A "second chance" program for vulnerable young adults: a program evaluation and a randomized controlled trial of adjunctive motivational interviewing to improve retention

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

A “SECOND CHANCE” PROGRAM FOR VULNERABLE YOUNG ADULTS: A
PROGRAM EVALUATION AND A RANDOMIZED CONTROLLED TRIAL OF
ADJUNCTIVE MOTIVATIONAL INTERVIEWING TO IMPROVE RETENTION

BY
Caitlin Smith

___________________________________________________

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
(PSYCHOLOGY)
May 2014
Copyright 2014 Caitlin Smith
ii
Acknowledgements
I am overwhelmed by the abundance of support I have received during the five-year
journey that has culminated in this dissertation. First, I owe great thanks to my advisor, Dr.
Stanley Huey, Jr., for seeing promise in me, steering my doctoral voyage, and sharing his
perceptive views on experimental design and clinical science. I am leaving USC a better writer
and a clearer thinker because of you. From the first conversation we had during my admissions
interview to our last dissertation revision meeting, you helped me balance a heart for activism
with the mind of a scientist, and I know this will allow me to make real positive changes in the
community moving forward. Also, Dr. Janet Schneiderman offered me so many opportunities to
develop as a researcher, that I could never fully express my thanks. I have learned to think
interdisciplinarily, work productively, and find joy in the academic process, under your
mentorship.
I deeply appreciate the wise advice and insightful questions from the other valuable
members of my dissertation guidance committee: Drs. Gayla Margolin, Donna Spruijt-Metz,
Wendy Wood, Margy Gatz, and Gale Sinatra. No matter how many hours I had spent writing and
re-writing my dissertation proposal or qualifying exam, the discerning and astute comments you
all made prompted me to think about my research in new ways and improve it beyond what I
could have imagined without you. Some of you also taught me in courses or supervised me
clinically, and the contributions you made in those contexts echo in this dissertation, as well.
I would also like to acknowledge my Huey lab-mates, especially Dr. Dawn McDaniel,
Taona Chithambo, and Eddie Jones. Dawn was my first mentor at USC, teaching me how to run
a study, see a client, and navigate South Central Los Angeles in one head-spinning summer.
iii
Taona and Eddie were my office-mates, and office-mates are the only people who really know
what you’re going through, what to say to lift your spirits, and when a conversation break is the
most important priority on the agenda. My cohort-mates (Larissa Borofsky Del Piero, Ilana
Kellerman Moss, Karan Singh, and the late John Keefe) are also some of the most important
people in my life, partly because we slogged through our academic and clinical training together,
but mostly because from the day we met, we formed a little family.
Family has always been my launching pad for great things, and I must acknowledge that
the love and support given me by Leroy, Kathi, Danny, and Matt Smith over my life have been at
the root of my capacity to complete a doctoral program. Their own accomplishments inspire me,
their proofreading improves my work, and time spent with them replenishes my energy. I am
very excited to add a new person to my family, Dr. Phil Sayegh, who I met through the USC
Psychology Department, as well. I feel so lucky to have a fiancé who knows why I’m stressing
out about non-normal distributions and can actually help me solve my statistical problems!
Funding for this dissertation was provided by the Dornsife College Doctoral Fellowship,
the Kellerman College Merit Fellowship, the Fahs-Beck Fund for Research and Experimentation
at the New York Community Trust, and the Society for the Psychological Study of Social Issues.
I am very thankful to these organizations for thinking this study was worthwhile and giving me
the tools to complete it. The research assistants who volunteered their time and intelligence to
this study were extremely important, as well: Erica Zara, Sarah Redmond, Nadia Naim, Nicole
Chow, Tonia Nguyen, Danielle Lindo, Michael Cox, Vicky Wen, Leslie Pennington, Ryan
Murakami, Robert “Augie” Craig, Darlyn Zenteno Aguilar, Carla Miller, Heidi Mayen, Holly
Schmidt, Amanda Haywood, Joanna Liang, Alex Ovadia, and Lily Mkhitarian. These students
took what they learned on this study, and are already in the process of becoming psychologists,
iv
social workers, marriage and family therapists, occupational therapists, and psychiatric nurses. I
can never thank them enough for their dedication and I am so excited to see what they achieve in
their careers. Also, this dissertation would not have been possible without the contributions of
Dr. Elizabeth Barnett. She volunteered hours of her time helping me with study design,
interventionist training, and making sense of the results we found. It has been so wonderful to
meet a kindred spirit and share the excitement of studying change talk with a colleague!
Finally, I would like to acknowledge and thank the Los Angeles Conservation Corps for
inviting me into their organization, especially Dan Knapp, Reyna Rauda-Trout, Andrea Klein,
Charles Ramos, and the 100 corps members who agreed to participate in this study. The trust and
openness you offered me is deeply appreciated.

v
Table of Contents
Acknowledgements ii
Table of Contents v
List of Tables vi
List of Figures vii
Dissertation Abstract viii
Chapter One: An Evaluation of an Education and Employment Program for Young 1
Adults: Promise and Challenges

Abstract 1
Background and Significance 2
Method 6
Results 12
Discussion 16
References 22
Chapter Two: A Randomized Controlled Trial of Motivational Interviewing to Improve 32
Program Retention
Abstract 32
Background and Significance 33
Method 36
Results 45
Discussion 51
References 58
Appendix A: MI Counseling Protocol 66
Appendix B: Placebo Counseling Protocol 68

vi
List of Tables
Table 1: Baseline characteristics 13
Table 2: High school diploma earning at LACC 14
Table 3: Pre-post tests for self-report outcomes 15
Table 4: Risk factors for program non-retention for negative or neutral reasons 17
Table 5: Baseline characteristics 46
Table 6: Intervention fidelity ratings 47
Table 7: Tests of MI effects on change talk, primary outcomes, and secondary 48
outcomes
Table 8: Links between change talk and outcomes 49
Table 9: Moderation results 50

vii
List of Figures
Figure 1: CONSORT flow diagram 37
Figure 2: Effect of MI on retention moderated by preference for consistency 51

viii
Dissertation Abstract
High school dropout puts young adults at risk for unemployment, criminal involvement,
and substance use. Education and employment programs, sometimes called “second chance”
programs, may help vulnerable young adults obtain high school diplomas, develop job skills, and
reduce problem behaviors. With roots in the United States depression era, Conservation Corps
programs are some of the most well-developed examples of education and employment
programs. However, the last published evaluation of the Conservation Corps was conducted in
the 1990s. Questions regarding the continuing effectiveness of the Conservation Corps must be
addressed, especially because there are some indications from the education and employment
program literature that retention may be a challenge. It is important to investigate whether failure
to complete programs like the Conservation Corps bodes poorly for participant outcomes, to
identify risk factors for program non-retention, and to test adjunctive strategies for assisting
participants in engaging and benefitting from these programs. Existing empirical and theoretical
work suggests Motivational Interviewing (MI) could encourage participants to use change talk,
or motivational language in favor of succeeding at the Conservation Corps. Change talk could
then trigger a drive in participants to reduce any cognitive dissonance arising from behaviors
inconsistent with that change talk, by working toward their stated educational and occupational
goals.
This dissertation consists of two studies conducted at the Los Angeles Conservation
Corps (LACC), with 100 young adults. The first study is an evaluation of LACC, which
characterized the educational attainment occurring in the program and examined pre-post
changes in problem behaviors. Program retention rates were calculated, and differences in pre-
post changes between program completers and non-completers were examined. Pre-treatment
ix
risk factors for program non-retention were also identified. Results indicated that a substantial
minority of participants succeeded in earning a high school diploma but that even more
participants failed to complete the program. Also, program completers succeeded in reducing
their own antisocial behavior and gang membership. Pre-treatment risk factors for program non-
retention included more years of education and more antisocial behaviors. Participants who were
gang members at program entry were more successful at completing the program and earning a
high school diploma than non-gang members. These results suggest that LACC offers some
vulnerable young adults an important educational opportunity and a possible pathway to criminal
and gang desistance. However, there appears to be a need for improving program retention, so
that more participants can benefit from the Conservation Corps.
The second study is a randomized controlled trial of MI for improving program retention.
After completing a measure of dissonance susceptibility, participants were randomly assigned to
one of three conditions: a) one session of MI designed to elicit change talk, b) one placebo
counseling session designed not to elicit change talk, or c) no additional treatment. Change talk
was tested as a mechanism of change and dissonance susceptibility was tested as a moderator of
MI efficacy. Results indicated that MI was not effective at improving program retention, even
though it did promote change talk. Participants who were the most susceptible to dissonance
induction were marginally more likely to complete the 8-week program orientation after
receiving MI counseling. These studies are helpful in illuminating the promise of education and
employment programs for vulnerable young adults. However, they highlight difficulties in
promoting retention. Future researchers should continue to investigate the efficacy of
Conservation Corps programs and to test MI in new contexts, with attention to guiding
theoretical frameworks and change talk as a mechanism of change.
1
CHAPTER ONE: AN EVALUATION OF AN EDUCATION AND EMPLOYMENT
PROGRAM FOR YOUNG ADULTS: PROMISE AND CHALLENGES
Abstract
Community programs, which offer educational opportunities and job training to
vulnerable young adults, aim to improve employment prospects and related outcomes. However,
research on the effects of these programs is limited. An evaluation of the Los Angeles
Conservation Corps (LACC) was conducted in order to assess the promise of education and
employment programs. Using archival and self-report data from 100 participants, this evaluation
assessed educational attainment during the program and tested for reductions in substance use,
antisocial behavior, and gang membership. Furthermore, pre-post changes were compared
between participants who completed versus did not complete 22 weeks at LACC, to assess the
importance of program retention. Finally, factors that reduced the likelihood of retention were
identified. Results indicated that 28% of participants earned a diploma through LACC within 22
weeks, and that participants achieved significant reductions in antisocial behavior and gang
involvement. However, only about three-quarters of participants completed the program
orientation, and only about two-thirds completed 22 weeks in the program. Participants who
were not retained did not reduce their gang membership or antisocial behavior. More antisocial
behavior and more years of education at program entry predicted a lower likelihood of retention.
Participants who were initially gang-involved were more likely to be retained and to earn
diplomas. These results suggest that LACC offers vulnerable young adults an important
educational opportunity and a possible pathway to criminal and gang desistance. Also, LACC
may best fit the needs of gang members and those with less education. Future research should
include investigations of whether program staff can implement strategies to improve retention.
2
Background and Significance
The United States (US) economy is characterized by declining earning capacity and
growing unemployment for young adults who have dropped out of high school (Sum,
Khatiwada, McLaughlin, & Palma, 2011). Even before the recent recession, young adults
without diplomas had difficulty securing well-paid positions (Farley, 1996; Morris & Western,
1999) or any type of employment at all (Fussell & Furstenberg, 2005). In the post-recession
environment, US unemployment rates range from 9.6-12.0% for individuals without high school
diplomas, compared to 6.3-8.3% for high school graduates, and 3.2-3.8% for college graduates
(US Department of Labor, 2014).
There is substantial evidence that high school dropout often precedes later criminal
involvement and substance use (e.g., Henry, Knight, & Thornberry, 2012). Qualitative studies
have found that criminal offenders often attribute their own antisocial behavior to a lack of
education (Laub & Sampson, 2003; Ronka, Oravala, & Pulkkinen, 2002), and quantitative
criminology research suggests unemployment subsequent to high school dropout may present a
path to criminal offending (Bernburg & Krohn, 2003; Cohen, Chen, Hamigami, Gordon, &
McArdle, 2000). High school dropout also increases risk for alcohol-related problems (Bingham,
Shope, & Tang, 2005; Casswell, Pledger, & Hooper, 2003; Crum, Helzer, & Anthony, 1993;
Harford, Hilton, & Yi, 2006) and other substance use (Fothergill et al., 2008; Obot, Hubbard, &
Anthony, 1999). Gang-involved young adults who drop out of high school may be at even
greater risk for antisocial behavior and substance use, because of the facilitative effect of gang
membership on these behaviors (Bendixen, Endresen, & Olweus, 2006; Gatti, Tremblay, Vitaro,
& McDuff, 2005; Harper, Davidson, & Hosek, 2008).
3
According to the most recent US census, Mexican Americans and African Americans
have the highest high school dropout rates (30.0% and 29.4%, respectively), compared to an
overall dropout rate across all ethnicities of 16.2% (Woo & Sakamoto, 2010). Subsequent
“idleness,” or disconnection from any educational or employment institutions, is also high
among African Americans (13.6%) and Mexican Americans (10.6%), as well as among Native
Americans (13.1%), other Latinos (10.1%), and Laotians (10.1%) (Woo & Sakamoto, 2010).
Furthermore, high school dropout status appears to confer more risk for ethnic minority
individuals compared to Whites. For example, 21% of African American male dropouts are later
imprisoned versus 2.9% of White male dropouts (Pettit & Western, 2004). This increased risk of
arrest for ethnic minority high school dropouts could contribute to what some have termed the
“school-to-prison pipeline” (Advancement Project, 2010).
Education and Employment Programs for Young Adults: Promise and Challenges
Most individuals without high school diplomas enter the adult labor market with no
previous part-time or summer employment, and are therefore disadvantaged in terms of both
education and work experience (Entwisle, Alexander, & Olson, 2000; Staff & Mortimer, 2007;
Sum et al., 2011). However, dropping out of high school does not necessarily doom young adults
to persistently low educational attainment and unemployment. Rather, they can improve their
outcomes by completing high school later, earning General Educational Development (GED)
certification, taking post-secondary courses, or obtaining early labor force experience (Bloom,
2010; Boudett, Murnane, & Willett, 2000; Kerckhoff & Bell, 1997). Some evidence suggests that
providing young unemployed adults with high quality, full-time work may be key to reducing
risk for criminal involvement and substance use (Lustig & Liem, 2010; Staff et al., 2010).
“Second chance” programs, which provide young adults with education and employment
4
opportunities, may be able to address skill and knowledge gaps, increase employability, and
reduce antisocial behavior and substance use (Bloom, 2010; Edelman, Holzer, & Offner, 2006).
The Los Angeles Conservation Corps (LACC), one such “second chance” program, is a
community-based education and employment organization modeled on President Franklin
Delano Roosevelt’s Civilian Conservation Corps (CCC). Between 1933 and 1942, the CCC
provided 2.5 million unemployed young men with work completing outdoor projects intended to
benefit the country (e.g., constructing hiking trails, planting trees, building dams) (Hendrickson,
2003). Today’s Conservation Corps programs, including LACC, often emphasize education and
community service in addition to paid work experience (The Corps Network, 2014). In a
randomized controlled trial of young adults across four Conservation Corps sites, corps members
had higher employment rates, worked more hours, and were less likely to be arrested at a 15-
month follow-up assessment compared to non-corps members (Jastrzab, Masker, Blomquist, &
Orr, 1996).
However, LACC faces a challenge in retaining participants. In 2011, LACC staff
reported a program retention rate of about 50%, prompting the board of directors to identify
improving retention as an important organizational goal. Client retention rates reported in the
education and employment program literature have been quite low. A large number of the
original CCC members “deserted” the corps (Hendrickson, 2003), only about 75% of participants
are retained in education and employment programs over the first three months (Cave, Bos,
Doolittle, & Toussaint, 1993; Schochet, Burghardt, & McConnell, 2008), and only about two-
thirds of participants complete these programs (Jastrab et al. 1996; Millenky, Bloom, Miller-
Ravett, & Broadus, 2011). In one evaluation, individuals who participated for the shortest
duration reported earning less income at a follow-up assessment, compared to those who
5
participated in the program for the longest duration (Cave et al., 1993). However, most
evaluations of educational and employment programs have given little attention to whether
intervention effects differ based on whether or how long participants are retained.
Furthermore, very little investigation has been conducted to identify which factors put
individuals at risk for non-retention in education and employment programs. Cave et al. (1993)
found that individuals who had been arrested previously spent the least amount of time in an
education and employment program, but they found no other significant predictors of the
duration of participation. It is possible that predictors of non-retention in other contexts could
inform our understanding of retention in programs like LACC. For example, the same factors
that predict high school dropout for adolescents, such as criminal involvement, gang
membership, and substance use (e.g., Hagedorn, 1998; Lynskey, Coffey, Degenhardt, Carlin, &
Patton, 2003; Newcomb et al., 2002; Pyrooz, 2013; Sweeten, 2006), could also put young adults
at risk for not completing education and employment programs. In the drug treatment literature,
belonging to a deviant peer group (Knight, Logan, & Simpson, 2001) and criminal involvement
(Brown, Zuelsdorff, & Gassman, 2009) predict program non-retention. Both drug use and having
deviant peers predict non-retention in "boot camp" prison alternatives, although gang
membership is protective against non-retention (Benda, Toombs, & Peacock, 2006). Fewer years
of education predicts non-retention in substance use treatment programs (Sayre et al., 2002). If
these factors are also related to retention in education and employment programs, staff members
could use this information to identify and engage participants at risk for not completing the
program.

6
Current Study
The current study is an evaluation of LACC, an education and employment program
serving ethnic minority young adults who have not yet graduated from high school. Using
archival and self-report data, we aimed to 1) determine the success rate for high school diploma
earning at LACC, 2) test whether participants showed significant reductions in substance use,
antisocial behavior, and gang membership over the course of participation in LACC, 3) examine
whether program completers showed greater pre-post reductions in substance use, antisocial
behavior, and gang membership than non-completers, and 4) identify pre-treatment risk factors
for program non-retention.
Method
Evaluation Setting
This study was conducted at LACC, a community-based employment and education
program for unemployed young adults without high school diplomas. LACC was founded in
1986 in order to provide work and service opportunities to young adults, and a charter school
was launched in 1997 to give corps members the opportunity to earn high school diplomas. The
current study took place at the South Los Angeles site, which consists of a two-story building
with classrooms, space for job training activities, and counseling rooms. Corps members wear
uniforms and participate in a wide variety of activities throughout the week, with a focus on
earning credits toward a high school diploma and travelling to job sites throughout the city as
members of paid work crews. Additional activities, such as book clubs or running groups, are
also available at the South Los Angeles site to provide corps members with opportunities for
self-improvement outside of work and school. LACC participants enter the program as “recruits”
7
and are promoted to “corps members” after 8 weeks of orientation, during which participants
engage in all non-paid LACC activities. Once recruits achieve corps member status, they are
offered paid work opportunities throughout the community. At the end of 22 weeks, corps
members receive their first performance evaluation and are given a raise if they receive positive
feedback. Eligibility criteria for enrolling in LACC resulted in participants who were all 1)
unemployed, 2) lacking a high school diploma, 3) between 18 and 24 years of age, and 4)
residents of South Los Angeles.
Design
Between July 2012 and May 2013, 160 young adults who enrolled in LACC at the South
Los Angeles site were invited to participate in this program evaluation. A randomized controlled
trial of an adjunctive motivational intervention was conducted with these same participants as
part of a separate study. However, because the motivational intervention had no impact on
retention or other participant outcomes (see Chapter Two), all conditions were combined in this
program evaluation. There were no exclusion criteria for study participation. The study was
approved by the University of Southern California Institutional Review Board. A certificate of
confidentiality was obtained from the US Department of Health and Human Services to protect
participants who reported engaging in criminal behavior or illegal substance use.
Recruitment occurred during the first week of each 8-week orientation period, and the
decision to enroll in the study was kept confidential from LACC staff. Immediately following
informed consent, 100 LACC recruits completed a questionnaire battery assessing demographic
and behavioral factors. Participants were asked to provide up to five telephone numbers (e.g., a
personal number, as well as a father’s, grandmother’s, sister’s, and boyfriend’s number), five
8
mailing addresses, and one email address to increase chances of successfully contacting them for
subsequent assessments. In order to enroll in the study, participants were required to sign release
of information forms to allow researchers access to LACC records. Research assistants
administered two follow-up assessments by phone, mail, or email, based on participant
preference. These follow-up assessments were scheduled in conjunction with two program
milestones: the completion of LACC orientation (8 weeks) and the first performance evaluation
(22 weeks). For compensation, participants received a $5 gift card at study enrollment, a $10 gift
card at the 8-week follow-up, and a $15 gift card at the 22-week follow-up. High school diploma
and retention status were extracted from LACC’s electronic archives at the conclusion of the
study.
Measures
Demographic information and personal characteristics. Participant age, ethnicity,
gender, and years of education were collected through self-report. Additionally, self-reported
child welfare system history, arrest history, relationship status, and parenthood status were
assessed.
High school diploma earning. LACC recruits spend a great deal of their time in class,
earning credits toward a high school diploma and studying for the California High School Exit
Exam (CAHSEE). After the 22-week follow-up, data was extracted from LACC records to
categorize participants as either having earned or not earned a high school diploma.
Substance use. Substance use was measured using items from the Youth Risk Behavior
Survey (YRBS; CDC, 2011), a questionnaire with demonstrated test-retest reliability, which is
administered twice annually to a nationally representative sample of high school students (Brener
9
et al., 2002; Zullig, Pun, Patton, & Ubbes, 2006). The YRBS has also been administered to
young adults (e.g., Schwartz et al., 2011). Items include, “During the past 30 days, how many
times did you use marijuana (also called grass or pot)?” and, “During the past 30 days, how
many times have you sniffed glue, breathed the contents of aerosol spray cans, or inhaled any
paints or sprays to get high?” The YRBS measures rates of binge drinking, tobacco use,
marijuana use, and other illegal drug use (e.g., heroin, cocaine) at each study time point. Most
analyses done using the YRBS rely on dichotomous measures of each type of substance use (e.g.,
Melnyk et al., 2013; Mercado-Crespo & Mbah, 2013), and this was the approach taken in the
current study as well.
Antisocial behavior. Self-reported antisocial behavior was measured at each assessment
period using the Self-Report Delinquency Scale (SRDS; Elliott, Huizinga, & Morse, 1986). This
measure includes 40 items, which mostly cover illegal activities but also include some legal
behaviors such as infidelity or begging. The SRDS was designed for use with juvenile
populations but has also been applied to adult samples (e.g., Langhinrichsen-Rohling, Arata,
Bowers, O'Brien, & Morgan, 2004). Sample items include, “In the past month, how many times
have you stolen (or tried to steal) something worth more than $50?” and, “In the past month, how
many times have you carried or hidden a weapon other than a plain pocket knife?” Self-report of
antisocial behavior is common (Aebi, 2010; Enzmann et al., 2010), and questionnaires like the
SRDS are significantly correlated with official reports of criminal activity (Hindelang, Hirschi,
& Weis, 1981; Kirk, 2006; Maxfield, Weiler, & Widom, 2000). The SRDS also has acceptable
test-retest reliability (Thornberry & Krohn, 2000). We used the variety scoring method, which
sums the number of different antisocial acts an individual reports, as this has been shown to be
10
more reliable and less skewed than other scoring methods (Bendixen, Endresen, & Olweus,
2003).
Gang membership. Gang membership was assessed using the Eurogang Youth Survey
(Weerman et al., 2009). This measure allows participants to endorse items derived from the
following definition: "a street gang... is any durable, street-oriented youth group whose
involvement in illegal activity is part of its group identity” (Weerman et al., 2009, p. 20). These
definitional elements are operationalized as belonging to a group of friends who, 1) are mostly
between 12 and 25 years old, 2) have been friends with each other for at least three months, 3)
spend time in public spaces, 4) commit illegal acts, and 5) consider illegal acts to be acceptable.
Self-labeling oneself as a “gang member” is not required.
Program retention. Participants may stop attending LACC for various reasons,
including leaving the program volitionally, breaking a “zero tolerance” rule, or being forcibly
terminated from the program (e.g., testing positive for drugs, fighting at the work site, receiving
three “docks” for bad behavior), or being unable to participate due to other circumstances (e.g.,
incarceration, illness, psychiatric hospitalization,). When a participant is terminated from or
leaves LACC, staff members record these incidents as “separations” in their electronic archives
and categorize them as negative, neutral, or positive. Negative reasons include gross misconduct
(e.g., testing positive for drugs or fighting at the work site) or having excessive absences. Neutral
reasons include medical circumstances (e.g., physical illness, mental illness, pregnancy) or
relocation (e.g., moving to a different state, being incarcerated). Positive reasons include
obtaining employment or pursuing higher education. For this evaluation, program non-retention
for negative, neutral, or positive reasons was extracted from the electronic records at 8 weeks
and 22 weeks.
11
Analysis
Descriptive univariate statistics were calculated to characterize the sample in terms of
age, ethnicity, gender, and other characteristics. The proportion of the sample that succeeded in
earning a high school diploma was calculated at 22 weeks. Then, χ
2
tests, independent-samples t-
tests, and Mann–Whitney U tests were used to test for participant factors associated with high
school diploma earning, based on whether the predictor variables were categorical, continuous,
and normal or non-normal. We tested for significant pre-post changes in self-report outcomes
using McNemar tests for categorical data, paired-samples t-tests for normal continuous data, and
Wilcoxon signed rank tests for non-normal continuous data. We examined pre-post changes in
the intent-to-treat sample, the sample of participants who completed 22 weeks in the program,
and the sample of those did not complete the full 22 weeks. We also tested whether pre-post
changes were significantly different by retention status using two-group McNemar tests for
categorical outcomes (Feuer & Kessler, 1989), mixed-model ANOVA tests for normal
continuous outcomes, and mixed-model ANOVA tests on rank-transformed data for non-normal
continuous outcomes (Conover & Iman, 1981). Finally, χ
2
tests, independent-samples t-tests, and
Mann–Whitney U tests were used to test which participant characteristics were related to
reduced likelihood of retention. We also assessed patterns in missing data and re-ran all analyses
involving self-report follow-up data on five separate multiply imputed datasets (m = 5). Graham
and Schafer (1999) have demonstrated that multiple imputation performs well even in samples as
small as N = 50. When results from multiply imputed datasets differed from those found in the
complete dataset, these deviations were reported.

12
Results
Missing Data
There was no missing data on high school diploma earning or program retention.
However, self-report data was missing for a substantial proportion of participants at both 8
weeks (51%) and 22 weeks (32%). Males were less likely to provide self-report data at 8 weeks,
χ
2
(1, 100) = 6.83, p < .01. Not completing the LACC program, χ
2
(1, 100) = 5.18, p = .02, being
married or living with a domestic partner, χ
2
(1, 82) = 4.04, p < .05, previous child welfare
system involvement, χ
2
(1, 97) = 4.17, p = .04, and pre-treatment binge drinking, χ
2
(1, 96) =
4.94, p = .03, predicted a higher likelihood of missing self-report data at 22 weeks.
Reliability and Distributional Properties
The antisocial behavior measure (SRDS; Elliot et al., 1986) showed high internal
consistency (α = .93), but had substantial skewness (2.55) and kurtosis (7.10). The D'Agostino-
Pearson test indicated that antisocial behavior was significantly skewed in this sample (p < .01)
and the Anscombe-Glynn test indicated a non-normal distribution (p < .01) (Wessa, 2014). Years
of education also had substantial skewness (-1.38, p < .01), kurtosis (1.46), and a non-normal
distribution (p < .01) (Wessa, 2014). Therefore, non-parametric tests were used for analyses of
antisocial behavior and years of education.
Baseline Characteristics
Table 1 provides information on participant demographics and baseline characteristics.
Upon entry into LACC, participants were about 20 years old and had attended school for 10
years, on average. Six participants reported completing more than 12 years of education, but had
13
not yet obtained high school diplomas. Sixty percent of participants were male, about two-thirds
were Latino, and one-third was African American. A substantial proportion of participants had
been arrested previously, and reported using alcohol, tobacco, and marijuana in the month before
enrolling in LACC.
Table 1: Baseline Characteristics
Characteristic Total (N = 100)
n %
Gender (male) 60 60.0
Ethnicity -- --
Latino 58 61.7
African American 29 30.9
Other 5 5.3
Married/Living with Partner 15 18.3
Parent 27 27.6
Child Welfare System
History
17 17.5
Arrest History 39 42.9
Gang Member 15 17.6
Binge Drinking 28 29.2
Tobacco Use 28 28.9
Marijuana Use 29 29.6
Other Illegal Drug Use 8 8.2
M SD
Age, years 19.91 1.56
Education, years 10.07 3.94
Antisocial Behavior 2.96 5.26
Note. Percentages were always calculated from the number of participants endorsing a behavior
divided by the number of participants who gave reports on that behavior, not the total number of
participants.

Educational and Problem Behavior Outcomes for LACC Participants
We evaluated the possible impact of LACC on educational attainment and reductions in
problem behaviors. Twenty-eight percent of participants earned a high school diploma through
LACC. Participants who reported gang-involvement prior to enrolling in LACC were more
14
likely to earn diplomas than participants who were not gang members, χ
2
(1, 85) = 4.44, p = .04
(Table 2).
Table 2: High School Diploma Earning at LACC
Predictors Diploma Earned
n = 28
Not Earned
n = 72
n % n %
Gender (male) 14 50.0 46 63.9
Ethnicity
Latino 18 66.7 40 59.7
Black 8 29.6 21 31.3
Married/Living with Partner 4 16.0 11 19.3
Parent 9 32.1 18 25.7
Child Welfare System History 4 14.3 13 18.8
Arrest History 9 34.6 30 46.2
Gang Member* 8 30.8 7 11.9
Binge Drinking 9 32.1 19 27.9
Tobacco Use 7 25.0 21 30.4
Marijuana Use 9 32.1 20 28.6
M SD M SD
Age, years 20.00 1.63 19.87 1.54
Education, years 10.46 2.90 9.92 4.28
Antisocial Behavior 3.32 6.86 2.81 4.50
* p < .05

Wilcoxon signed rank tests and exact McNemar tests performed on the intent-to-treat
sample indicated that participants significantly reduced their antisocial behavior over 8 weeks, Z
= 2.34, p = .02, and their gang membership over 22 weeks (p = .01). Re-analyses using multiple
imputation no longer indicated significant reductions (p < .05) in either outcome, although the
reduction in gang membership over 22 weeks was marginally significant (p = .06).
Seventy-four percent of recruits completed the program orientation, and 61% completed
22 weeks in the program. Before orientation was complete, one participant left for a job, which
was considered a positive reason by LACC. Another 25 participants were not retained in the
program for negative reasons. Between orientation and the first performance evaluation, two
15
recruits left for reasons considered positive by LACC—one for a job and another to enroll in a
different school. Three recruits were not retained for neutral reasons, and 8 participants were not
retained for negative reasons. Those participants who completed 22 weeks in the program
showed significant reductions in antisocial behavior over 8 weeks, Z = 2.02, p = .04, and gang
membership over 22 weeks (p = .04). Non-completers failed to show these reductions (Table 3).
However, two-group McNemar tests and mixed-model ANOVA tests using rank-transformed
data did not indicate significant differences in pre-post changes by retention status. Results from
the re-analyses using multiple imputation indicated that the reduction in antisocial behavior was
no longer significant for program completers, but that the reduction in gang membership was
significant.
Table 3: Pre-Post Tests for Self-Report Outcomes
Program Completers (n = 61) Program Non-Completers
a
(n = 39)
Pre Post Pre Post
8-Week Follow-Up n % n % n % n %
Binge Drinking 6 17.6 4 11.8 6 40.0 5 33.3
Tobacco Use 8 23.5 5 14.7 5 35.7 5 35.7
Marijuana Use 3 9.1 1 3.0 7 46.7 4 26.7
Gang Membership 5 15.2 2 6.1 1 8.3 2 16.7
M SD M SD M SD M SD
Antisocial Behavior 1.35 4.27 .26 .75 5.13 5.22 3.33 4.10
22-Week Follow-Up n % n % n % n %
Binge Drinking 9 20.5 9 20.5 5 26.3 2 10.5
Tobacco Use 12 26.1 9 19.6 6 30.0 9 45.0
Marijuana Use 9 19.6 5 10.9 7 35.0 9 45.0
Gang Membership 9 20.9 2 4.7 2 11.8 0 0.0
M SD M SD M SD M SD
Antisocial Behavior 1.47 4.27 .80 1.82 4.10 4.12 2.55 3.83
Note. Bolded cells indicate within-group significant differences (p < .05). There were no
significant differences in pre-post changes between completers and non-completers.
a
Participants who were not retained in the program for any reason (i.e., positive, negative, or
neutral) were included in this group to examine changes made only by participants who
completed all 22 weeks versus those who did not.

16
Risk Factors for Program Non-Retention
Risk factors for program non-retention (excluding non-retention for positive reasons)
were examined. Non-retention for a positive reason was not considered problematic in these
analyses, because job attainment and school enrollment were consistent with the goals of LACC.
Table 4 provides information about which pre-treatment characteristics predicted non-retention.
More years of education at treatment entry predicted non-retention at LACC at 8 weeks, U(88) =
541.5, Z = 2.19, p = .03, and at 22 weeks, U(88) = 695.0, Z = 1.97, p < .05. More antisocial
behavior before entering the program also predicted non-retention at LACC at 8 weeks, U(97) =
636, Z = 2.16, p = .03, but not at 22 weeks. Not only was gang membership not a significant risk
factor for non-retention, participants who reported being gang members before entering LACC
were significantly more likely to complete 22 weeks in the program, χ
2
(1, 85) = 4.54, p = .03.
When non-retention for negative reasons only was examined, initial years of education did not
predict retention status at 22 weeks. No other risk factors for negative or neutral non-retention
differed from risk factors for negative non-retention only.
Discussion
This evaluation highlighted the promising opportunities that LACC may offer to
vulnerable young adults, but also uncovered a challenge in program retention. Almost the same
number of participants to earn a high school diploma, did not even complete program orientation.
The prospect of earning a high school diploma is an important societal good; graduating high
school at LACC could confer benefits for earning potential and psychosocial adjustment to
vulnerable young adults. Furthermore, there was evidence that LACC participants, particularly
those who completed 22 weeks in the program, successfully reduced their own antisocial
17
Table 4: Risk Factors for Program Non-Retention for Negative or Neutral Reasons
8 Weeks 22 Weeks
Baseline Characteristics Retained
a
n = 75
Not Retained
n = 25
Retained
a
n = 64
Not Retained
n = 36
n % n % n % n %
Gender (male) 43 57.3 17 77.3 35 56.3 25 69.4
Ethnicity
Latino 42 59.2 16 69.6 38 61.3 20 62.5
Black 22 31.0 7 30.4 18 29.0 11 34.4
Married/Living with Partner 9 14.8 6 28.6 7 13.7 8 35.8
Parent 20 27.4 7 28.0 15 24.2 12 33.3
Child Welfare System History 12 16.4 5 20.8 8 12.9 9 25.7
Arrest History 30 44.8 9 37.5 24 41.4 15 45.5
Gang Member 14 21.5 1 5.0 14 23.3 1 4.0
Binge Drinking 20 27.4 8 34.8 17 27.4 11 32.4
Tobacco Use 20 27.4 8 33.3 16 25.4 12 35.3
Marijuana Use 19 25.7 10 41.7 17 27.0 12 34.3
M SD M SD M SD M SD
Age, years 20.00 1.66 19.64 1.22 20.03 1.69 19.69 1.28
Education, years 9.72 3.79 11.00 4.25 9.78 3.69 10.53 4.32
Antisocial Behavior 2.49 5.08 4.38 5.64 2.53 5.33 3.71 5.11
Note. Bolded cells indicate significant differences (p < .05).
a
Participants who left the program for a job or school were included in the “retained” group for
these analyses in order to identify factors which predicted non-retention for negative or neutral
reasons only, as these reasons were conceptually more problematic than not completing the
program after obtaining a job or transferring to a new school.

behavior and gang membership over the course of participation in the program. However, LACC
program retention rates were moderately low. About a quarter of participants did not complete
program orientation, and over a third did not complete 22 weeks in LACC. These retention rates
are similar to those reported in previous evaluations (Cave et al., 1993; Jastrab et al. 1996;
Millenky et al., 2011; Schochet et al., 2008), and may indicate that education and employment
programs, broadly, face a challenge in retaining participants. This study found that participants
with higher antisocial behavior, more years of education, and who were non-gang members at
program entry, were at increased risk for non-retention at LACC. Targeting these participants
18
with retention promotion strategies could improve the effectiveness of education and
employment programs.
The results regarding antisocial behavior warrant attention. It is possible that LACC staff
will feel cautious about enrolling recruits who are criminally active because greater antisocial
behavior put participants at risk for non-retention. This finding is consistent with Cave and
colleagues’ (1993) result showing that a previous arrest predicted shorter program participation.
LACC has finite resources and it is prudent to consider whether potential recruits will benefit
from being given a program slot. If a participant reports heavy current criminal involvement, this
recruit may not complete the program, leaving a valuable classroom and work crew space empty.
However, excluding criminally active recruits from LACC could be harmful. The significant pre-
post reductions in antisocial behavior for at least some participants suggest LACC could be an
effective crime intervention. Jastrab et al. (1996) found that Conservation Corps programs
reduced the risk for arrest for corps members, and LACC may be producing similar results. For
that reason, enrolling recruits who report antisocial behavior may provide an important pathway
to desistance from crime and have positive effects on the surrounding South Los Angeles
community.
The high levels of diploma earning and program retention for gang members at LACC
were unexpected. Gang membership can facilitate negative behaviors, such as drug use and
antisocial behavior (Bendixen et al., 2006; Gatti et al., 2005; Harper et al, 2008), which could
have been barriers to success at LACC. However, it is possible that LACC provided pro-social
peers and positive adult role models to gang members that they lacked. Vigil (2004) has argued
that schools can engage gang members better if they provide some of the positive qualities of
gangs, such as respect, friendships, security, and affection. Programs like LACC, which are
19
delivered in a group context, may provide new pro-social settings to gang members and offer a
pathway toward educational and occupational success. These aspects of LACC may also explain
why gang membership significantly declined according to pre-post analyses. Furthermore, the
structure of LACC may be especially effective with gang members. Benda and colleagues (2006)
suggested that “boot camp” programs may be appealing to gang members because of the
hypermasculine, group-oriented environment. LACC participants wear uniforms, meet for
morning calisthenics, and engage in difficult physical labor. These elements could be a good fit
for gang-affiliated young adults.
It was also surprising that participants who had completed more years of education before
entering LACC were less likely to complete the program. This result contradicts previous
research showing that fewer years of education puts individuals at risk for program non-retention
(Sayre et al., 2002). We considered whether this result emerged because participants with more
years of education were able to earn the last few credits they needed to graduate, and then left the
program without engaging in the work opportunities offered at LACC. However, years of
education did not predict high school diploma earning. It is possible that these LACC
participants became demoralized while working to pass the CAHSEE. English-language learners
and African Americans have a significantly harder time passing the CAHSEE than native
English speakers or Whites (Zau & Betts, 2008). Some students may have already attended five
or six years of traditional high school without graduating (Becker, Wise, & Watters, 2010) and
may have entered LACC already feeling discouraged about earning a diploma and training for a
career. Finally, the course work may have been geared toward corps members with lower levels
of educational attainment, which could have made instruction less appealing to participants with
higher levels of academic achievement. For example, if students at a 9
th
grade level are attending
20
class with students at a 12
th
grade level, the teacher may have to target instruction to the 9
th
grade
level. A mismatch between academic needs and curriculum decisions could have made the
educational opportunities less attractive to corps members with more education.
Limitations
The results of this study should be interpreted with caution because of the small sample
size. Also, bivariate analyses were the most appropriate tools for testing study hypotheses,
because statistical power was limited. Future research should access larger samples so that
integrated models can be constructed to more fully characterize program effects and retention
prediction. Furthermore, the lack of a control group limits the conclusions that can be drawn
from the pre-post analyses. Participants who chose to enroll in LACC may have reduced their
antisocial behavior and gang membership even without the program. However, the fact that we
observed significant reductions in problem behaviors for program completers, but not for those
who failed to complete the program suggests LACC may have provided essential tools for
criminal and gang desistance. Also, the missing self-report data may have introduced bias. Re-
analyses using multiple imputation indicated that the reductions in antisocial behavior may not
have been significant across the entire sample. For example, criminal desistance may not occur
for participants who engage in more binge drinking, since using pre-treatment binge drinking to
impute missing data resulted in analyses that did not confirm the reductions in antisocial
behavior found in the complete dataset. Therefore, the generalizability of these results may be
limited. Finally, educational attainment and reductions in problem behaviors were only assessed
over 22 weeks. After this point, some corps members may have continued to participate in
LACC or made progress outside the program. Long-term effects of enrolling in LACC were not
explored in this study.
21
Conclusion
The results of this evaluation suggest that the iteration of the Conservation Corps in
South Los Angeles may provide an important service to ethnic minority, unemployed young
adults, providing them with an opportunity to earn a high school diploma and offering some a
pathway toward criminal and gang desistance. However, program retention is a challenge. It may
be beneficial to explore adjunctive strategies staff can implement that will keep participants
engaged in the education and employment opportunities provided at LACC.
22
References
Advancement Project. (2010). Test, punish, and pushout: How 'zero-tolerance' and high-stakes
testing funnel youth into the 'school-to-prison pipeline'. Retrieved from http://www.
advancementproject.org/sites/default/files/publications/rev_fin.pdfhttp://www.advancement
project.org/sites/default/files/publications/rev_fin.pdf
Aebi, M. F., Andersson, L., & Carra, C. (2010). Self-reported crime and deviance studies in Europe.
In R. Zauberman (Ed.), Current state of knowledge and review of use (pp. 11-50). Brussels:
VUBPRESS Brussels University Press.
Becker, D. S., Wise, L. L., & Watters, C. (2010). Independent Evaluation of the California High
School Exit Examination (CAHSEE): 2010 Biennial Report. Alexandria, VA: Human
Resources Research Organization.
Benda, B. B., Toombs, N. J., & Peacock, M. (2006). Distinguishing graduates from dropouts and
dismissals: Who fails boot camp?. Journal of Criminal Justice, 34(1), 27-38. doi:10.1016/j.
jcrimjus.2005.11.003
Bendixen, M., Endresen, I. M., & Olweus, D. (2003). Variety and frequency scales of antisocial
involvement: Which one is better?. Legal and Criminological Psychology, 8(2), 135-150.
doi:10.1348/135532503322362924
Bendixen, M., Endresen, I. M., & Olweus, D. (2006). Joining and leaving gangs: Selection and
facilitation effects on self-reported antisocial behaviour in early adolescence. European
Journal of Criminology, 3(1), 85-114. doi:10.1177/1477370806059082
23
Bernburg, J. G. & Krohn, M. D. (2003). Labeling, life chances, and adult crime: The direct and
indirect effects of official intervention in adolescence on crime in early adulthood.
Criminology, 41, 1287-317. doi:10.1111/j.1745-9125.2003.tb01020.x
Bingham, C. R., Shope, J. T., & Tang, X. (2005). Drinking behavior from high school to young
adulthood: Differences by college education. Alcoholism: Clinical & Experimental
Research, 29, 2170-2180. doi:10.1097/01.alc.0000191763.56873.c4
Bloom, D. (2010). Programs and policies to assist high school dropouts in the transition to
adulthood. The Future of Children, 20, 89-108. doi:10.1353/foc.0.0039
Boudett, K.P., Murnane, R.J., & Willett, J.B. (2000). ‘Second chance’ strategies for women who
drop out of school. Monthly Labor Review, 123, 19–32.
Brener, N. D., Kann, L., McManus, T., Kinchen, S. A., Sundberg, E. C., & Ross, J. G. (2002).
Reliability of the 1999 youth risk behavior survey questionnaire. Journal of Adolescent
Health, 31, 336-342. doi:10.1016/S1054-139X(02)00339-7
Brown, R. T., Zuelsdorff, M., & Gassman, M. (2009). Treatment retention among African
Americans in the Dane County Drug Treatment Court. Journal of Offender
Rehabilitation, 48(4), 336-349. doi:10.1080/10509670902851042
Casswell, S., Pledger, M., & Hooper, R. (2003). Socioeconomic status and drinking patterns in
young adults. Addiction, 98, 601-610. doi:10.1046/j.1360-0443.2003. 00331.x
Cave, G., Bos, H. Doolittle, F., & Toussaint, C. (1993) JOBSTART: Final report on a program for
school dropouts. New York: Manpower Demonstration Research Corporation.
24
CDC. (2011). State and local youth risk behavior survey. Retrieved from http://www.cdc.gov/
healthyyouth/yrbs/pdf/questionnaire/2011_hs_questionnaire.pdf.
Cohen, P., Chen, H., Hamigami, F., Gordon, K., & McArdle, J. J. (2000). Multilevel analyses for
predicting sequence effects of financial and employment problems on the probability of
arrest. Journal of Quantitative Criminology, 16, 223-235. doi:10.1023/A:1007568606759
Conover, W. J. & Iman, R. L. (1981). Rank transformations as a bridge between parametric and
nonparametric statistics. The American Statistician, 35(3), 124-129. doi:10.2307/2683975
Crum, M., Helzer, J. E., & Anthony, J. C. (1993). Level of education and alcohol abuse and
dependence in adulthood: A further inquiry. American Journal of Public Health, 83, 830-
837. doi:10.2105/AJPH.83.6.830
Edelman, P. B., Holzer, H. J., & Offner, P. (2006). Reconnecting disadvantaged young men.
Washington, DC: The Urban Institute Press.
Elliott, D.S., Huizinga, D., & Morse, B. (1986). Self-reported violent offending: A descriptive
analysis of juvenile violent offenders and their offending careers. Journal of Interpersonal
Violence, 1, 472-514. doi:10.1177/088626086001004006
Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2000). Early work histories of urban
youth. American Sociological Review, 65, 279-297. doi:10.2307/2657441
Enzmann, D., Marshall, I. H., Killias, M., Junger-Tas, J., Steketee, M., & Gruszczynska, B. (2010).
Self-reported youth delinquency in Europe and beyond: First results of the second
25
international self-report delinquency study in the context of police and victimization
data. European Journal of Criminology, 7, 159-183. doi:10.1177/ 1477370809358018
Farley, R. (1996). The new American reality. New York: Russell Sage Foundation.
Feuer, E. J., & Kessler, L. G. (1989). Test statistic and sample size for a two-sample McNemar test.
Biometrics, 45(2), 629-636. doi:10.2307/2531505
Fothergill, K. E., Ensminger, M. E., Green, K. M., Crum, R. M., Robertson, J., & Juon, H. (2008).
The impact of early school behavior and educational achievement on adult drug use
disorders: A prospective study. Drug and Alcohol Dependence, 92, 191-199. doi:10.1016/j.
drugalcdep.2007.08.001
Fussell, E., & Furstenberg, F.. (2005). The transition to adulthood during the twentieth century:
Race, nativity, and gender. In F. F. Furstenberg, R. G. Rumbaut, & R. A. Setterson (Eds.),
On the frontier of adulthood: Theory, research, and public policy (pp. 29-75). Chicago: The
University of Chicago Press.
Gatti, U., Tremblay, R. E., Vitaro, F., & McDuff, P. (2005). Youth gangs, delinquency and drug
use: A test of the selection, facilitation, and enhancement hypotheses. Journal of Child
Psychology and Psychiatry, 46(11), 1178-1190. doi:10.1111/j.1469-7610.2005.00423.x
Graham, J. W., & Schafer, J. L. (1999). On the performance of multiple imputation for multivariate
data with small sample size. In R. Hoyle (Ed.), Statistical Strategies for Small Sample
Research (pp. 1–29). Thousand Oaks, CA: Sage.
26
Hagedorn, J. M. (1998). People and folks: Gangs, crime and the underclass in a rustbelt city (2
nd

ed). Chicago: Lakeview Press.
Harford, T. C., Yi, H., & Hilton, M. E. (2006). Alcohol abuse and dependence in college and
noncollege samples: A ten-year prospective follow-up in a national survey. Journal of
Studies on Alcohol, 67, 803-809. doi:10.1177/0146167201277008
Harper, G. W., Davidson, J., & Hosek, S. G. (2008). Influence of gang membership on negative
affect, substance use, and antisocial behavior among homeless African American male
youth. American Journal of Men's Health, 2(3), 229-243. doi:10.1177/1557988307312555
Hendrickson, Jr., K. E. (2003). Replenishing the soil and the soul of Texas: The civilian
conservation corps in the lone star state as an example of state-federal work relief during the
great depression. The Historian, 65, 801-817. doi:10.1111/1540-6563.00038
Henry, K. L., Knight, K. E., & Thornberry, T. P. (2012). School disparticipation as a predictor of
dropout, delinquency, and problem substance use during adolescence and early adulthood.
Journal of Youth and Adolescence, 41, 156-166. doi:10.1007/s10964-011-9665-3
Hindelang, M.J., T. Hirschi, & Weis, J.G. (1981). Measuring delinquency. Beverly Hills: Sage
Publications.
Jastrab, J., Masker, J., Blomquist, J., & Orr, L. (1996). Evaluation of national and community
services programs: Impacts of service: Final report on the evaluation of American
conservation and youth service corps. Cambridge, MA: Abt Associates.
27
Kerckhoff, A. C., & Bell, L. (1997). Early adult outcomes of students at ‘risk.’ Social Psychology of
Education: An International Journal, 2, 81-102. doi:10.1023/A: 1009605618662
Kirk, D. S. (2006). Examining the divergence across self-report and official data sources on
inferences about the adolescent life-course of crime. Journal of Quantitative Criminology,
22, 107-129. doi:10.1007/s10940-006-9004-0
Knight, D. K., Logan, S. M., & Simpson, D. D. (2001). Predictors of program completion for
women in residential substance abuse treatment. The American Journal of Drug and Alcohol
Abuse, 27(1), 1-18. doi:10.1081/ADA-100103116
Landis, J. R. & Koch, G. G. (1977). The measurement of observer agreement for categorical data.
Biometrics, 33(1), 159–174. doi:10.2307/2529310
Langhinrichsen-Rohling, J., Arata, C., Bowers, D., OBrien, N., & Morgan, A. (2004). Suicidal
behavior, negative affect, gender, and self-reported delinquency in college students. Suicide
and Life-Threatening Behavior, 34, 255-266. doi:10.1521/suli.34.3.255.42773
Laub, J. H., & Sampson, R. J. (2003). Shared beginnings, divergent lives: Delinquent boys to age
70. Boston, MA: Harvard University Press.
Lustig, K., & Liem, J. H. (2010). Quality of employment and delinquency during the adolescent to
young adult transition. The New School Psychology Bulletin, 8, 4-14.
Lynskey, M. T., Coffey, C., Degenhardt, L., Carlin, J. B., & Patton, G. (2003). A longitudinal study
of the effects of adolescent cannabis use on high school completion. Addiction, 98(5), 685-
692. doi:10.1046/j.1360-0443.2003.00356.x
28
Maxfield, M. G., Weiler, B. L., & Widom, C. S. (2000). Comparing self-reports and official records
of arrests. Journal of Quantitative Criminology, 16, 87-110. doi:10.1023/A: 1007577512038
Melnyk, B. M., Jacobson, D., Kelly, S., Belyea, M., Shaibi, G., Small, L., . . . Marsiglia, F. F.
(2013). Promoting healthy lifestyles in high school adolescents: A randomized controlled
trial. American Journal of Preventive Medicine, 45(4), 407-415. doi:10.1016/j.amepre.2013.
05.013
Mercado-Crespo, M., & Mbah, A. K. (2013). Race and ethnicity, substance use, and physical
aggression among U.S. high school students. Journal of Interpersonal Violence, 28(7),
1367-1384. doi:10.1177/0886260512468234
Millenky, M., Bloom, D., Muller-Ravett, S., & Broadus, J. Three-year results of the National Guard
Youth ChalleNGe Evalution. New York: Manpower Demonstration Research Corporation.
Morris, M., & Western, B. (1999). Inequality in earnings at the close of the twentieth century.
Annual Review of Sociology, 25, 623– 657. doi:10.1146/annurev.soc.25.1. 623
Newcomb, M. D., Abbott, R. D., Catalano, R. F., Hawkins, J. D., Battin-Pearson, S., & Hill, K.
(2002). Mediational and deviance theories of late high school failure: Process roles of
structural strains, academic competence, and general versus specific problem
behavior. Journal of Counseling Psychology, 49(2), 172-186. doi:10.1037/0022-0167.49.
2.172
Obot, I. S., Hubbard, S., & Anthony, J. C. (1999). Level of education and injecting drug use among
African Americans. Drug and Alcohol Dependence, 55, 177-182. doi:10. 1016/S0376-
8716(98)00168-9
29
Pettit, B., & Western, B. (2004). Mass imprisonment and the life course: Race and class inequality
in U.S. incarceration. American Sociological Review, 69, 151-169. doi:10.1177/
000312240406900201
Pyrooz, D. C. (2013). The non-criminal consequences of gang membership: Impacts on education
and employment in the life-course. Retrieved from http://search.
proquest.com/docview/1346798791?accountid=14749. (1346798791; 2013-99030-353).
Ronka, A., Oravala, S., & Pulkkinen, L. (2002). ‘I met this wife of mine and things got onto a better
track’: Turning points in risk development. Journal of Adolescence, 25, 47–63. doi:
10.1006/jado.2001.0448
Sayre, S. L., Schmitz, J. M., Stotts, A. L., Averill, P. M., Rhoades, H. M., & Grabowski, J. J. (2002).
Determining predictors of attrition in an outpatient substance abuse program. The American
Journal of Drug and Alcohol Abuse,28(1), 55-72. doi:10. 1081/ADA-120001281
Schochet, P. Z., Burghardt, J., & McConnell, S. (2008). Does job corps work? Impact findings from
the National Job Corps Study. The American Economic Review, 98(5), 1864-1886. doi:
10.1257/aer.98.5.1864
Schwartz, S. J., Weisskirch, R. S., Zamboanga, B. L., Castillo, L. G., Ham, L. S., Huynh, Q., . . .
Cano, M. A. (2011). Dimensions of acculturation: Associations with health risk behaviors
among college students from immigrant families. Journal of Counseling Psychology, 58, 27-
41. doi:10.1037/a0021356
30
Staff, J., & Mortimer, J. T. (2007). Educational and work strategies from adolescence to early
adulthood: Consequences for educational attainment. Social Forces, 85, 1169-1194.
doi:10.1353/sof.2007.0057
Staff, J., Schulenberg, J. E., Maslowsky, J., Bachman, J. G., O’Malley, P. M., Maggs, J. L., &
Johnston, L. D. (2010). Substance use changes and social role transitions: Proximal
developmental effects on ongoing trajectories from late adolescence through early
adulthood. Development and Psychopathology, 22, 917–932. doi:
10.1017/S0954579410000544
Sum, A., Khatiwada, I., McLaughlin, J., & Palma, S. (2011). No country for young men:
Deteriorating labor market prospects for low-skilled men in the United States. Annals of the
American Academy of Political and Social Science, 635, 24-55. doi:10.1177/
0002716210393694
Sweeten, G. (2006). Who will graduate? Disruption of high school education by arrest and court
involvement. Justice Quarterly, 23(4), 462-480. doi:10.1080/ 07418820600985313
The Corps Network. (2014). The corps network: Strengthening America through service and
conservation. Retrieved from http://www.nascc.org/about
Thornberry, T.P., & Krohn, M.D. (2000). The self-report method for measuring delinquency and
crime. In U.S. National Institute of Justice (Ed.), Measurement and analysis of crime and
justice: Criminal justice series (Vol. 4, pp. 33-83). Washington, DC: National Institute of
Justice.
31
US Department of Labor. (2014). Employment status of the civilian population 25 years and over by
educational attainment. Retrieved from http://www.bls.gov/news. release/empsit.t04.htm
Vigil, J. D. (2004). Gangs and group membership: Implications for schooling. In M.A. Gibson, P.
Gándara, and J.P. Koyama (Eds.). School connections: U.S. Mexican youth, peers, and
school achievement (pp. 87-106). New York: Teachers College Press.
Weerman, F. M., Maxson, C. L., Esbensen, F.-A., Aldridge, J., Medina, J., & van Gemert, F. (2009)
Eurogang Program Manual. Retrieved at http://www.umsl.edu/ccj/eurogang/
EurogangManual.pdf
Wessa, P. (2014), Free Statistics Software, Office for Research Development and Education, version
1.1.23-r7, Retrieved from http://www.wessa.net/
Woo, H., & Sakamoto, A. (2010). Racial and ethnic differentials in idleness, highest-risk idleness,
and dropping out of high school. Race and Social Problems, 2, 115-124.
doi:10.1007/s12552-010-9029-8
Zau, A. C., & Betts, J. R. (2008). Predicting Success, Preventing Failure: An Investigation of the
California High School Exit Exam. San Francisco, CA: Public Policy Institute of California.
Zullig, K. J., Pun, S., Patton, J. M., & Ubbes, V. A. (2006). Reliability of the 2005 middle school
youth risk behavior survey. Journal of Adolescent Health, 39, 856-860. doi:
10.1016/j.jadohealth.2006.07.008

32
CHAPTER TWO: A RANDOMIZED CONTROLLED TRIAL OF MOTIVATIONAL
INTERVIEWING TO IMPROVE PROGRAM RETENTION
Abstract
The Los Angeles Conservation Corps (LACC) is a program that provides education and
employment opportunities to young adults. This study tested the efficacy of adjunctive
Motivational Interviewing (MI) for improving program retention at LACC. Participants (N =
100) were randomly assigned to one of three conditions: one session of MI designed to elicit
change talk (language in favor of succeeding at LACC), one placebo counseling session designed
not to elicit change talk, or no additional treatment. Effects on program retention and high school
diploma earning at LACC, as well as on substance use, antisocial behavior, and gang
membership were examined. Furthermore, change talk was examined as a mechanism of change.
Finally, to investigate the role of cognitive dissonance in MI, preference for consistency (an
index of dissonance susceptibility) was tested as a moderator of treatment efficacy. Results
showed that even though MI sessions elicited more change talk than placebo sessions, MI did not
improve retention, facilitate diploma earning, or reduce substance use, antisocial behavior, or
gang membership. There was limited evidence for preference for consistency as a moderator of
MI’s impact on program retention, with a trend toward increased retention for MI participants
with higher preference for consistency. Mostly, these results do not support the use of MI as an
intervention to improve LACC retention, but there is some evidence that MI works for
individuals who are susceptible to dissonance induction. Future researchers should continue to
explore the role of change talk and develop effective strategies for improving client retention in
education and employment programs, which may or may not include MI.

33
Background and Significance
High school dropout puts young adults at risk for unemployment (Sum, Khatiwada,
McLaughlin, & Palma, 2011), criminal involvement (Bernburg & Krohn, 2003), and substance
use (Fothergill et al., 2008). “Second chance” programs, which provide young adults with
education and employment opportunities, may be able to address skill and knowledge gaps,
increase employability, and reduce antisocial behavior and substance use (Bloom, 2010).
Conservation Corps organizations are exemplars of this type of program. In a randomized
controlled trial of young adults across four Conservation Corps sites, corps members were less
likely to be unemployed or arrested compared to non-corps members (Jastrzab, Masker,
Blomquist, & Orr, 1996). However, these programs often suffer from low retention rates (e.g.
Jastrzab et al., 1996). Participants who do not complete education and employment programs
may miss out on opportunities for positive self-development (Uggen, 2000).
Motivational Interviewing (MI) is an intervention that explicitly aims to encourage
targeted behaviors by eliciting and strengthening motivation to change (Miller & Rollnick,
2012). A meta-analysis showed that MI can help people with a wide variety of problems (e.g.,
substance use, medical adherence) and that it has small positive effects on client engagement in
other interventions (Lundahl, Kunz, Brownell, Tollefson, & Burke, 2010). MI has never been
studied as an adjunctive program retention strategy in the context of education and employment
programs, but its successful implementation in a number of other domains suggests it may prove
helpful at retaining participants in LACC.
Change talk (CT), or clients' verbal arguments for change, has been identified as a
potential mechanism underlying MI effects (Miller & Rose, 2009). The link between MI
34
adherence and increased CT is well supported by empirical research (e.g., Gaume, Bertholet,
Faouzi, Gmel, & Daeppen, 2010; Moyers & Martin, 2006), as is the link between increased CT
and positive behavioral outcomes (e.g., Baer et al., 2008; Moyers et al., 2007). Cognitive
Dissonance Theory (Festinger, 1957) offers an explanation of how CT might function as a
change mechanism. Traditional dissonance studies suggest that asking participants to freely
advocate a position makes their attitudes more consistent with their advocacy (e.g., Festinger &
Carlsmith, 1959). CT is essentially advocacy for a certain course of action, and therefore, may
promote future behaviors consistent with that advocacy (Draycott & Dabbs, 1998). That is,
eliciting CT during MI may induce dissonance by highlighting the discrepancy between an
advocated position and inconsistent behaviors (e.g., a statement in favor of arriving on time vs.
actual frequent tardiness). Individuals may then be motivated to reduce dissonance through
behavior change (e.g., arriving on-time in the future). Dissonance induction has promoted
various health and conservation behaviors (Stone & Focella, 2011).
The extant MI literature has not yet conclusively answered the question of whether CT is
a change mechanism. Three within-group studies have shown CT to mediate the relationship
between MI adherence and client changes (Barnett et al., 2013; Moyers, Martin, Houck,
Christopher, & Tonigan, 2009; Pirlott, Kisbu-Sakarya, DeFrancesco, Elliot, & MacKinnon,
2012), although another study did not find evidence of mediation (Vader, Walters, Prabhu,
Houck, & Field, 2010). All four studies were limited by the lack of experimental manipulation of
CT. Two studies have experimentally manipulated CT, but only one of these assessed CT as a
mediator. Glynn and Moyers (2010) showed that when counselors alternated between MI and
Functional Analysis in a single session, more CT occurred during the MI portions of the session.
However, because all participants were exposed to both MI and Functional Analysis, a mediation
35
test involving outcomes was not possible. Morgenstern et al. (2012) randomly assigned
participants to 8 weeks of MI, counseling using MI relational skills only, or a self-change
condition. They found that MI produced the most CT and that CT mediated the effects of MI on
drinking outcomes, but only during the first week of treatment. Since only one study has
simultaneously brought CT under experimental control and tested whether it mediated MI’s
effects on client behaviors, more studies of this design are needed to understand whether CT is
an important mechanism of change in MI.
Furthermore, no studies have investigated whether dissonance is induced in MI.
Examining whether MI promotes change by inducing dissonance is difficult because the very
measurement of dissonance through self-report appears to reduce the magnitude of subsequent
change (Galinsky, Stone, & Cooper, 2000). An alternative to examining MI-related dissonance
through self-report is to measure susceptibility to dissonance induction prior to delivering MI.
Preference for consistency (PFC), a personality trait defined as motivation to be and appear
consistent, has been shown to moderate the magnitude of dissonance effects (Cialdini, Trost, &
Newson, 1995). For example, in a study of adults with high levels of prejudice, individuals with
greater PFC reported less prejudice after advocating a non-prejudicial attitude, compared to low-
PFC individuals who performed the same advocacy (Heitland & Bohner, 2010). Therefore, if
high-PFC participants show higher program retention rates after receiving MI than low-PFC
participants, the dissonance interpretation of MI would be supported.
Current Study
The chief purpose of this study was to bring methodological rigor to research on CT as a
mechanism of MI effects. Therefore, we designed a study that would contrast an MI exercise
36
specifically designed to elicit CT (Values Card Sort; Miller, C’de Baca, Matthews, & Wilbourne,
2001) with a structurally similar control condition that was not designed to elicit CT. In addition,
we included a no-treatment control group to account for the effects of simply meeting with a
counselor. The resulting randomized controlled trial tested the efficacy of this CT-focused MI
exercise for improving program retention and high school diploma earning. MI’s effects on
secondary outcomes, including antisocial behavior, gang membership, and substance use, were
also examined. This study followed in Morgenstern and colleagues’ footsteps by employing a
randomized design and measuring client motivational language in two separate conditions,
allowing for a rare experimental test of CT as a mechanism of change in MI. Furthermore, this
study examined whether it is plausible that MI promotes behavior change by inducing
dissonance, one proposed explanation of CT’s role in MI.
Method
Participants and Setting
This study was conducted at the South Los Angeles site of the Los Angeles Conservation
Corps (LACC). Between July 2012 and May 2013, 160 individuals who enrolled in LACC at the
South Los Angeles site were invited to participate in this randomized trial. There were no
exclusion criteria. However, eligibility criteria for enrolling in LACC resulted in participants
who were all 1) unemployed, 2) lacking a high school diploma, 3) between 18 and 24 years old,
and 4) residents of South Los Angeles. The 100 recruits who participated in this study were
randomly assigned to one of three conditions: 1) MI (n = 34), 2) placebo counseling (PL; n =
38), or 3) no treatment (NT; n = 28) (see Figure 1). Ninety-four percent of MI participants and
89% of PL participants attended their assigned counseling session.
37

Figure 1. Consolidated Standards of Reporting Trials (CONSORT) flow diagram. Participants who were lost to follow-up for self-report data did
not respond to repeated phone, mail, or email contacts from study personnel. No participants who were successfully contacted declined to provide
self-report data.
Declined to participate (n = 60)
Allocated to MI (n = 34)
Received allocated intervention (n = 32)
Did not receive allocated intervention
(failure to schedule session; n = 2)
Lost to follow-up for secondary outcomes
Time 1 (n = 15)
Time 2 (n = 8)
Analyzed for primary outcomes (n = 28)
Randomized (n = 100)
Allocated to PL (n = 38)
Received allocated intervention (n = 34)
Did not receive allocated intervention
(failure to schedule session; n = 4)
Allocated to NT (n = 28)
Received allocated intervention (n = 28)
Did not receive allocated intervention
(n = 0)
Recruited (n = 160)
Lost to follow-up for secondary outcomes
Time 1 (n = 15)
Time 2 (n = 12)
Analyzed for primary outcomes (n = 34)
Lost to follow-up for secondary outcomes
Time 1 (n = 21)
Time 2 (n = 12)
Analyzed for primary outcomes (n = 38)

37
38

Design
The University of Southern California Institutional Review Board approved all study
procedures. A certificate of confidentiality was obtained from the US Department of Health and
Human Services to protect participants who reported illegal activity. Recruitment and consent
took place in a classroom at the beginning of six consecutive LACC orientation periods. In order
to enroll in the study, each participant was required to sign a release of information form that
permitted researchers to access their LACC records. Recruits who agreed to participate
completed questionnaires assessing demographic and other factors. After completing the pre-
treatment questionnaires, participants were randomly assigned to one of the three conditions
using a random number generator. Those participants who were assigned to MI or PL met with a
study interventionist within the first two weeks of LACC. Sessions were audio-recorded.
Research assistants administered two follow-up assessments by phone, mail, or email, according
to participant preference. These follow-up assessments were linked with two program
milestones: the completion of orientation (8 weeks) and the first performance evaluation (22
weeks). Compensation was $5 at pretreatment, $10 at 8 weeks, and $15 at 22 weeks.
Conditions
MI condition. The MI condition consisted of one 30-minute session during which the
Values Card Sort (Miller et al., 2001) was administered using the MI counseling style (e.g.,
evocative, autonomy-enhancing, directed). The Values Card Sort is an exercise designed to (1)
facilitate a discussion of values in relation to target behaviors, (2) elicit CT, and ultimately (3)
enhance perceived discrepancy between stated values and inconsistent behavior (Moyers &
Martino, 2006). Participants were given cards with written values (e.g., independence,
responsibility) and then asked to sort the cards into three stacks labeled “very important to me,”
39
“important to me,” or “not important to me.” Afterward, participants were asked to rank the
cards in the “very important” stack from most-to-least important. Once participants had a final
ranking of their very important values, the interventionists asked open-ended questions exploring
what the top three values meant to the participants, how well the participants saw themselves as
living out these values, how these values were relevant to LACC, and how committed they were
to each value. The interventionists used MI-adherent techniques (e.g., complex reflections, open-
ended questions) to elicit CT and focused the session on how participants’ values were relevant
to succeeding at LACC. The MI protocol is included in Appendix A.
Placebo condition. The PL condition was structurally similar to MI, but differed in
several key ways. Again, interventionists asked the participants to sort the cards into three
categories. However, interventionists then asked the participants to order the values in the “not
important to me” category and explore how these least important values might be important to
other people. In the PL condition, the interventionists used MI-style, but did not direct the
session toward CT. Instead, they directed the session toward elaborations on why other people
might find the bottom three values important, with a target of enhancing perspective-taking. The
PL protocol is included in Appendix B.
No treatment condition. In the NT condition, participants engaged in LACC as usual,
except for responding to questionnaires at pre-treatment, 8 weeks, and 22 weeks.
Interventionist Training and Fidelity
Interventionists were six advanced undergraduate students, who received 16 hours of
instruction from a member of the Motivational Interviewing Network of Trainers (MINT) and
four hours of instruction about delivering the PL condition. PL was presented as a viable
40
alternative to MI, in order to keep interventionists from expecting MI to outperform PL. All six
interventionists conducted both MI and PL sessions throughout the trial. Training for both
interventions involved reviewing treatment protocols, listening to recorded model sessions, and
receiving live supervision of role-played sessions. Interventionists’ practice sessions were coded
using the Motivational Interviewing Treatment Integrity (MITI) Code 3.1.1 (Moyers, Martin,
Manuel, Miller, & Ernst, 2010). The MITI produces behavioral counts and global scores of
therapist behavior. Behavioral counts measure the proportions of open-ended questions, complex
reflections, and MI-adherent behaviors, as well as the ratio of reflections to questions. A global
spirit rating was calculated as a composite score consisting of three global scores: evocation
(eliciting the client's own ideas), collaboration (maintaining equitable power between therapists
and clients), and autonomy/support (emphasizing the client's independence). Interventionists
were trained until they achieved beginning proficiency levels on these measures in both
conditions, so that differences between MI and PL could be understood based on differences in
CT, and not differences in proportions of reflective questions, for example.
Throughout the study, coded transcripts were returned to interventionists as feedback
because previous research has shown this to improve fidelity (Miller, Yahne, Moyers, Martinez,
& Pirritano, 2004). Interventionists attended weekly supervision with the author, a clinical
psychology doctoral student, and participated in a four-hour re-training session with the MINT
trainer mid-way through the study. In studies using the Values Card Sort, high school graduates
taught by MINT trainers have shown proficiency in delivering MI (Naar-King, Outlaw, Green-
Jones, Wright, & Parsons, 2009). In a meta-analysis, the degree held by the counselor did not
moderate treatment efficacy; undergraduates, and those with BAs, MAs, PhDs, or MDs, were
equally effective at delivering MI (Lundahl et al., 2010).
41
Measures
Participant age, ethnicity, gender, and education level were collected through self-report
upon enrollment. Additionally, child welfare system history, arrest history, relationship status,
and parenthood status were assessed. We also measured self-reported program motivation with
Miller and Johnson’s (2008) Motivational Screening Measure, a short three-item measure using
simple language. This measure requires participants to rate how important a target is for them,
how able they feel they are to reach this target, and how committed they are to reaching this
target, on a scale from 0-10. We defined the target as “succeeding at the Conservation Corps.”
Primary outcomes. Two primary program outcomes were extracted from LACC’s
electronic archives. Program retention was measured as a dichotomous variable—participants
were categorized as either “still enrolled in LACC” or “no longer enrolled in LACC” through
orientation (8 weeks) and through the first performance evaluation (22 weeks). Whenever a
participant is expelled from LACC or leaves volitionally, a program staff member records a
“separation” in the electronic archives. Each separation is categorized as negative (e.g., gross
misconduct, attendance issues), neutral (e.g, medical reason, relocation), or positive (e.g.,
employment, enrolling in another school). In this study, we examined the effect of MI on
program non-retention for any negative or neutral reason, as positive separations were consistent
with the goals of LACC. Another program outcome was high school diploma earning. LACC
recruits spend a great deal of their time in class earning credits toward a high school diploma.
After 22 weeks, data was extracted from the LACC records to categorize participants as either
having earned or not earned a high school diploma.
42
Secondary outcomes. Participants provided self-report data on substance use, antisocial
behavior, and gang membership at each time point. Substance use was measured using items
from the Youth Risk Behavior Survey (YRBS; CDC, 2011), a questionnaire with demonstrated
reliability (Zullig, Pun, Patton, & Ubbes, 2006). The YRBS provided a way to measure rates of
binge drinking, tobacco use, marijuana use, and other illegal drug use (e.g., heroin, cocaine) in
the sample. Antisocial behavior was measured using the Self-Report Delinquency Scale (SRDS;
Elliott, Huizinga, & Morse, 1986), a questionnaire with acceptable test-retest reliability
(Thornberry & Krohn, 2000). We used the variety scoring method, which sums the number of
different antisocial acts an individual reports, as this has been shown to be more reliable and less
skewed than other scoring methods (Bendixen, Endresen, & Olweus, 2003). Gang membership
was measured using the Eurogang Youth Survey (Weerman et al., 2009), which asks participants
to endorse definitional components of gang membership without requiring self-labeling as a
“gang member.” Items assess whether the individual belongs to a group of friends who 1) are
mostly between 12 and 25 years old, 2) have been friends with each other for at least three
months, 3) spend time in public spaces, 4) commit illegal acts, and 5) consider illegal acts to be
acceptable.
Change talk. CT was coded from session recordings by three undergraduate research
assistants. They used the Client Behavior Counts section of the Motivational Interviewing Skill
Code 2.1 (MISC; Miller, Moyers, Ernst, & Amrhein, 2008). Coders were trained in bi-weekly
four-hour group meetings until they achieved reliability with master-coded transcripts. Client
CT related to “succeeding at LACC” was identified and summed to create a CT frequency (CTF)
score, the most common measure of CT in MI process research (e.g., Moyers et al., 2009; Gaume
43
et al., 2010). In prior studies, interrater reliability for this variable has ranged from ICC = .41
(Glynn & Moyers, 2010) to ICC = .88 (Moyers et al., 2009).
Preference for consistency. We measured the extent to which participants were
susceptible to dissonance induction with the 9-item Preference for Consistency Scale-Brief
(PFC-B; Cialdini et al., 1995). High scores on this scale predict stronger responses to dissonance
induction (Bator & Cialdini, 2006; Nail et al., 2001). The reliability (α = .84) and validity of the
PFC-B have been established with undergraduates (e.g., Cialdini et al., 1995; Nail et al., 2001),
and the PFC-B has been administered to young adults without high school diplomas (Brown,
Asher, & Cialdini, 2004). One sample item is, “I get uncomfortable when I find my behavior
contradicts my beliefs.”
Data Analysis
Descriptive statistics were calculated to characterize the sample in terms of age, ethnicity,
gender, and other characteristics. We tested whether continuous variables were significantly
skewed and non-normal using D'Agostino-Pearson and Anscombe-Glynn tests (Wessa, 2014).
Baseline characteristics were compared across all three conditions for statistically significant
differences using χ
2
tests for categorical variables, one-way analysis of variance (ANOVA) tests
for continuous variables with normal distributions, and Kruskal–Wallis one-way ANOVA tests
for continuous variables with non-normal distributions. The main effects of MI on primary
outcomes (i.e., retention, high school diploma earning) and secondary outcomes (i.e., antisocial
behavior, substance use, gang membership) were examined using χ
2
tests for categorical
outcomes, t-tests for normal continuous outcomes, and Mann-Whitney U tests for non-normal
continuous outcomes. Fisher’s Exact Test was used when conducting χ
2
tests with fewer than five
44
observations per cell. MI participants were contrasted with PL participants, NT participants, and
a combined group of PL and NT participants. For hypotheses regarding CT, only data from
participants assigned to the MI and PL conditions who received their allocated counseling were
used (n = 66). The effect of MI on CT was assessed with an independent samples t-test, and the
effects of CT on primary and secondary outcomes were assessed using independent samples t-
tests and Spearman’s rank correlation coefficients (r
s)
, a nonparametric measure of correlation.
All analyses were conducted based on intent-to-treat principles. If significant baseline
differences between conditions were identified, we stratified analyses based on the participant
characteristic and examined whether the same results occurred in both groups, because χ
2
tests,
independent samples t-tests, Mann-Whitney U tests, and r
s
do not allow for control variables.

We examined CT as a mediator and PFC as a moderator of MI effects on primary
outcomes, using binary logistic regression models. If there was a significant main effect of MI on
any primary outcome, we added CTF as a second predictor and inspected whether this changed
the significance of the main effect. Regardless of whether there was a main effect of MI on each
primary outcome, we added PFC and the interaction between PFC and MI as additional
predictors. The significance of the interaction term was examined for evidence of moderation.
We included any characteristics with significant baseline differences between conditions in these
logistic regression models, because control variables are possible in this type of analysis.
In order to examine the effect of missing data on results, we explored how participants
who did not provide follow-up self-report data differed from those who did using pretreatment
characteristics. Then, we re-ran all analyses on five separate (m = 5) multiply imputed datasets
(Schlomer, Bauman, & Card, 2010), using secondary outcomes and significant predictors of
45
missing data to impute missing data. We reported any conflicts between results from the analyses
conducted on the complete dataset versus the imputed datasets.
Results
On average, participants were about 20 years old and had completed the 10
th
grade before
dropping out of high school. Sixty percent of the sample was male, about two-thirds were Latino,
and one-third was African American (Table 5). The only participant characteristic that showed
significant baseline differences across the three conditions was gender, χ
2
(2, 100) = 7.01, p =
.03, with the highest proportion of males in the PL group and the lowest proportion in the MI
group. The SRDS showed high internal consistency in this sample (α = .93), but was
significantly skewed (p < .01) and had a non-normal distribution (p < .01). The PFC-B had high
internal consistency (α = .80). Neither the PFC-B nor CTF was significantly skewed or non-
normal. Therefore, non-parametric statistical tests were always used for antisocial behavior, and
conventional analyses were used for the PFC-B and CTF.
Missing Data
There was no missing data for primary outcomes from the LACC archives. CTF was
missing for the 6 participants who did not attend their allocated counseling session. Ten
participants left the PFC-B blank. Self-report data on secondary outcomes was missing for a
substantial proportion of participants at 8 weeks (51%) and 22 weeks (32%). Males were less
likely to provide self-report data at 8 weeks, χ
2
(1, 100) = 6.83, p < .01. Program non-retention at
22 weeks, χ
2
(1, 100) = 5.18, p = .02, being married or living with a domestic partner, χ
2
(1, 82) =
4.04, p < .05, previous child welfare system involvement, χ
2
(1, 97) = 4.17, p = .04, and

46

Table 5: Baseline Characteristics
Total (N = 100) MI (n = 34) PL (n = 38) NT (n = 28)
Characteristic n % n % n % n %
Gender (male)* 60 60.0 14 41.2 27 71.1 18 64.3
Ethnicity
Latino 58 61.7 19 55.9 21 55.3 18 64.3
African American 29 30.9 9 26.5 12 31.6 8 28.6
Other 5 5.3 3 8.8 2 5.2 0 0.0
Married/Living with Partner 15 18.3 5 20.6 5 13.2 3 10.7
Parent 27 27.6 8 23.5 7 18.4 12 42.9
History of Child Welfare System Involvement 17 17.5 4 11.8 7 18.4 6 21.4
History of Arrest 39 42.9 13 38.2 15 39.5 11 39.3
Gang Member 15 17.6 7 25.0 4 12.1 4 16.7
Binge Drinking 28 29.2 8 35.0 11 30.6 9 32.1
Tobacco Use 28 28.9 11 33.3 10 27.8 7 25.0
Marijuana Use 29 29.6 9 27.3 10 27.0 10 35.7
Illegal Drug Use 8 8.2 4 12.1 2 5.4 2 7.1
M SD M SD M SD M SD
Age, years 19.91 1.56 19.97 1.29 19.95 1.77 19.79 1.60
Education, years 10.07 3.94 9.87 3.52 10.32 4.10 9.96 4.38
Antisocial Behavior 2.96 5.26 3.00 4.72 3.03 5.86 2.82 5.22
Program Motivation 9.46 1.18 9.57 .60 9.20 1.84 9.64 .48
Preference for Consistency 5.69 1.49 5.42 1.65 5.80 1.41 5.90 1.39
* p < .05, significant difference by condition.
Note. Percentages were always calculated from the number of participants endorsing a behavior divided by the number of participants
who gave reports on that behavior, not the total number of participants.
46

47

pre-treatment binge drinking, χ
2
(1, 96) = 4.94, p = .03, predicted missing self-report data at 22
weeks.
Intervention Fidelity
Using Landis and Koch’s (1977) criteria, coders achieved fair to good reliability (mean κ
= .63, ranging from .55-.68) when coding individual counselor behaviors. According to
Cicchetti’s (1994) guidelines, ICCs for summary scores were fair to excellent (Table 6). Our goal
was to deliver both MI and PL sessions with proficient use of open-ended questions, reflections,
complex reflections, MI adherence, and MI global spirit in order to equalize the conditions as
much as possible, aside from the amount of CT elicited. Interventionists were judged as
proficient on all five MITI measures when conducting MI sessions, and on four out of five MITI
measures when conducting PL. There were a few significant differences in fidelity between MI
and PL sessions. MI sessions had higher fidelity according to two measures (i.e., reflection to
question ratio, t(64) = 2.93, p = .01; proportion of MI adherence, t(64) = 2.04, p < .05) and PL
sessions had higher fidelity according to one measure (i.e., proportion of complex reflections,
t(64) = 2.64, p = .01), than PL sessions. Due to scheduling conflicts between interventionists and
participants, the author conducted two sessions. Results did not differ when these sessions were
excluded from analyses.
Table 6: MITI Fidelity Ratings
Global
Spirit
Rating
Reflection
To
Question
Ratio
Percent
Open
Questions
Percent
Complex
Reflections
Percent
MI
Adherence
Reliability, ICC .84 .99 .90 .66 .40
Proficiency Standard 3.5 1.00 .50 .40 .90
Average Score, M(SD) 3.52 (.73) 1.39 (.59) .76 (.11) .68 (.15) .94 (.08)
MI Average Score, M(SD) 3.66 (.80) 1.59 (.72) .75 (.13) .64 (.16) .96 (.06)
PL Average Score, M(SD) 3.38 (.65) 1.19 (.34) .77 (.09) .73 (.12) .92 (.09)
Condition Effect * * *
* p < .05
48
Intervention Effects
Twenty-five percent of recruits did not complete the 8-week LACC orientation, and 36%
did not complete 22 weeks in the program. Four of those participants were not retained at 22
weeks for a neutral reason. Twenty-eight percent of recruits earned a high school diploma within
22 weeks at LACC. There were no effects of MI on primary or secondary outcomes compared
with PL, NT, or PL and NT combined (Table 7). MI significantly increased CTF, t(64) = 12.50,
p < .01, compared to PL, although CTF did not predict any outcomes (Table 8). Because illegal
drug use (excluding marijuana) was rare, effects on this outcome were not examined. When the
analyses were stratified by gender in order to investigate the impact of baseline differences in
this variable across conditions, the results were the same. Results did not differ when analyses
were re-run using imputed data.
Table 7: Tests of MI Effects on Change Talk, Primary Outcomes, and Secondary Outcomes
Total (N = 100) MI (n = 34) PL (n = 38) NT (n = 28)
Session Characteristics M SD M SD M SD
Change Talk Frequency 15.88 15.70 29.41 11.79 3.15 3.24 -- --
Session Duration, min 28.65 8.05 27.79 9.15 29.47 6.90 -- --
8 Week Follow-Up n % n % n % n %
Retention 75 75.0 26 76.5 29 76.3 20 71.4
Binge Drinking 9 18.0 3 15.8 2 11.1 4 30.8
Tobacco Use 10 20.4 2 10.5 4 23.5 4 30.8
Marijuana Use 5 10.2 1 5.3 1 5.9 3 23.1
Gang Membership 4 8.0 1 5.3 0 0.0 3 23.1
M SD M SD M SD M SD
Antisocial Behavior 1.18 2.69 1.58 3.72 .72 1.56 1.23 2.17
22 Week Follow-Up n % n % n % n %
Retention 64 64.0 23 67.6 25 65.8 16 57.1
Earned Diploma 28 28.0 10 29.4 11 28.9 7 25.0
Binge Drinking 13 19.4 3 13.6 8 30.8 2 10.5
Tobacco Use 20 29.4 6 27.3 8 30.8 6 30.0
Marijuana Use 15 22.1 4 18.2 5 19.2 6 30.0
Gang Membership 2 2.9 0 0.0 1 3.8 1 5.0
M SD M SD M SD M SD
Antisocial Behavior 1.51 2.84 1.41 3.39 1.46 2.40 1.70 2.85
Note. Bolded cells indicate a significant difference (p < .05).
49

Table 8: Links between Change Talk and Outcomes
Outcomes Change Talk Frequency
Not In Category In Category
n M SD n M SD
8 Week Follow-Up
Retention 16 18.19 18.84 50 15.14 14.70
Binge Drinking 30 16.17 15.54 5 15.40 13.13
Tobacco Use 28 16.79 13.87 6 12.83 18.17
Marijuana Use 32 16.31 15.27 2 15.60 20.51
Gang Membership 34 15.06 14.05 1 50.00 --
22 Week Follow-Up
Retention 23 16.30 16.59 43 15.65 15.40
Earned Diploma 48 15.40 15.59 18 17.17 16.37
Binge Drinking 33 15.85 15.45 10 8.50 9.54
Tobacco Use 30 13.77 15.01 13 15.00 13.93
Marijuana Use 35 13.80 14.91 8 15.63 13.57
Gang Membership 42 14.40 14.61 1 3.00 --
Note. CTF did not significantly predict antisocial behavior at 8 weeks, r
s
= .11, p = .53, or 22
weeks, r
s
= -.06, p = .72.

Mediation and Moderation
Because there were no main effects of MI or CT on primary outcomes, mediation
analyses were not warranted. However, moderation analyses were possible. PFC was a
marginally significant moderator of MI’s effects on retention at 8 weeks (p < .10; Table 9).
Participants who received MI and reported higher PFC were marginally more likely to complete
the 8-week orientation than other participants (Figure 2). However, moderation models explained
very little variance in retention. Results from the imputed datasets indicated the same marginally
significant moderation for MI’s impact on 8-week program retention.

50

Table 9: Moderation Results
Models Predictor β SE β Walds’s χ
2
df OR (95% CI) p
Predicting 8-Week Retention
Step 1 (R
2
= .01) MI vs. Controls -.02 .52 .00 1 .98 (.36-2.69) .97
Step 2 (R
2
= .06) MI vs. Controls
†
-3.81 2.05 3.45 1 .02 (.00-1.23) .06
PFC -.18 .23 .64 1 .83 (.53-1.30) .42
MI x PFC
†
.72 .37 3.76 1 2.05 (.99-4.22) .05
Predicting 22-Week Retention
Step 1 (R
2
= .02) MI vs. Controls .12 .47 .07 1 1.13 (.45-2.84) .79
Step 2 (R
2
= .05) MI vs. Controls -2.18 1.84 1.40 1 .11 (.00-4.15) .24
PFC -0.06 .20 .09 1 .94 (.64-1.39) .76
MI x PFC .43 .33 1.77 1 1.54 (.82-2.92) .18
Predicting Diploma Earning
Step 1 (R
2
= .03) MI vs. Controls .04 .50 .01 1 1.04 (.40-2.76) .93
Step 2 (R
2
= .03) MI vs. Controls -.19 2.01 .01 1 .83 (.02-42.38) .93
PFC .09 .23 .17 1 1.10 (.71-1.71) .68
MI x PFC .05 .34 .02 1 1.05 (.54-2.04) .88
†
p < .10
Note. All analyses were conducted controlling for gender. Ten participants did not fill out the PFC-B measure at pre-treatment, and
were missing from these moderation analyses. Explained variance is estimated with Cox and Snell R
2
.
50

51

Figure 2. Effect of MI on retention moderated by preference for consistency. The 8-week
program retention rate for participants assigned to MI versus control conditions and who
reported preference for consistency scores one standard deviation above versus below average.

Post-hoc Power Analysis
There was sufficient statistical power to detect medium and large effects using χ
2
tests
and large effects using correlational tests. At the most, power was only .52 for finding large
effects with t-tests and .54 with Mann-Whitney U tests. When there were unequal group sizes for
some t-tests, the power was even further reduced. Additionally, there was only sufficient
statistical power to detect medium or large moderation effects in the logistic regression models.
Discussion
This randomized controlled trial was informed by Cognitive Dissonance Theory and
employed a rigorous design to test CT as a mechanism of change. Young adults participating in
an education and employment program were assigned to one of three adjunctive conditions: 1)
one session of MI focused on eliciting CT, 2) a structurally similar counseling session which did
not elicit CT, or 3) no additional treatment. In this trial, MI did not have a significant effect on
0
25
50
75
100
High PFC and
Received MI
Low PFC and
Received MI
High PFC and
Received Control
Low PFC and
Received Control
8-Week Retention Rate
52
program retention or other outcomes, even though it promoted CT. Furthermore, CT did not
predict any participant outcomes. The results of this study did not provide evidence for CT as a
change mechanism, but did show a trend toward MI improving program retention rates for
individuals with higher PFC. This marginally significant moderation offers some support for the
dissonance interpretation of MI’s effects.
Change Talk and Dissonance Induction in Motivational Interviewing
Although there are diverse theoretical explanations for MI’s effects, this trial was
designed to test the plausibility of Cognitive Dissonance Theory as a framework (Draycott &
Dabbs, 1998), and CT as a mechanism of change (Miller & Rose, 2009). The evidence for a
dissonance explanation was mixed. We found that PFC, an index of dissonance susceptibility,
was a marginally significant (p < .10) moderator of MI’s effect on program retention. There was
a trend showing that participants allocated to MI with greater PFC were more likely than other
participants to complete the 8-week program orientation. This provides a rationale for including
PFC measures in future MI research, to clarify these results and further explore the possibility
that dissonance induction is what makes MI effective.
Most previous research on the mediating role of CT in MI has not included experimental
manipulations of CT. Rather, most researchers have compared highly adherent MI sessions with
less adherent MI sessions (Barnett et al., 2013; Moyers et al., 2009; Pirlott et al., 2012). This
precluded researchers’ ability to determine whether MI was effective at eliciting CT, or, rather,
whether clients using more CT were simply more effective at eliciting MI adherence from their
counselors. One promising approach to unraveling causality in this relationship is using
sequential analysis to understand the likelihood of CT following particular counselor behaviors
53
(e.g., Barnett et al., in press; Gaume et al., 2010; Moyers & Martin, 2006). Another is bringing
CT under experimental control. Taking this approach, we demonstrated a causal relationship
between MI and CT, which complements findings by Glynn and Moyers (2009) and
Morgenstern et al. (2012). However, in the current study, CT did not predict greater program
retention, which is inconsistent with Morgenstern’s slight evidence supporting CT as a
mechanism of change. Furthermore, our hypothesis that MI would be more effective for
participants who were susceptible to dissonance induction arose from an interpretation of CT as
advocacy that would lead to dissonance if future behaviors were inconsistent with that advocacy.
The lack of predictive relationships between CT and outcomes does not support this hypothesis,
weakening the evidence for Cognitive Dissonance Theory as an explanation of MI’s effects.
The Utility of Motivational Interviewing for Improving Program Retention
This trial was the first to apply MI in the context of an educational and employment
program. We found that MI was not effective at improving program retention, except for those
participants who were highly motivated to be and appear consistent. There are several reasons
why MI may not have promoted program retention. First, although MI has been an effective
intervention for target behaviors outside of its original context, substance use (Miller & Rollnick,
2012), its applicability to novel contexts may be limited if there are major differences in how
individuals think about these new target behaviors. For example, MI is conceptualized as helping
individuals resolve ambivalence (Miller & Rollnick, 2012). When applying MI to substance use,
individuals often have strong reasons in favor of quitting (e.g., wanting to improve health,
wanting to follow the law) and continuing to use (e.g., enjoying the feeling of being intoxicated,
wanting to avoid withdrawal symptoms). In contrast, motivation to succeed in an education and
54
employment program may not involve much ambivalence (Bridgeland, Dilulio, & Burke
Morison, 2006), making MI less appropriate in this context.
In addition, this intervention only included one 30-minute MI exercise. Although about
50% of MI trials have involved one-session interventions and the number of sessions has not
been associated with treatment effectiveness (Lundahl et al., 2010), it is atypical to limit MI to
only the Values Card Sort (Miller, Zweben, DiClemente, & Rychtarik, 1995). This choice was
made in order to isolate CT as an active ingredient, allowing for a rigorous investigation of this
proposed mechanism of change. However, this focus on CT may have resulted in a version of MI
that was dissimilar from the versions implemented in other research.
Furthermore, the delivery of MI at the beginning of the program orientation may have
been less effective than offering MI at a later stage to address dips in motivation. Implementing
MI as a prelude to other interventions is a common approach and we thought delivering MI at the
beginning could inoculate participants against threats to motivation throughout the program.
However, considering the overwhelmingly high program motivation reported by participants in
the pretreatment screening measure (i.e., a mean of 9.46 on a 10-point scale), there may not have
been much room for improvement. Rather than delivering MI at a stage in the program when
most participants were looking forward optimistically to educational and occupational success, it
may have been more effective to time the delivery of MI in response to declines in motivation.
Finally, motivational deficits may not be the most important risk factor for failing to
complete an education and employment program. Participants in this study were experiencing
significant life stressors. A substantial proportion of corps members had children, and many
described other family responsibilities in their counseling sessions, such as caring for younger
55
siblings or contributing financially to their parents. There were no exclusionary criteria for
medical, psychiatric, or academic disorders, and many participants described such barriers to
their interventionists. Neighborhood characteristics like poor transportation or community
violence may have interfered with program retention, regardless of motivation to succeed. Some
researchers have argued that MI may not be sufficient in helping individuals overcome socio-
economic or other life stressors (e.g., Befort et al., 2008). If this is the case, LACC may be better
able to improve program retention by teaching corps members problem-solving skills to directly
address stressors and barriers, because MI does not include this type of intervention.
Limitations
Several limitations warrant caution in the interpretation of these results. First, missing
self-report data at 8 and 22 weeks introduced possible bias, so that participants unwilling or
unable to provide data were excluded from the analyses of the complete dataset. Countervailing
strengths were the focus on archival data, which was not missing for any participants and the re-
analysis using multiple imputation. However, the archival data was limited to high school
diploma earning taking place at LACC; if participants left the program early, but made education
or employment gains outside in the community, we were not able to include this in our analyses.
Additional limitations include low statistical power for some analyses, due to small
sample size, low frequencies for some outcomes, and non-normal variables. Although there was
sufficient power to test for main effects of MI and CT on primary outcomes and moderating
effects of PFC, some of the tests of effects on secondary outcomes were underpowered. In these
cases, non-significant findings may have resulted from lack of power rather than a lack of a true
effect. Furthermore, the small sample and reduced power required that we use bivariate analyses
56
in many cases. This limits the nuance of our results, because we could not test multivariate
models or understand the unique relationships between variables, adjusting for covariates.
Finally, although MI and PL sessions were rated as proficient in almost all cases, there
were several unintended differences in counselor behaviors between conditions, but due to
insufficient statistical power, we were unable to control for specific fidelity measures or clinician
effects. However, the goal of equalizing MI and PL conditions on all aspects aside from CT-
elicitation was mostly achieved, and allowed for an important, rigorous test of CT as a
mechanism of change in MI.
Conclusion
This randomized controlled trial moved the MI literature forward in three important
ways. First, this trial demonstrated that MI can cause clients to produce CT, but that CT does not
necessarily lead to positive client outcomes. These results cast doubt on the hypothesis that CT is
a mechanism of change. Second, the marginally significant moderation of MI efficacy by a
measure of dissonance susceptibility suggests Cognitive Dissonance Theory may be an
appropriate framework for understanding MI effects. Third, the lack of efficacy of this particular
iteration of MI for improving retention in an education and employment program highlights that
the implementation of MI to address wide-ranging, novel target behaviors should be continued
with caution and scrutiny. Already, MI has proven effective in several arenas, and there are
likely many undiscovered uses for MI which will be identified in the future. Yet, researchers
should apply rigorous tests to assess how helpful MI can be in each new context. Although MI
did not improve program retention in this education and employment program for young adults,
it is possible that other adjunctive interventions might have an impact on participants’ ability to
57
benefit from programs like LACC. Whether those interventions include problem-solving
training, more intensive MI, or other strategies, future researchers should continue to explore
how young adults can be assisted in making the most of these “second chance” programs.
58
References
Baer, J. S., Beadnell, B., Garrett, S. B., Hartzler, B., Wells, E. A., & Peterson, P. L. (2008).
Adolescent change language within a brief motivational intervention and substance use
outcomes. Psychology of Addictive Behaviors, 22, 570-575. doi:10.1037/a0013022
Barnett, E., Moyers, T. B., Sussman, S., Smith, C., Rohrbach, L. A., Sun, P., & Spruijt-Metz, D.
(2013). From counselor skill to decreased marijuana use: Does change talk
matter? Journal of Substance Abuse Treatment. Advance online publication. doi:10.1016/
j.jsat.2013.11.004
Barnett, E., Spruijt-Metz, Moyers, T. B., Smith, C., Rohrbach, L. A., Sun, P., & Sussman, S. (in
press). Bi-directional relationships between counselor and client speech: The importance
of reframing. Psychology of Addictive Behaviors.
Befort, C. A., Nollen, N., Ellerbeck, E. F., Sullivan, D. K., Thomas, J. L., & Ahluwalia, J. S.
(2008). Motivational interviewing fails to improve outcomes of a behavioral weight loss
program for obese African American women: a pilot randomized trial. Journal of
Behavioral Medicine, 31(5), 367-377. doi:10.1007/s10865-008-9161-8
Bendixen, M., Endresen, I. M., & Olweus, D. (2003). Variety and frequency scales of antisocial
involvement: Which one is better?. Legal and Criminological Psychology, 8(2), 135-150.
doi:10.1348/135532503322362924
Bernburg, J. G. & Krohn, M. D. (2003). Labeling, life chances, and adult crime: The direct and
indirect effects of official intervention in adolescence on crime in early adulthood.
Criminology, 41, 1287-317. doi: 10.1111/j.1745-9125.2003.tb01020.x
59
Bloom, D. (2010). Programs and policies to assist high school dropouts in the transition to
adulthood. The Future of Children, 20, 89-108. doi:10.1353/foc.0.0039
Bridgeland, J. M., Dilulio, Jr., J. J., & Burke Morison, K. (2006) The silent epidemic:
Perspectives of high school dropouts. Washington, DC: Civic Enterprises.
Brown, S. L., Asher, T., & Cialdini, R. B. (2005). Evidence of a positive relationship between
age and preference for consistency. Journal of Research in Personality, 39, 517-533.
doi:10.1016/j.jrp.2004.07.001
CDC. (2011). State and local youth risk behavior survey. Retrieved from http://www.cdc.
gov/healthyyouth/yrbs/pdf/questionnaire/2011_hs_questionnaire.pdf.
Cialdini, R. B., Trost, M. R., & Newsom, J. T. (1995). Preference for consistency: The
development of a valid measure and the discovery of surprising behavioral implications.
Journal of Personality and Social Psychology, 69, 318-328. doi:10.1037/0022-3514.69.2.
318
Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and
standardized assessment instruments in psychology. Psychological Assessment, 6(4), 284.
doi:10.1037//1040-3590.6.4.284
Draycott, S., & Dabbs, A. (1998). Cognitive dissonance 2: A theoretical grounding of
motivational interviewing. British Journal of Clinical Psychology, 37, 355-364. doi:
10.1111/j.2044-8260.1998.tb01391.x
60
Elliott, D.S., Huizinga, D., & Morse, B. (1986). Self-reported violent offending: A descriptive
analysis of juvenile violent offenders and their offending careers. Journal of
Interpersonal Violence, 1, 472-514. doi:10.1177/088626086001004006
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson & Co.
Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. The
Journal of Abnormal and Social Psychology, 58, 203-210. doi:10.1037/h0041593
Fothergill, K. E., Ensminger, M. E., Green, K. M., Crum, R. M., Robertson, J., & Juon, H.
(2008). The impact of early school behavior and educational achievement on adult drug
use disorders: A prospective study. Drug and Alcohol Dependence, 92, 191-199. doi:10.
1016/j.drugalcdep.2007.08.001
Galinsky, A. D., Stone, J., & Cooper, J. (2000). The reinstatement of dissonance and
psychological discomfort following failed affirmation. European Journal of Social
Psychology, 30, 123-147. doi:10.1002/(SICI)1099-0992(200001/02)30
Gaume, J., Bertholet, N., Faouzi, M., Gmel, G., & Daeppen, J. (2010). Counselor motivational
interviewing skills and young adult change talk articulation during brief motivational
interventions. Journal of Substance Abuse Treatment, 39, 272-281. doi: 10.1016/j.jsat.
2010.06.010
Glynn, L. H., & Moyers, T. B. (2010). Chasing change talk: The clinicians role in evoking client
language about change. Journal of Substance Abuse Treatment, 39, 65-70. doi:10.1016/j.
jsat.2010.03.012
61
Heitland, K., & Bohner, G. (2010). Reducing prejudice via cognitive dissonance: Individual
differences in preference for consistency moderate the effects of counter-attitudinal
advocacy. Social Influence, 5, 164-181. doi:10.1080/ 15534510903332261
Jastrab, J., Masker, J., Blomquist, J., & Orr, L. (1996). Evaluation of national and community
services programs: Impacts of service: Final report on the evaluation of American
conservation and youth service corps. Cambridge, MA: Abt Associates.
Landis, J. R. & Koch, G. G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 33(1), 159–174. doi:10.2307/2529310
Lundahl, B. W., Kunz, C., Brownell, C., Tollefson, D., & Burke, B. L. (2010). A meta-analysis
of Motivational Interviewing: Twenty-five years of empirical studies. Research on Social
Work Practice, 20, 137-160. doi:10.1177/1049731509347850
Miller, W. R., & Johnson, W. R. (2008). A natural language screening measure for motivation to
change. Addictive Behaviors, 33(9), 1177-1182. doi:10.1016/j.addbeh. 2008.04.018
Miller, W. R., & Rollnick, S. (2012). Motivational interviewing: Helping people change (3
rd
ed.).
New York: Guilford Press.
Miller, W. R., & Rose, G. S. (2009). Toward a theory of motivational interviewing. American
Psychologist, 64, 527-537. doi:10.1037/a0016830
Miller, W. R., C’de Baca, J., & Matthews, D. B. & Wilbourne, P.L. (2001). Personal values card
sort. Retrieved from http://casaa.unm.edu/inst/Personal%20Values%20Card%20Sort.pdf.
62
Miller, W. R., Moyers, T. B., Ernst, D. & Amrhein, P. (2008). Manual for the motivational
interviewing skill code. University of New Mexico: Center on Alcoholism, Substance
Abuse and Addictions. Retrieved from http://casaa.unm.edu/download/misc.pdf
Miller, W. R., Yahne, C. E., Moyers, T. B., Martinez, J., & Pirritano, M. (2004). A randomized
trial of methods to help clinicians learn motivational interviewing. Journal of Consulting
and Clinical Psychology, 72(6), 1050-1062. doi:10.1037/ 0022-006X.72.6.1050
Miller, W. R., Zweben, A., DiClemente, C. C., & Rychtarik, R. G. (1995). Motivational
enhancement therapy manual. In Project MATCH Monograph Series, Vol. 2. Rockville,
MD: National Institute on Alcohol Abuse and Alcoholism.
Morgenstern, J., Kuerbis, A., Amrhein, P., Hail, L., Lynch, K., & McKay, J. R. (2012).
Motivational interviewing: A pilot test of active ingredients and mechanisms of
change. Psychology of Addictive Behaviors, 26(4), 859-869. doi:10.1037/a0029674
Moyers, T. B. & Martin, T. (2006). Therapist influence on client language during motivational
interviewing sessions. Journal of Substance Abuse Treatment, 30, 245-251. doi:10.1016/
j.jsat.2005.12.003
Moyers, T. B., & Martino, S. (2006). What’s Important in My Life: The Personal Goals and
Values Card Sorting Task for Individuals with Schizophrenia. Retrieved from http://
casaa.unm.edu/inst/Values%20Card%20Sorting%20Task%20for%20Individuals%20with
%20Schizophrenia.pdf
63
Moyers, T. B., Martin, T., Christopher, P. J., Houck, J. M., Tonigan, J. S., & Amrhein, P. C.
(2007). Client language as a mediator of motivational interviewing efficacy: Where is the
evidence? Alcoholism: Clinical and Experimental Research, 31, 40S-47S. doi:10.1111/j.
1530-0277.2007.00492.x
Moyers, T. B., Martin, T., Houck, J. M., Christopher, P. J., & Tonigan, J. S. (2009). From in-
session behaviors to drinking outcomes: A causal chain for motivational interviewing.
Journal of Consulting and Clinical Psychology, 77, 1113-1124. doi:10.1037/a0017189
Moyers, T. B., Martin, T., Manuel, J. K., Miller, W. R., & Ernst, D. (2010). Revised global
scales: Motivational interviewing treatment integrity 3.1.1. University of New Mexico:
Center on Alcoholism, Substance Abuse and Addictions. Retrieved from http://casaa.
unm.edu/download/miti3_1.pdf
Naar-King, S., Outlaw, A., Green-Jones, M., Wright, K., & Parsons, J. T. (2009). Motivational
interviewing by peer outreach workers: A pilot randomized clinical trial to retain
adolescents and young adults in HIV care. AIDS Care, 21, 868-873. doi:10.1080/
09540120802612824
Nail, P. R., Correll, J. S., Drake, C. E., Glenn, S. B., Scott, G. M., & Stuckey, C. (2001). A
validation study of the preference for consistency scale. Personality and Individual
Differences, 31, 1193-1202. doi:10.1016/S0191-8869(00)00218-X
Pirlott, A. G., Kisbu-Sakarya, Y., DeFrancesco, C. A., Elliot, D. L., & MacKinnon, D. P. (2012).
Mechanisms of motivational interviewing in health promotion: A bayesian mediation
64
analysis. The International Journal of Behavioral Nutrition and Physical Activity, 9.
doi:10.1186/1479-5868-9-69
Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best practices for missing data management
in counseling psychology. Journal of Counseling Psychology, 57(1), 1-10. doi:10.1037/
a0018082
Stone, J., & Focella, E. (2011). Hypocrisy, dissonance and the self-regulation processes that
improve health. Self and Identity, 10, 295-303. doi:10.1080/15298868.2010.538550
Sum, A., Khatiwada, I., McLaughlin, J., & Palma, S. (2011). No country for young men:
Deteriorating labor market prospects for low-skilled men in the United States. Annals of
the American Academy of Political and Social Science, 635, 24-55. doi:10.1177/
0002716210393694
Thornberry, T.P., & Krohn, M.D. (2000). The self-report method for measuring delinquency and
crime. In U.S. National Institute of Justice (Ed.), Measurement and analysis of crime and
justice: Criminal justice series (Vol. 4, pp. 33-83). Washington, DC: National Institute of
Justice.
Uggen, C. (2000). Work as a turning point in the life course of criminals: A duration model of
age, employment, and recidivism. American Sociological Review, 65, 529-546. doi:10.
2307/2657381
Vader, A. M., Walters, S. T., Prabhu, G. C., Houck, J. M., & Field, C. A. (2010). The language
of motivational interviewing and feedback: Counselor language, client language, and
client drinking outcomes. Psychology of Addictive Behaviors, 24, 190-197. doi:10.1037/
a0018749
65
Weerman, F. M., Maxson, C. L., Esbensen, F.-A., Aldridge, J., Medina, J., & van Gemert, F.
(2009). Eurogang Program Manual. Retrieved at http://www.umsl.edu/ccj/eurogang/
EurogangManual.pdf
Wessa, P. (2014), Free Statistics Software, Office for Research Development and Education,
version 1.1.23-r7, Retrieved from http://www.wessa.net/
Zullig, K. J., Pun, S., Patton, J. M., & Ubbes, V. A. (2006). Reliability of the 2005 middle school
youth risk behavior survey. Journal of Adolescent Health, 39, 856-860. doi:10.1016/j.
jadohealth.2006.07.008
66
Appendix A: MI Counseling Protocol
(Adapted from Naar-King et al., 2009; Sanchez, 2001)
First, explain to the client that you’d like to know more about what’s most important to
them, and that to do that you would like to try a little exercise, if that’s OK with them. Once the
client indicates interest, ask him or her to sort the values cards into three piles—most important,
important, and not as important to them. Then ask them to rank-order the values in their “most
important” pile from most to least important. This can be very difficult to do! (It’s OK to reflect
this if they’re struggling.) Help them to at least identify their top three values. Then initiate a
discussion with the following questions. You don’t have to rigidly adhere to these questions or
this order, but you should not deviate much. In general, you would only skip a question if the
client has already addressed it completely in answering an earlier question. Use empathic
reflection to help you gain (and express!) understanding of what is truly important for this client.
Summaries may also be quite helpful for the client.
For each of the top 3 values:
1. What does this value mean to you?
2. How are you doing at living out this value?
3. How much time and effort are you putting into living this value?
4. (If necessary.) What behaviors or attitudes prevent you from achieving this value?
For all 3 values, combined:
5. How is your involvement in the Conservation Corps relevant to your living out these
values?
6. How could the Conservation Corps help you promote these values in your life?
67
7. How will you know if you are achieving these values?
8. What has the Conservation Corps told you so far about what you need to do to
succeed?________________________ (If the participant does not describe good
attendance, wearing uniform, no fighting, or not using drugs, say: Is it ok if I explain
what the Conservation Corps told me about what it takes to succeed here? Part of doing
well at the Conservation Corps is coming on time every day, wearing the uniform, not
fighting, and not using drugs.) How do you see these things playing into your ability to
live out these values?
9. (If necessary) What, if anything, do you want to do differently from here on out in order
to live out these values the way you would like?
10. How committed are you to achieving this value (open-ended)? How much do you agree
with this statement on a scale of 1 to 10: “I am very committed to achieving this value”?
Why a ___ and not a ____(lower number)? (Ex: Why a 2 and not a 1? Why a 4 and not a
2?)
*Ask client to write his/her top 3 values on a card and take it home to review periodically. *
68
Appendix B: Placebo Counseling Protocol
(Adapted from Naar-King et al, 2009; Cohen et al., 2006)
First, explain to the client that you’d like to get to know their opinions about different
values, and that you would like to try a little exercise, if that’s OK with them. Once the client
indicates interest, ask him or her to sort the values cards into three piles—most important,
important, and not as important to them. Then ask them to rank-order the values in their “not
important” pile from least to most important. This can be very difficult to do! (It’s OK to reflect
this if they’re struggling.) Help them to at least identify their bottom three values. Use empathic
reflection to help you gain (and express!) understanding of your client’s personal opinions on
these values, even if you disagree. Let them disapprove of the values they choose as
unimportant, but encourage them to come up with at least some understanding of why other
people might care about these values. Summaries and reflections may also be quite helpful for
the client. Your main goal is to promote perspective-taking skills, so that the participant can
come up with some rational understanding of why other people think these values are important.
For each of the 3 least important values:
1. How do you think other people define this value?
2. Why would this value matter to someone else?
3. If someone told you they really cared about this value, what would you think of them?
4. How much time and effort do other people put into living out this value? Why that much
time and effort?
5. What do you think gets in the way of other people trying to live out these values?
69
6. Would it be ok for me to ask you to try a little exercise? Imagine you were someone who
cared a lot about this value. Put yourself in their shoes. Now from their perspective,
what is the most important thing about this value?
7. How much do you understand why someone else would think this value is important on a
scale of 1 to 10? Why a ____ and not a ____?
For all 3 least important values, combined:
8. Sometimes at the Conservation Corps, you have to work with other people with very
different values from yourself. How do you think it would be for you to sit next to
someone in class or work with someone on a crew who thinks _________ (your
unimportant value) is very important?
*Ask client to write the 3 values which may be important to others on a card and take it home to
review periodically.*

Abstract (if available)

Abstract High school dropout puts young adults at risk for unemployment, criminal involvement, and substance use. Education and employment programs, sometimes called “second chance” programs, may help vulnerable young adults obtain high school diplomas, develop job skills, and reduce problem behaviors. With roots in the United States depression era, Conservation Corps programs are some of the most well‐developed examples of education and employment programs. However, the last published evaluation of the Conservation Corps was conducted in the 1990s. Questions regarding the continuing effectiveness of the Conservation Corps must be addressed, especially because there are some indications from the education and employment program literature that retention may be a challenge. It is important to investigate whether failure to complete programs like the Conservation Corps bodes poorly for participant outcomes, to identify risk factors for program non‐retention, and to test adjunctive strategies for assisting participants in engaging and benefitting from these programs. Existing empirical and theoretical work suggests Motivational Interviewing (MI) could encourage participants to use change talk, or motivational language in favor of succeeding at the Conservation Corps. Change talk could then trigger a drive in participants to reduce any cognitive dissonance arising from behaviors inconsistent with that change talk, by working toward their stated educational and occupational goals. ❧ This dissertation consists of two studies conducted at the Los Angeles Conservation Corps (LACC), with 100 young adults. The first study is an evaluation of LACC, which characterized the educational attainment occurring in the program and examined pre‐post changes in problem behaviors. Program retention rates were calculated, and differences in pre‐post changes between program completers and non‐completers were examined. Pre‐treatment risk factors for program non‐retention were also identified. Results indicated that a substantial minority of participants succeeded in earning a high school diploma but that even more participants failed to complete the program. Also, program completers succeeded in reducing their own antisocial behavior and gang membership. Pre‐treatment risk factors for program non‐retention included more years of education and more antisocial behaviors. Participants who were gang members at program entry were more successful at completing the program and earning a high school diploma than non‐gang members. These results suggest that LACC offers some vulnerable young adults an important educational opportunity and a possible pathway to criminal and gang desistance. However, there appears to be a need for improving program retention, so that more participants can benefit from the Conservation Corps. ❧ The second study is a randomized controlled trial of MI for improving program retention. After completing a measure of dissonance susceptibility, participants were randomly assigned to one of three conditions: a) one session of MI designed to elicit change talk, b) one placebo counseling session designed not to elicit change talk, or c) no additional treatment. Change talk was tested as a mechanism of change and dissonance susceptibility was tested as a moderator of MI efficacy. Results indicated that MI was not effective at improving program retention, even though it did promote change talk. Participants who were the most susceptible to dissonance induction were marginally more likely to complete the 8‐week program orientation after receiving MI counseling. These studies are helpful in illuminating the promise of education and employment programs for vulnerable young adults. However, they highlight difficulties in promoting retention. Future researchers should continue to investigate the efficacy of Conservation Corps programs and to test MI in new contexts, with attention to guiding theoretical frameworks and change talk as a mechanism of change.

Linked assets

University of Southern California Dissertations and Theses

Asset Metadata

Creator Smith, Caitlin A. (author)

Core Title A "second chance" program for vulnerable young adults: a program evaluation and a randomized controlled trial of adjunctive motivational interviewing to improve retention

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Publication Date 07/07/2014

Defense Date 05/12/2014

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

Tag commitment language,high school dropout,motivational interviewing,OAI-PMH Harvest,preference for consistency,retention

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Huey, Stanley J., Jr. (committee chair), Margolin, Gayla (committee member), Sinatra, Gale M. (committee member), Wood, Wendy (committee member)

Creator Email caitlin.alka.smith@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-431866

Unique identifier UC11287902

Identifier etd-SmithCaitl-2622.pdf (filename),usctheses-c3-431866 (legacy record id)

Legacy Identifier etd-SmithCaitl-2622.pdf

Dmrecord 431866

Document Type Dissertation

Format application/pdf (imt)

Rights Smith, Caitlin A.

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