E., Hinterland, K., Myers, C., Gupta, L., Harris, T. G., & Konty, K. J. (2016). A School-
Level Proxy Measure for Individual-Level Poverty Using School-Level Eligibility for
Free and Reduced-Price Meals. Journal of School Health, 86(3), 204-214.
https://doi.org/10.1111/josh.12371
Dewey, J. (1916). Democracy and Education. The Macmillan Company.
Dickson-Gomez, J., Convey, M., Hilario, H., Corbett, A. M., & Weeks, M. (2007). Unofficial
policy: access to housing, housing information and social services among homeless drug
users in Hartford, Connecticut. Substance Abuse Treatment, Prevention, and Policy, 2(1),
8. https://doi.org/10.1186/1747-597X-2-8
Directory of Programs Serving Children & Families of the Incarcerated. (n.d.). Retrieved May 5,
2019 from https://nrccfi.camden.rutgers.edu/resources/directory/
Domanski, D., Howaldt, J., & Kaletka, C. (2020). A comprehensive concept of social innovation
and its implications for the local context – on the growing importance of social
innovation ecosystems and infrastructures. European Planning Studies, 28(3), 454-474.
https://doi.org/10.1080/09654313.2019.1639397

143
Donovan, S. & Galante, C. (2011, June 17). Letter from Secretary Donovan to PHAs.
Department of Housing and Urban Development. Retrieved from
http://www.usich.gov/resources/uploads/asset_library/Rentry_letter_from_Donovan_to_P
HAs_6-17-11.pdf
Downey, S., Delamatre, J., & Jones, J. (2007). Measuring the impact of museum-school
programs: Findings and implications for practice. The Journal of Museum Education
32(2), 175-187. https://doi.org/10.1080/10598650.2007.11510567
Edwards-Schachter, M., & Wallace, M. L. (2017). ‘Shaken, but not stirred’: Sixty years of
defining social innovation. Technological Forecasting and Social Change, 119, 64-79.
https://doi.org/10.1016/j.techfore.2017.03.012
Emerson, K., & Gerlak, A. K. (2014). Adaptation in collaborative governance regimes.
Environmental Management, 54(4), 768-781. https://doi.org/10.1007/s00267-014-0334-7
Emerson, K., Nabatchi, T., & Balogh, S. (2012). An integrative framework for collaborative
governance. Journal of Public Administration Research and Theory, 22(1), 1-29.
https://doi.org/10.1093/jopart/mur011
Eshach, H. (2007). Bridging in-school and out-of-school learning: formal, non-formal, and
informal education. Journal of Science Education and Technology, 16(2), 171-190.
https://doi.org/10.1007/s10956-006-9027-1
Feldman, L., Bowes, J., Meads, C., Gabriel, K., & Mathison, C., (2010). School in the Park end-
of-year report 2009-10. School in the Park.
http://schoolinthepark.net/SITP/Reports/SITP%20End%20of%20Year%20Rpt%202009-
10.pdf

144
Evans, D. N., Blount-Hill, K.-L., & Cubellis, M. A. (2019). Examining housing discrimination
across race, gender and felony history. Housing Studies, 34(5), 761–778.
https://doi.org/10.1080/02673037.2018.1478069
Evans, D., & Porter, J. (2015). Criminal history and landlord rental decisions: A New York
quasi-experimental study. Journal of Experimental Criminology, 11, 21–42.
http://doi.org/10.1007/s11292-014-9217-4
Fischer, W., & Sard, B. (2017). Chartbook: Federal housing spending is poorly matched to need.
Center on Budget and Policy Priorities. Retrieved from
https://www.cbpp.org/research/housing/chart-book-federal-housing-spending- is-poorly-
matched-to-need
Frable, D. E., Platt, L., & Hoey, S. (1998). Concealable stigmas and positive self-perceptions:
feeling better around similar others. Journal of Personality and Social Psychology, 74(4),
909. https://doi.org/10.1037/0022-3514.74.4.909
Geller, A., Cooper, C. E., Garfinkel, I., Schwartz-Soicher, O., & Mincy, R. B. (2012). Beyond
absenteeism: Father incarceration and child development. Demography, 49(1), 49-76.
https://doi.org/10.1007/s13524-011-0081-9
Geller, A. & Curtis, M. (2011). A sort of homecoming: Incarceration and the housing security of
urban men. Social Science Research 40, 1196–1213.
https://doi.org/10.1016/j.ssresearch.2011.03.008
Geller, A., Garfinkel, I., Cooper, C. E., & Mincy, R. B. (2009). Parental incarceration and child
well‐being: Implications for urban families. Social Science Quarterly, 90(5), 1186-1202.

145
Geller, A., Garfinkel, I., & Western, B. (2011, February). Paternal incarceration and support for
children in fragile families. Demography, 48(1), 25-47. https://doi.org/ 10.1007/s13524-
010-0009-9
Gardner, H. (1991). The unschooled mind, Basic Books, New York.
Giordano, P.C., Copp, J.E., Manning, W.D., & Longmore, M.A. (in press). Linking parental
incarceration and family dynamics associated with intergenerational transmission: A life-
course perspective. Criminology. https://doi.org/10.1111/1745-9125.12209
Glaser, B.G., & Strauss, A.L. (1967) The Discovery of Grounded Theory: Strategies for
qualitative research. Routledge.
Glaze, L. E., & Kaeble, D. (2014). Correctional Populations in the United States, 2013. Bureau
of Justice Statistics, Washington, DC. Retrieved from
http://www.bjs.gov/content/pub/pdf/cpus13.pdf
Grieb, S. M. D., Crawford, A., Fields, J., Smith, H., Harris, R., & Matson, P. (2014). “The stress
will kill you”: Prisoner reentry as experienced by family members and the urgent need for
support services. Journal of Health Care for the Poor and Underserved, 25(3), 1183–
1200. https://doi.org/10.1353/hpu.2014.0118
Grimm, R., Fox, C., Baines, S., & Albertson, K. (2013). Social innovation, an answer to
contemporary societal challenges? Locating the concept in theory and
practice. Innovation: The European Journal of Social Science Research, 26(4), 436-455.
https://doi.org/10.1080/13511610.2013.848163
Gustafson, K. S. (2011). Cheating welfare: Public assistance and the criminalization of poverty.
NYU Press.

146
Hagan, J., & Dinovitzer, R. (1999). Collateral consequences of imprisonment for children,
communities, and prisoners. Crime and Justice, 26, 121-162.
https://doi.org/10.1086/449296
Hagan, J., & Foster, H. (2012). Intergenerational educational effects of mass imprisonment in
America. Sociology of Education, 85, 259–286.
https://doi.org/10.1177/0038040711431587
Hamlin, M. (2020). Second chances in the second city: Public housing and prisoner reentry in
Chicago. Environment and Planning D: Society and Space, 38(4), 587–606.
https://doi.org/10.1177/0263775820918274
Harwell, M., & Lebeau, B. (2010). Student Eligibility for a Free Lunch as an SES Measure in
Education Research. Educational Researcher, 39(2), 120-131.
https://doi.org/10.3102/0013189x10362578
Haskins, A. R. (2014). Unintended consequences: effects of paternal incarceration on child
school readiness and later special education placement. Sociological Science, 1(1), 141-
58. Retrieved from
https://sociologicalscience.com/download/volume%201/april/unintended-consequences-
effects-of-paternal-incarceration.pdf
Haskins, A. R., & Jacobsen, W. C. (2017). Schools as surveilling institutions? Paternal
incarceration, system avoidance, and parental involvement in schooling. American
Sociological Review, 82(4), 657-684. https://doi.org/10.1177/0003122417709294
Hein, G. E. (2004). John Dewey and museum education. Curator: The Museum Journal, 47(4),
413-427.

147
Herbert, C. W., Morenoff, J. D., & Harding, D. J. (2015). Homelessness and housing insecurity
among former prisoners. RSF: The Russell Sage Foundation Journal of the Social
Sciences, 1(2), 44-79. https://doi.org/10.7758/RSF.2015.1.2.04
Higdon, D., Goldstone, J., Meads, C., Gabriel, K., Jepson., T., & Mathison, C., (2017). SITP
final report 2016-17. School in the Park.
https://drive.google.com/file/d/0BwUBHlOrZnegOGlvOW5wY1JMUWM/view
Higdon, D., Goldstone, J., Meads, C., Gabriel, K., Jepson., T., & Mathison, C., (2018). SITP
final report 2017-18. School in the Park. http://schoolinthepark.net/wp-
content/uploads/2018/09/FINAL-End-of-Year-Rpt-17-18-with-Appendices-Reduced-
File-Size.pdf
Hiller, M. L., Knight, K., & Simpson, D. D. (1999). Prison-based substance abuse treatment,
residential aftercare and recidivism. Addiction, 94(6), 833-842.
https://doi.org/10.1046/j.1360-0443.1999.9468337.x
Hinojosa, M.S., Hinojosa, R., Bright, M., & Nguyen, J. (2019). Adverse childhood experiences
and grade retention in a national sample of US children. Sociological Inquiry, 89.
https://doi.org/ 10.1111/soin.12272
Hoffmann, H. C., Byrd, A. L., & Kightlinger, A. M. (2010). Prison programs and services for
incarcerated parents and their underage children: Results from a national survey of
correctional facilities. The Prison Journal, 90(4), 397-416.
https://doi.org/10.1177/0032885510382087
Huxham, C. (1996). Collaboration and collaborative advantage. In C. Huxham (Ed.), Creating
collaborative advantage (pp. 1-18). Sage Publications.

148
Huxham, C., & Vangen, S. (2005). Managing to collaborate: The theory and practice of
collaborative advantage. Routledge.
The International Coalition for Children with Incarcerated Parents. (n.d.) Find an organization.
Retrived from http://inccip.org/find-an-organization-2/by-location/north-america/
Johnson, E. I., & Easterling, B. (2012). Understanding unique effects of parental incarceration on
children: Challenges, progress, and recommendations. Journal of Marriage and Family,
74(2), 342-356. https://doi.org/10.1111/j.1741-3737.2012.00957.x
Johnston, D. (2012) Services for children of incarcerated parents. Family Court Review 50(1),
91–105. https://doi.org/10.1111/j.1744-1617.2011.01431.x
Jones, G. M., Jones, B. D., & Hargrove, T. (2003). The unintended consequences of high-stakes
testing. Rowman & Littlefield Publishers.
Kanevsky, L., Corke, M., & Frangkiser, L. (2008). The academic resilience and psychosocial
characteristics of inner-city English learners in a museum-based school program.
Education and Urban Society 40(4), 452-475.
Keene, D. E., Rosenberg, A., Schlesinger, P., Guo, M., & Blankenship, K. M. (2018). Navigating
limited and uncertain access to subsidized housing after prison. Housing Policy
Debate, 28(2), 199–214. https://doi.org/10.1080/10511482.2017.1336638
Keene, D. E., Smoyer, A. B., & Blankenship, K. M. (2018). Stigma, housing and identity after
prison. The Sociological Review, 66(4), 799–815.
https://doi.org/10.1177/0038026118777447
Kisida, B., Bowen, D. H., & Greene, J. P. (2016). Measuring critical thinking: Results from an
art museum field trip experiment. Journal of Research on Educational Effectiveness,
9(sup1), 171-187, http://doi.org/10.1080/19345747.2015.1086915

149
Kisida, B., Greene, J. P., & Bowen, D. H. (2014). Creating cultural consumers: The dynamics of
cultural capital acquisition. Sociology of Education, 87(4), 281-295,
http://doi.org/10.1177/0038040714549076
Kopak, A. M., & Smith-Ruiz, D. (2015). Criminal justice involvement, drug use, and depression
among African American children of incarcerated parents. Race and Justice, 6(2), 89-
116. https://doi.org/10.1177/2153368715586633
Korver-Glenn, E. (2018). Compounding inequalities: How racial stereotypes and discrimination
accumulate across the stages of housing exchange. American Sociological Review, 83(4),
627–656. https://doi.org/10.1177/0003122418781774
Kratz, S., & Merritt, E. (2012). On the Horizon: Museums and the future of education. An
American Alliance of Museums conceptual paper. Retrieved from http://www.aam-
us.org/docs/default-document-library/on-the-horizon-web-version.pdf?sfvrsn=0.
La Vigne, N.G., Visher, C. & Castro, J. (2004). Chicago prisoners’ experiences returning home.
Washington, DC: Urban Institute.
Ladd, H. F. (2012). Education and poverty: Confronting the evidence. Journal of Policy Analysis
and Management, 31: 203–227. https://doi.org/10.1002/pam.21615
Lageson, S. E. (2016). Found out and opting out: The consequences of online criminal records
for families. The ANNALS of the American Academy of Political and Social
Science, 665(1), 127-141. https://doi.org/10.1177/0002716215625053
Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2010). Prisoner reentry in the first decade of
the twenty-first century. Victims and Offenders, 5(3), 253-267.
https://doi.org/10.1080/15564886.2010.485907

150
Lee, R. D., Fang, X., & Luo, F. (2013). The impact of parental incarceration on the physical and
mental health of young adults. Pediatrics, 131(4), 1188-1195.
https://doi.org/10.1542/peds.2012-0627
Lutze, F. E., Rosky, J.W., & Hamilton Z. K. (2014). A multi-site outcome evaluation of
Washington State’s Reentry Housing Pilot Program for High Risk Offenders. Criminal
Justice and Behavior, 41(4), 471–491. https://doi.org/10.1177/0093854813510164
Major, B., & O'Brien, L. T. (2005). The social psychology of stigma. Annual Review of
Psychology, 56, 393-421. https://doi.org/10.1146/annurev.psych.56.091103.070137
Maruschak, L. M., & Minton, T. D. (2020). Correctional populations in the United States, 2017-
2018. Bureau of Justice Statistics, Washington, DC. Retrieved from
https://www.bjs.gov/content/pub/pdf/cpus1718.pdf
Marzano, R. J. (2004). Building background knowledge for academic achievement: Research on
what works in schools. Alexandria, VA: Association of Supervision and Curriculum
Development. http://www.ascd.org/ASCD/pdf/siteASCD/video/buildingacademic.pdf
Matheson, F.I., Doherty, S., & Grant, B. A. (2011). Community-based aftercare and return to
custody in a national sample of substance-abusing women offenders. American Journal
of Public Health, 101(6), 1126–1132. https://doi.org/10.2105/AJPH.2010.300094
Melber, L. M. (2003). Partnerships in science learning: Museum outreach and elementary gifted
education. Gifted Child Quarterly 47(4), 251-258.
https://doi.org/10.1177/001698620304700402
Metraux, S., & Culhane, D. P. (2004). Homeless shelter use and reincarceration following prison
release. Criminology Public Policy, 3(2), 139-160.
https://doi.org/10.1111/j.17459133.2004.tb00031.x

151
Metraux, S., Roman, C. G., & Cho, R. (2007). National symposium on homelessness research.
Washington, DC: U.S. Department of Housing and Urban Development.
Mowen, T. J., Stansfield, R., & Boman IV, J. H. (2019). Family matters: Moving beyond “if”
family support matters to “why” family support matters during reentry from
prison. Journal of Research in Crime and Delinquency, 56(4), 483-523.
https://doi.org/10.1177/0022427818820902
Mulgan, G. (2006). The process of social innovation. Innovations: Technology, governance,
globalization, 1(2), 145-162.
https://www.mitpressjournals.org/doi/pdf/10.1162/itgg.2006.1.2.145
Murphey, D., & Cooper, P. M. (2015). Parents behind bars. Child Trends. Washington, DC.
Retrieved from http://www.childtrends.org/wp-content/uploads/2015/10/2015-
42ParentsBehindBars.pdf
Murray, J., Farrington, D. P., & Sekol, I. (2012). Children's antisocial behavior, mental health,
drug use, and educational performance after parental incarceration: a systematic review
and meta-analysis. Psychological Bulletin, 138(2), 175-210.
http://doi.org/10.1037/a0026407
Murray, J., & Farrington, D. P. (2008). The effects of parental imprisonment on children. Crime
and Justice, 37(1), 133-206. https://doi.org/10.1086/520070
Naser, R. L., & La Vigne, N. G. (2006). Family support in the prisoner reentry process. Journal
of Offender Rehabilitation, 43(1), 93–106. https://doi.org/10.1300/j076v43n01_05
National Gallery of Art. (2019). Art Around the Corner. Retrieved from
https://www.nga.gov/education/teachers/art-around-the-corner.html

152
Navarro, M. (2014, Nov. 9). Push to bring ex-convicts and families together lags. The New York
Times. Retrieved from http://www.nytimes.com/2014/11/10/nyregion/ push-to-bring-ex-
convicts-and-families-together-lags.html
Nelson, M., Deess, P., & Allen, C. (1999). The first month out: Post-incarceration experiences
in New York City. New York: Vera Institute of Justice.
Nichols, E. B., & Loper, A. B. (2012). Incarceration in the household: Academic outcomes of
adolescents with an incarcerated household member. Journal of Youth and Adolescence,
41(11), 1455-1471. https://doi.org/10.1007/s10964-012-9780-9
Ondrich, J., Ross, S. L., & Yinger, J. (2000). Now you see it, now you don't: Why do real estate
agents withhold available houses from Black customers? Review of Economics and
Statistics, 85(4), 854–73. https://doi.org/10.2139/ssrn.1808064
Page, S. (2004). Measuring accountability for results in interagency collaboratives. Public
Administration Review, 64(5), 591-606.
https://doi.org/10.1111/j.1540-6210.2004.00406.x
Paris, S. G., Yambor, K. M., & Packard, B. W. (1998). Hands-on biology: A museum-school-
university partnership for enhancing students' interest and learning in science. The
Elementary School Journal 98(3), 267-288. https://doi.org/10.1086/461894
Parke, R.D. & Clarke-Stewart, A. (2003). The effects of parental incarceration on children:
perspectives, promises, and policies. In J. Travis & M. Waul (Eds.), Prisoners once
removed: The impact of incarceration and reentry on children, families, and
communities. The Urban Institute.
Pew Charitable Trusts. (2010). Collateral costs: Incarceration's effect on economic mobility.
Washington, DC: Author.

153
Phillips, S. D., & Gates, T. (2011). A conceptual framework for understanding the stigmatization
of children of incarcerated parents. Journal of Child and Family Studies, 20(3), 286-294.
https://doi.org/10.1007/s10826-010-9391-6
Pol, E., & Ville, S. (2009). Social innovation: Buzz word or enduring term?. The Journal of
Socio-economics, 38(6), 878-885. https://doi.org/10.1016/j.socec.2009.02.011
POPS the Club. (2020). Our Clubs. https://popsclubs.org/clubs/
POPS the Club. (2019). POPS Curriculum. http://popstheclub.com/the-writing-program/
Pumpian, I., Fisher, D., & Wachowiak, S. (Eds.). (2006). Challenging the classroom standard
through museum-based education: School in the Park. Routledge.
https://doi.org/10.4324/9781410617279
Quinn, D. M., & Chaudoir, S. R. (2009). Living with a concealable stigmatized identity: the
impact of anticipated stigma, centrality, salience, and cultural stigma on psychological
distress and health. Journal of personality and social psychology, 97(4), 634.
https://doi.org/10.1037/a0015815
Randi Korn & Associates, Inc. (2004). Dinosaurs and More school tour: Findings from
questionnaires, interviews and observations.
https://www.informalscience.org/sites/default/files/Peabody_Final_distribution_copy.pdf
Randi Korn & Associates, Inc. (2010). Educational research: Evaluation of Traveling Through
Time, a school program of the Museum of the City of New York.
https://www.informalscience.org/sites/default/files/2010_RKA_MCNY_TTT_Ed_Rsch_f
inal.pdf

154
Ricks, M. (2017, January 23). “Reentry program inks public housing deal.” New Haven
Independent. Retrieved from
http://www.newhavenindependent.org/index.php/archives/entry/prison_entry/
Robrecht, L.C. (1995). Grounded theory: Evolving methods. Qualitative Health Research, 5(2),
169-177. https://doi.org/10.1177/104973239500500203
Roettger, M. E., & Boardman, J. D. (2012). Parental incarceration and gender-based risks for
increased body mass index: Evidence from the National Longitudinal Study of
Adolescent Health in the United States. American Journal of Epidemiology, 175(7), 636-
644. https://doi.org/10.1093/aje/kwr409
Roman, C.G., & Travis, J. (2004). Taking stock: housing, homelessness, and prisoner reentry.
Washington, DC: The Urban Institute.
Rosen, E. (2014). Rigging the Rules of the Game: How Landlords Geographically Sort Low-
Income Renters. City & Community, 13(4), 310–340. https://doi.org/10.1111/cico.12087
Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004). Evaluation: A systematic approach. Sage
publications.
San Antonio Housing Authority. (2020). SAHA 2025: Impact and Innovation Plan (Draft).
Retrieved from https://www.saha.org/wp-
content/uploads/2020/02/DRAFT_StrategicPlan2020-2025.pdf
San Antonio Housing Authority. (2019). Moving to work: Annual MTW plan – FY 2020.
Retrieved from https://saha.org/wp-content/uploads/2020/01/sanantoniofy20plan.pd
San Diego Unified School District. (n.d.). About San Diego Unified.
https://www.sandiegounified.org/about-san-diego-unified

155
Scally, C. P., Batko, S., Popkin, S. J., & DuBois, N. (2018). The case for more, not less:
Shortfalls in federal housing assistance and gaps in evidence for proposed policy
changes. Urban Institute, 3. Retrieved from
Schneider, M., Scholz, J., Lubell, M., Mindruta, D., & Edwardsen, M. (2003). Building
consensual institutions: Networks and the National Estuary Program. American Journal
of Political Science 47(1), 143–58. https://doi.org/10.1111/1540-5907.00010
School in the Park. About School in the Park. (n.d.). Retrieved from
http://schoolinthepark.net/about-sitp/
Shlafer, R., Duwe, G., & Hindt, L. (in press). Parents in prison and their minor children:
Comparisons between state and national estimates. The Prison Journal.
https://doi.org/10.1177/0032885519836996
Snyder, T., & Muso-Gillete, L., (2015, April 16). Free or reduced price lunch: A proxy for
poverty? National Center for Educational Statistics Blog. Retrieved from
https://nces.ed.gov/blogs/nces/post/free-or-reduced-price-lunch-a-proxy-for-poverty
Solomon, A., Visher, C., La Vigne, N., & Osborne, J. (2006). Understanding the challenges of
prisoner reentry: Research findings from the Urban Institute's prisoner reentry portfolio.
Urban Institute. Retrieved from http://www.urban.org/url.cfm?ID=411289
Springer, D. W., Lynch, C., & Rubin, A. (2000). Effects of a solution-focused mutual aid group
for Hispanic children of incarcerated parents. Child and Adolescent Social Work Journal,
17(6), 431-442. https://doi.org/10.1023/A:1026479727159
Sugie, N. F. (2015). Chilling effects: Diminished political participation among partners of
formerly incarcerated men. Social Problems, 62(4), 550-571.
https://doi.org/10.1093/socpro/spv017

156
Swisher, R. R., & Waller, M. R. (2008). Confining fatherhood incarceration and paternal
involvement among nonresident white, African American, and Latino fathers. Journal of
Family Issues, 29(8), 1067-1088. https://doi.org/10.1177/0192513X08316273
Thomson, A. M. (1999). AmeriCorps organizational networks: Six case studies of Indiana
AmeriCorps programs. National Service Fellows Program. Report for the Corporation
for National Service, Washington, DC: Corporation for National and Community
Service.
Thomson, A. M., & Perry, J. L. (2006). Collaboration processes: Inside the black box. Public
administration review, 66(s1), 20-32. https://doi.org/10.1111/j.1540-6210.2006.00663.x
Thomson, A. M., Perry, J. L., & Miller, T. K. (2009). Conceptualizing and measuring
collaboration. Journal of Public Administration Research and Theory, 19(1), 23-56.
https://doi.org/10.1093/jopart/mum036
Tracey, P. & Stott, N. (2017). Social innovation: a window on alternative ways of organizing and
innovating. Innovation, 19(1), 51-60. https://doi.org/10.1080/14479338.2016.1268924
Tran-Leung, M. (2015). When discretion means denial. A national perspective on criminal
records barriers to federally subsidized housing. Chicago, IL: Sargent Shriver National
Center on Poverty Law.
Trice, A. D., & Brewster, J. (2004). The effects of maternal incarceration on adolescent children.
Journal of Police and Criminal Psychology, 19(1), 27-35.
https://doi.org/10.1007/BF02802572
Turney, K. (2017). The unequal consequences of mass incarceration for
children. Demography,54(1), 361-389. https://doi.org/10.1007/s13524-016-0543-1

157
Turney, K., & Goodsell, R. (2018). Parental incarceration and children’s wellbeing. The Future
of Children, 28(1), 147-164. https://doi.org/10.1353/foc.2018.0007
Turney, K., & Haskins, A. R. (2014). Falling behind? Children’s early grade retention after
paternal incarceration. Sociology of Education, 87(4), 241-258.
https://doi.org/10.1177/0038040714547086
Turney, K., & Schneider, D. (2016). Incarceration and household asset ownership. Demography,
53(6), 2075-2103. https://doi.org/10.1007/s13524-016-0519-1
Turney, K., & Wildeman, C. (2013). Redefining relationships explaining the countervailing
consequences of paternal incarceration for parenting. American Sociological Review,
78(6), 949-979. https://doi.org/ 10.1177/0003122413505589
Turney, K., & Wildeman, C. (2015). Detrimental for some? Heterogeneous effects of maternal
incarceration on child wellbeing. Criminology & Public Policy, 14(1), 125-156.
https://doi.org/10.1111/1745-9133.12109
U.S. Government Accounting Office. (2009). Access to arts education: Inclusion of additional
questions in education’s planned research would help explain why instruction time has
decreased for some students (GAO-09-286). https://www.gao.gov/new.items/d09286.pdf
Visher, C. A., Lindquist, C., & Brumbaugh, S. (2007). Lessons learned from SVORI: Program
director perspectives on implementing reentry programming. Corrections Today
Magazine, 69(2), 92-97.
Voorberg, W. H., Bekkers, V. J. J. M., & Tummers, L. G. (2014). A systematic review of co-
creation and co-production: Embarking on the social innovation journey. Public
Management Review, 17(9), 1333–1357. https://doi.org/ 10.1080/14719037.2014.930505

158
Wakefield, S., & Wildeman, C. (2014). Children of the prison boom: Mass incarceration and the
future of American inequality. Oxford University Press.
Weaver, V. M., & Lerman, A. E. (2010). Political consequences of the carceral state. American
Political Science Review, 104(4), 817-833. https://doi.org/10.2307/40982899
Weinstein, M., Whitesell, E. R., & Schwartz, A. E. (2014). Museums, Zoos, and Gardens How
Formal-Informal Partnerships Can Impact Urban Students’ Performance in
Science. Evaluation review, 38(6), 514-545. https://doi.org/10.1177/0193841X14553299
Western, B., Braga, A. A., DRvis, J., & Sirois, C. (2015). Stress and hardship after
prison. American Journal of Sociology, 120(5), 1512-1547.
https://doi.org/10.1086/681301
Whitten, T., Burton, M., Tzoumakis, S., & Dean, K. (2019). Parental offending andcChild
physical health, mental health, and drug use outcomes: A systematic literature
review. Journal of Child and Family Studies, 28(5), 1155-1168.
https://doi.org/10.1007/s10826-019-01388-7
Wildeman, C. (2010). Paternal incarceration and children's physically aggressive behaviors:
Evidence from the Fragile Families and Child Wellbeing Study. Social Forces, 89(1),
285-309. https://doi.org/10.1353/sof.2010.0055
Wildeman, C., & Turney, K. (2014). Positive, negative, or null? The effects of maternal
incarceration on children’s behavioral problems. Demography, 51(3), 1041-1068.
Winner, E., & Cooper, M. (2000). Mute those claims: No evidence (yet) for a causal link
between arts study and academic achievement. Journal of Aesthetic Education, 34(3/4),
11-75, http://doi.org/10.2307/3333637
Yin, R. K. (2009). Case study research: Design and methods (5
th
ed.). Sage Publications.

159
Yinger, J. O. (1986). Measuring racial discrimination with fair housing audits: Caught in the
act. The American Economic Review, 76(5), 881–893.
https://www.jstor.org/stable/1816458
Zhang, S. X., Roberts, R. E. L., & Callanan, V. J. (2006). Preventing parolees from returning to
prison through community-based reintegration. Crime and Delinquency 52. 551–71.
https://doi.org/10.1177/0011128705282594
Ziegler, R. (2017). Social innovation as a collaborative concept. Innovation: The European
Journal of Social Science Research, 30(4), 388-405.
https://doi.org/10.1080/13511610.2017.134

Abstract (if available)

Abstract Social innovation, which consists of both the process and the solutions that result from the process itself, has the potential to address heretofore intractable social problems and result in more equitable outcomes and fundamental system change. This dissertation evaluates three such programs intended to assist different at-risk populations in Southern California. While these populations vary as do the means of the intervention as well as the form of social innovation typified by each, all three programs use cross-sectoral partnerships and attempt to leverage existing resources in order to address the issue at hand. These evaluations are intended not only to ascertain the effectiveness of the programs themselves but also to contribute to the literature on how best to utilize existing resources in novel and replicable ways in order to improve the lives of marginalized groups.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Technological innovation in public organizations

PDF

Faith-based promising practices: innovative forms of collaborative social services

PDF

The context of leadership in the development of California’s innovation hubs

PDF

The functions of the middleman: how intermediary nonprofit organizations support the sector and society

PDF

Philanthropic foundations and social impact bonds: understanding impact investment approaches among philanthropic foundations

PDF

Levels of interest: the effects of teachers' unions on Congress, statehouses, and schools

PDF

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

PDF

Creativity and innovation in undergraduate education: an innovation study

PDF

Reproducing fear amid fears of reproduction: the Black maternal body in U.S. law, media, and policy

PDF

Girls in STEM: the underrepresented trajectory in Tennessee: an innovation study

PDF

The role of public policy in the decisions of parents and caregivers: an examination of work, fertility, and informal caregiving

PDF

Racial disparities in U.S. K-12 suburban schools: an evaluation study

PDF

Improving post-secondary success for first generation college students through community partnerships: programming practices for charter high schools

PDF

The perception of innovation in the delivery of services for Hawaiian students

PDF

The institutional context of Korean philanthropy and the role of government and (quasi-) community foundations

PDF

Location of warehouses and environmental justice: Three essays

PDF

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

PDF

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

PDF

Adoption and implementation of innovative diagnostic tools for Alzheimer's Disease: challenges and barriers in primary care

Asset Metadata

Creator Williams, Brittany Danielle (author)

Core Title Innovation and good intentions: evaluations of three cross-sectoral programs for at-risk populations in Southern California

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Public Policy and Management

Publication Date 03/09/2021

Defense Date 01/19/2021

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

Tag children of the incarcerated,cross-sectoral partnerships,familial incarceration,formerly incarcerated,Housing,museum-based education,OAI-PMH Harvest,organizational analysis,prisoner re-entry,school outcomes,Social Innovation

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Painter, Gary (committee chair), Beckman, Christine (committee member), Esparza, Nicole (committee member)

Creator Email bdaniellewilliams@gmail.com,bdwillia@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-425522

Unique identifier UC11668572

Identifier etd-WilliamsBr-9310.pdf (filename),usctheses-c89-425522 (legacy record id)

Legacy Identifier etd-WilliamsBr-9310.pdf

Dmrecord 425522

Document Type Dissertation

Rights Williams, Brittany Danielle

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