Psychosocial Intervention (2018) 27(1) 22-34

Vol. 27. No. 1, April 2018

ISSN: 1132-0559

Psychosocial Intervention Psychosocial Intervention

Editor
Enrique Gracia

Associate Editors
Fernando Chacón
Manuel García-Ramírez
Marisol Lila
Gonzalo Musitu
Douglas D. Perkins

h t t p : / / j o u r n a l s. c o p m a d r i d. o r g / p i
Consejo General
de la Psicología
ESPAÑA

Promoting Positive Youth Development: A Psychosocial Intervention Evaluation

Kyle Eichasa, William M. Kurtinesb, Roberto L. Rinaldic, and A. Celeste Farrc
a
Tarleton State University, USA; bFlorida International University, USA; cOakland University, William Beaumont School of Medicine, USA

ARTICLE INFO A B S T R A C T

Article history: Convergence among treatment, prevention, and developmental psychosocial intervention approaches has led to the
Received 23 June 2017 recognition of the need for evaluation models and research designs that employ a full range of evaluation information
Accepted 21 September 2017 to provide an empirical basis for enhancing the efficiency, efficacy, and effectiveness of prevention and psychosocial
Available online 14 March 2018
positive youth development interventions (PYD). This study reports an investigation of a PYD psychosocial program
using an Outcome Mediation Cascade (OMC) evaluation model, which is an integrated model for evaluating the empirical
Keywords:
intersection between intervention and developmental processes. The Changing Lives Program (CLP) is a community
Positive youth development
engaged PYD psychosocial intervention for multi-ethnic, multi-problem at risk youth in urban alternative high schools.
Psychosocial intervention
Outcome mediation Findings indicated that the hypothesized model fit the data, χ2(7) = 6.991, p = .43, RMSEA = .00, CFI = 1.00, WRMR = .459.
cascade model Findings also provided preliminary evidence consistent with the hypothesis that, in addition to having effects on targeted
positive outcomes, PYD interventions are likely to have progressive cascading effects on untargeted problem outcomes (in
this case, behavioral outcomes) that operate through effects on positive outcomes.

La promoción del desarrollo positivo en la juventud: evaluación de una

intervención psicosocial

R E S U M E N
Palabras clave:
Desarrollo positivo en la juventud
La convergencia entre enfoques de tratamiento, prevención e intervención psicosocial evolutiva ha dado lugar a que se
Intervención psicosocial reconozca que se necesitan modelos de evaluación y diseños de investigación que utilicen una gama completa de infor-
Modelo de mediación en cascada mación sobre evaluación que siente la base empírica para mejorar la eficiencia, la eficacia y la efectividad de la prevención
y de las intervenciones psicosociales en el desarrollo positivo de la juventud (DPJ). Este estudio trata de la investigación
sobre un programa psicosocial sobre DPJ utilizando el Outcome Mediation Cascade (OMC), modelo integrado para eva-
luar la intersección empírica entre la intervención y los procesos de desarrollo. El Changing Lives Program (CLP) es una
intervención psicosocial de DPJ en la que está implicada la comunidad para jóvenes multiétnicos en riesgo en colegios
urbanos de secundaria alternativos. Los resultados indican que el modelo postulado se ajusta a los datos, χ2(7) = 6.991,
p = .43, RMSEA = .00, CFI = 1.00, WRMR = .459. Igualmente, los resultados aportan pruebas preliminares consistentes con
la hipótesis de que además de tener influencia en los resultados positivos planeados es probable que las intervenciones
DPJ tengan efectos progresivos en cascada en los resultados no previstos de los problemas (en este caso, resultados en el
comportamiento) que operan a través de sus efectos en los resultados positivos.

PYD Interventions: Targeting Identity Formation Processes that target identity formation has also been discussed extensively
within the identity literature (Archer, 1994; Erikson & Erikson, 1957;
Initial evidence of the effectiveness of positive youth development Hernandez, Montgomery, & Kurtines, 2006; Kurtines et al., 2008;
(PYD) programs in promoting positive change in concepts and Montgomery, Hernandez, & Ferrer-Wreder, 2008; Montgomery et al.,
constructs related to positive identity has led to an interest in 2008).
developing, implementing, and evaluating interventions that As Ferrer-Wreder, Montgomery, and Lorente (2003) point out,
specifically target core components of identity formation in the literature on identity formation has a potential contribution
adolescents as a strategy for promoting positive outcomes and to make to efforts to promote positive development and prevent
preventing negative outcomes. The potential utility of interventions problematic outcomes, in part because it offers guidance for the

Cite this article as: Eichas, K., Kurtines, W. M., Rinaldi, R. L., & Farr, A. C. (2018). Promoting positive youth development: A psychosocial intervention evaluation. Psychosocial
Intervention, 27, 22-34. https://doi.org/10.5093/pi2018a5

Correspondence: eichas@tarleton.edu (K. Eichas).

ISSN:1132-0559/© 2018 Colegio Oficial de Psicólogos de Madrid. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
23 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

effort to identify and investigate the component processes of identity also known as identity exploration (Adams et al., 2001; Waterman,
formation. Montgomery et al. (2008) suggest that the inclusion 1999), is considered a normal and healthy developmental challenge
of variables tapping key identity processes and outcomes, such as during adolescence that involves processes necessary for finding,
identity cohesion, style, distress, and turning points, in psychosocial assessing, and establishing identity commitments.
intervention research would make these interventions for youth more
effective and potent, in part because this type of research would add Self-Construction: The Identity Style Model
to knowledge about for whom interventions work and why they work.
Within the identity literature, two core identity processes have Within the identity exploration literature, a significant body
been discussed extensively: identity exploration and identity of work has drawn upon the constructivist tradition in asserting
commitment (Marcia, 1966, 1980, 1988). Elaborating upon Erikson’s that the individual is an intentional agent who participates in the
writing, Marcia (1966) described behavioral markers of the processes construction of his or her world. The individual is depicted in the role
by which an adolescent consolidates a coherent sense of identity. of a scientist, proactively making identity-related choices by forming
Adolescents face the challenge of first exploring multiple possible and testing hypotheses in a cognitive, rational, and dispassionate
identity alternatives in order to make decisions about life choices manner (Berman, Schwartz, Kurtines, & Berman, 2001; Berzonsky,
and then choosing one or more of the multiple possible identity 1989; Grotevant, 1987). Identity theories in this vein have emphasized
alternatives and following through with them (Grotevant, 1987; abilities and orientations (Grotevant, 1987), cognitive problem-
Marcia, 1980, 1988; Schwartz, 2001). According to this formulation, solving competence (Berman et al., 2001), and cognitive processing
identity exploration is the search for a revised and updated sense of styles (Berzonsky, 1989).
self, and identity commitment is the adherence to a particular course Berzonsky’s (1989) constructivist approach to identity formation
of action characterized by a self-selected specific set of goals, values, emphasizes the cognitive processing orientations with which a
and beliefs (Marcia, 1988; Schwartz, 2001). person forms and maintains an identity. Drawing on the work of Kelly
As Bosma and Kunnen (2001) point out, a number of theoretical (1955), Berzonsky (1993) proposed that people are self-theorists in
perspectives on identity development (e.g., Adams & Marshall, 1996; that they create a conceptual structure that helps them make sense
Breakwell, 1988; Grotevant, 1987; Kerpelman, Pittman, & Lamke, of their experience. There are three types of self-theorists: (1) the
1997; Kroger, 1997) have sought to extend and refine Marcia’s (1966) scientific information-oriented, (2) the dogmatic normative-oriented,
conceptualization of identity commitment and identity exploration and (3) the ad hoc diffuse/avoidant-oriented. When applied to the
by emphasizing the transaction between the individual and the psychosocial task of identity formation, these orientations toward
environment. According to Bosma and Kunnen (2001), conflict or self-construction are termed identity styles. The informational
mismatch between existing identity commitments and information orientation is characterized by the seeking out and utilizing of self-
from an individual’s environment represents the “starting point” relevant information when making decisions related to identity
for the identity process, a conceptualization that is consistent (Berzonsky, 1993), while the normative orientation is characterized
with the view of adolescence as a developmental period with by conformation to the expectations of others or of reference groups.
immense personal and contextual change, including biological, The diffuse avoidant identity style is often distinguished by avoidance
cognitive, and social changes. Whereas the conflict between of identity-related problems.
existing identity commitments and the environment has been Most individuals have the ability to use all three identity
described as the “starting point” for the identity process (Bosma orientations by late adolescence (Berzonsky, 1990). Though not
& Kunnen, 2001), identity exploration has been described as the necessarily more advanced than normative style, the informational
“work” of the identity process (Grotevant, 1987) and includes both style has been associated with several indices of positive
cognitively focused and emotion focused components (Schwartz, psychosocial adjustment in adolescence, including: openness
2002; Soenens, Berzonsky, Vansteenskiste, Beyers, & Goosens, 2005). to ideas and experience and the need to exert cognitive effort
Marcia’s conceptualization of identity exploration maps directly (Berzonsky, 1990); active, problem-focused coping strategies and
onto Erikson’s description of the identity crisis, and the terms are social support seeking (Berzonsky & Sullivan, 1992); and self-
usually used interchangeably (Adams et al., 2001; Kidwell, Dunham, reflective tendencies, salience of personal identity content, and
Bacho, Pastorino, & Portes, 1995; Waterman, 1999). openness to feelings and fantasies (Berzonsky & Sullivan, 1992).
Adolescents who score higher on measures of informational
Self-Construction and Self-Discovery in Adolescence style tend to come from authoritative homes (Berzonsky, 2004),
have well-defined educational plans, and have high levels of
According to Erikson (1963), the ascendant psychosocial task of both conscientiousness and goal-directedness (Berzonsky & Kuk,
adolescence is the integration of the roles, skills, and identifications 2000). Though evidence suggests that while both normative and
youth have learned in childhood with the expectations of the informational styles are generally more adaptive than diffuse
adult world into a coherent sense of identity. The search for this avoidant style, the informational style is more adaptive than
sense of identity is characterized by the crisis between the ego the normative style in situations in which youths must assume
syntonic tendency toward identity synthesis, a self-constructed personal responsibility for academic priorities and monitor their
dynamic organization of drives, abilities, beliefs, and individual own activities and progress (Berzonsky & Kuk, 2000).
history, and the ego dystonic tendency toward identity confusion,
the lack of such an organization. Crisis is thought to be prompted Self-Discovery: Feelings of Personal Expressiveness
by either an identity deficit, in which the adolescent encounters
circumstances in which he or she lacks enough of a sense of Within the identity literature, a body of work is emerging in
identity to make important life decisions, or an identity conflict, in the humanist tradition, resonant with the rapidly growing interest
which the adolescent experiences circumstances that bring to light in positive psychology (Seligman & Csikszentmihalyi, 2000), to
the incompatibility of two or more aspects of his or her identity. emphasize the consideration of personal strengths and creative
Both syntonic and dystonic elements are necessary in the positive potentials. The individual is depicted as becoming or growing toward
development of the adolescent as the dialectic tension between the fulfillment of his or her potential. Theories in this vein have
them represents a time of increased vulnerability and potential asserted the importance of self-actualization (Maslow, 1968), flow
for development (Erikson, 1968, 1985). As such, the identity crisis, (Csikszentmihalyi, 1990), and feelings of personal expressiveness
 Positive Youth Development 24

(Waterman, 1990). Flow is an affective state characterized by a Evaluating the Formation of a Self-Structure:
balance between the challenge at hand and the skills one brings to it Relational Data Analysis of Life Goals
(Csikszentmihalyi, 1990), while self-actualization refers to fulfilling
one’s potentials and living up to one’s ideals on a consistent basis It is important to note that the construction of this self-structure
(Maslow, 1968). Personal expressiveness emphasizes the subjective has many measureable dimensions (e.g., identity exploration
experience associated with engaging in identity-related activities and commitment, identity style), any of which may potentially
(Waterman, 2005) and the self-discovery process (Waterman, 1984). be measured and evaluated with quantitative strategies for
Personal expressiveness is considered the second of these analyzing linear, additive change. However, a strictly quantitative
successively more integrated levels of affective processing, approach cannot capture changes in the meaning and significance
more integrated than the experience of flow and less integrated of critical experiential components of an individual’s sense of
than self-actualization (Schwartz, 2002, 2006; Waterman, 1990). self and identity because these changes are subjective in nature
Feelings of personal expressiveness are defined as the positive, and characterized by an increasingly complex (i.e., integrated)
subjective state characterized by the deep satisfaction that structural organization with emergent new properties. A
accompanies engagement in activities or goals that utilize one’s quantitative approach is not capable of the detection of qualitative
unique potentials and that are hypothesized to represent one’s change that involves the emergence of new content domains or
basic purpose in living (Waterman, 1993). While performing new structural organizations outside empirically or theoretically
activities that evoke feelings of personal expressiveness, pre-selected content domains because fixed response quantitative
individuals experience (a) an unusually intense involvement, (b) a measures are by design not capable of collecting and capturing
special fit or meshing with the activities, (c) a feeling of intensely data outside of their pre-specified content domains. On the other
being alive, (d) a feeling of completeness or fulfillment, (e) an hand, as discussed previously, free-response qualitative measures
impression that this is what one was meant to do, and (f) a feeling are capable of capturing the emergence of new content domains or
that this is who one really is (Waterman, 2005). new structural organizations.
Existing research suggests that feelings of personal While previous studies have evaluated the effect of the CLP
expressiveness are associated with many positive life outcomes intervention on change in identity processes (e.g., Albrecht, 2007;
(Csikszentmihalyi, 1975, 1990; Waterman, 1993, 2004). Eichas et al., 2010) or change in subjective sense of self and identity
Participation in personally expressive activities is related to higher through the examination of narrative expressions of meaning and
levels of intrinsic motivation to accomplish life tasks (Waterman, significance (e.g., Arango, Kurtines, Montgomery, & Ritchie, 2008;
2005), as well as higher scores on measures of perceived Kortsch, Kurtines, & Montgomery, 2008), and intervention change
competence and self-realization values and importance. Evidence in identity processes as mediators of intervention change in sense of
suggests a strong association between personal expressiveness self and identity and sense of self and identity (Eichas et al., 2010),
and self-determination (Waterman et al., 2003). Recent studies no studies, to date, have incorporated both quantitative measures of
conducted with high school students have also found evidence identity processes and multiple qualitative assessments of subjective
of the association between personal expressiveness and several sense of self and identity. The current study sought to contribute
indices of positive psychosocial adjustment in adolescence. Feelings to the evaluation of Eichas et al., (2010) outcome mediation model
of personal expressiveness, in combination with flow and goal- (OMC) and, toward that end, had two main research aims.
directed behavior, have been found to be significantly associated The first aim of this study was to investigate CLP intervention
with higher levels of adolescent-reported psychological well-being effectiveness in promoting positive change in self-constructive and
and lower levels of adolescent reported problem behavior (Palen self-discovery identity exploration processes and level of identity
& Coatsworth, 2007). While there appear to be gender differences resolution in a psychosocial intervention and effects on behavioral
in the types of activities that males and females find personally outcomes. The second aim was to investigate examined intervention
expressive (socializing, instrumental, and literary activities for change in identity processes as mediators of intervention change in
females and sports/physical activities for males), reported levels subjective (qualitative) sense of self and identity.
of personal expressiveness within those activities have been found This study furthered the evaluation of the OMC by identifying
to be similar across gender (Sharp, Coatsworth, Darling, Cumsille, which core identity processes (i.e., cognitive identity exploration,
& Ranieri, 2007). Research also suggests that there are more identity commitment, and emotion focused identity evaluation)
similarities in feelings of personal expressiveness across countries serve as mediators of effects of participation in the CLP on change
and cultures than differences (Coatsworth et al., 2005; Sharp et al., in sense of self and identity and behavioral outcomes.
2007).
Conceptual and Theoretical Analyses:
Self-Construction and Self-Discovery Identification of Properties
in Positive Youth Development
The short-term outcome study reported here was conducted
Because adolescents must begin to choose and commit to the as part of the evaluation of the Changing Lives Program (CLP)
goals, roles, and beliefs about the world that give life direction and participatory transformative intervention approach in promoting
purpose as well as coherence and integration (Erikson, 1968), models both positive change in sense of self and identity among troubled
of identity development that integrate the self-construction and adolescent youth. The CLP was implemented by the Miami Youth
self-discovery processes underlying identity exploration may be of Development Project as a PYD psychosocial program targeting
particular utility in psychosocial interventions, such as the PYD, that multi-problem youth in alternative high schools in the Miami Dade
focus on self-construction and self-discovery among adolescents. County Public Schools (M-DCPS), the fourth largest school system in
Previous positive youth development research has recognized the United States. The CLP is a long-standing community engaged,
the potential of models integrating self-construction and self- gender, and ethnic inclusive PYD psychosocial intervention. The
discovery processes to guide the development, implementation, CLP specifically targets these youths because students who come to
and evaluation of psychosocial developmental interventions with alternative schools are on a negative life course pathway and at risk
adolescents exposed to risks associated with living in low-income for multiple negative developmental outcomes and/or engaged in
urban community contexts (Albrecht, 2007). multiple problem behaviors.
25 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

The CLP provides an on-site (in school) psychosocial intervention e1 e2

in all of the M-DCPS voluntary alternative high schools. The primary
PE n Life Goal
intervention goal is to create contexts that empower troubled at FU1 at FU1
adolescents in ways the promote the development of a positive r
q
identity and, as described next, also result in positive change in h t
problem outcomes, thereby changing their “negative” life pathways
b i s INT
into positive ones. e3
c at FU2
e4
Youth in alternative high schools are at increased risk for l
multiple problem behaviors (Grunbaum et al., 2000; Windle, 2003). a IDCR
CLP m
Although the effects of treatment and prevention interventions on at FU1
e j
problem behaviors have been widely researched, evidence for PYD k
EXT
intervention effects (direct or cascading) on problematic outcomes o at FU2 e5
d
is scant or nonexistent (Schwartz et al., 2007). The literature that has
begun to emerge has largely focused on the potential direct effects p
of PYD programs in reducing internalizing problem behaviors (e.g.,
EXT EXT
depression, anxiety, etc.) (McWhinnie, Abela, Hilmy, & Ferrer, 2008). at FU1 at FU1
There has been little research on the mediational processes by which
these PYD effects may occur. e6 e7
A main conceptual focus of the ongoing program of developmental
intervention research on the Changing Lives Program (CLP)
intervention has been on the theory informed selection of mediators Figure 2. Hypothesized OMC Evaluation Model.
and positive outcome variables, as well as a broad and representative
array of problem outcome variables for evaluating CLP outcomes.
The CLP uses a participatory transformative intervention process to
target the formation of a positive sense of self and identity among Method
multi-problem youth attending alternative high schools. Thus, this
study used quantitative and qualitative indices of positive identity Participants
development drawn from the literature on self and identity.
Participants for this study were drawn from the Changing
Lives Program. In addition to a core family of measures (group
Positive administered self-report, individually administered open ended,
Identy and semi-structured interview measures, etc.), a core of specialty
at FU1
measures associated with specific evaluation projects (such as this
b+
study) were utilized. The sample for this study was comprised
d+
h-
of 259 White/Non-Hispanic, African-American, and Hispanic
adolescents. Due to under-representation, the 21 White/Non-
Identy
a+ Crisis/ Hispanic individuals were not included in the sample. Analyses
CLP
Exploration were conducted with 238 African-American and Hispanic
at FU1 g+
adolescents aged 14-18 who had completed the intervention, 98
c-
e+
of whom participated in a two-semester non-intervention (non-
random) comparison control condition. The sample consisted of
137 females (87 African American, 50 Hispanic) and 101 males (52
Problem f+/- Problem
Behaviors Behaviors African American, 49 Hispanic). With regard to the socio-economic
at FU1 at FU2 characteristics of the sample, 38% of annual family incomes were
below $21,000, while 17% were over $41,000. Seventy-four percent
of the participants had at least one parent who completed high
Figure 1. Conceptualized Pathways of Intervention Change. school, 50% had two parents who completed high school, and 31%
had at least one parent who completed a bachelor’s degree.
Figure 1 visually depicts the basic conceptualized pathways of
intervention change, including (1) direct effects on positive identity Procedure
outcomes, (2) direct effects on problem behavior outcomes, (3) indi-
rect effects on positive identity and problem behavior outcomes me- Intervention procedure. Following the Youth Development
diated by hypothesized identity formation processes, and (4) indirect Project’s established procedures, each intervention group was led
effects on problem behavior outcomes mediated by positive identity by an intervention team that consisted of one group facilitator, one
outcomes, that is, cascading positive change that spills over across co-facilitator, and one or two group assistants. All groups shared this
positive and problem outcomes. structure and format. All group facilitators and co-facilitators were
graduate level students enrolled in either a doctoral or a master’s
Hypothesized Pathways of Intervention Change: Direct, level program. Group assistants were undergraduate psychology
Mediated, and Cascade Effects students who had been trained in the administration of the measures
and in participant tracking procedures.
Figure 2 visually presents the Outcome Mediation Cascade (OMC) The group facilitators and co-facilitators served as counselors and
evaluation model used to investigate intervention effects (including used the CLP’s intervention strategy, a participatory transformative
measures of both positive and problem outcomes). Measured varia- approach (Montgomery et al., 2008). The intervention groups met for
bles were selected to operationalize the conceptual model described approximately 45 minutes to 1 hour every week for approximately 8
above (see Measures section for a description of the measures). to 12 weeks in either the fall or spring semester.
 Positive Youth Development 26

Assessment procedure. The students were assessed by under- significant or important to you? How significant or important is this
graduate psychology students serving as research trainees. Their to you?” The meaning and significance questions were then followed
training took place at the beginning of each semester and included by three neutral probes (e.g., “Can you say more about that?”, “Is there
instruction concerning confidentiality issues, assessment adminis- anything else?”) to request secondary elaboration when necessary.
tration, dress code, high school regulations, interviewing strate- Analysis of inter-coder agreement for level 1 subcategories
gies, and role-playing of interviews. Assessments were conducted indicated agreement of 96% and a Fleiss’ kappa of .84, suggesting
at three times during the school year on school grounds and during almost perfect agreement (correcting for chance). The Personally
school hours. Assessments took place the week preceding the com- Expressive category had two nested subcategories: Personally
mencement of the semester sessions and the week after the end of Expressive through Others and Personally Expressive through
each semester’s sessions. Self. The Non-Personally Expressive category had three nested
subcategories: Prove to Others, Self-Satisfying, Benefit of Others;
Measures and two mixed nested subcategories: Mixed Prove to Others/Self
Satisfying and Mixed Self Satisfying/Benefit of Others. Analysis of
Psychosocial mediator: Identity conflict resolution. The Erikson inter-coder agreement for level 2 subcategories indicated agreement
Psycho-Social Stage Index (EPSI; Rosenthal, Gurney, & Moore, 1981) of 86% and a Fleiss’ kappa of .69, suggesting substantial agreement
is a 72-item self-report measure that includes six subscales, each (correcting for chance). Analyses of the full model used level 1
consisting of 12 items indicating how well respondents have resolved subcategories, that is, personally expressive versus non-personally
conflicts indicative of Erikson’s stages of psychosocial development. expressive life goals (Life Goal).
This study used the identity conflict resolution subscale (IDCR). Items Problem outcomes: Internalizing and externalizing problem
on this subscale utilize key words and statements from Erikson’s behaviors. The Behavior Problem Index (BPI; Peterson & Zill,
characterizations of the identity synthesis versus identity confusion 1986) was used to assess internalizing and externalizing problem
stage. Sample items included “I know what kind of person I am.” behaviors. The 32 items of the BPI were taken from the Achenbach
Items are rated on a 5-point Likert scale from 1 (almost never true) to Behavior Problems Checklist (Achenbach & Edelbrock, 1981), a more
5 (almost always true) with half of the items representing resolution extensive measure that is widely used with children and adolescents.
of the identity crisis, and half representing identity confusion. Items The BPI items were included in the Panel Study of Income Dynamics
representing identity confusion were reverse-coded prior to analysis. – Child Development Supplement II (PSID-CDS II; Mainieri, 2006), a
Thus, mean scores for the IDCR subscale yield an index of degree of national survey of children and adolescents, to describe a number
resolution, with high scores indicating higher levels of resolution. of behavior problem areas over the past 3 months, which are scored
Rosenthal et al. (1981) reported a correlation coefficient of .56 for Internalizing (INT) and Externalizing (EXT) problem behaviors.
between the identity conflict resolution subscale of the EPSI and Mainieri (2006) reported Cronbach’s alphas of .86 for EXT and .83
the identity subscale of Greenberger, Knerr, Knerr, and Brown (1974) for INT in the PSID-CDS II sample. Cronbach’s alphas for the present
Psychosocial Maturity Inventory (PSM), Form D. Alpha coefficients sample were .81 and .85 for the EXT and INT, respectively.
of .74 have been reported in an adolescent sample (Ferrer-Wreder, Moderators of outcome (gender and ethnicity). Background Infor-
Palchuk, Poyrazli, Small, & Domitrovich, 2008) and in a large youth mation Form (BIF) is a record of demographic information completed
sample ranging across middle school through college (Montgomery, by all participants in the YDP program. It provided the data used in
2005). In this study, Cronbach’s alpha coefficient was .77. analyses of gender and ethnicity (Hispanic/Latino, African American,
Positive outcome: Feelings of personal expressiveness. The Non-Hispanic White, Bi-ethnic, and Other) as exogenous moderators.
Personally Expressive Activities Questionnaire (PEAQ; Waterman,
1995) is a 14-item self-report measure that includes three subscales
Results
indicating feelings of personal expressiveness, hedonic enjoyment,
and flow challenge. This study used the feelings of personal Descriptive Statistics
expressiveness subscale (PE) as a measure of differential intervention
outcome in the positive domain. The PEAQ has been adapted for use in Table 1 presents the means, standard deviations, skewness, and
the evaluation of CLP such that questions refer to activities associated kurtosis for the continuous outcome variables at Baseline (BL), Follow-
with the long-term life goals of the respondents. Thus, the PE subscale up 1 (FU1), and Follow-up 2 (FU2). These include the subscales of
indicates the degree to which respondents feel that the pursuit of life the Personally Expressive Activities Questionnaire (PEAQ): Personal
goals is personally satisfying and expressive of their unique potentials. Expressiveness (PE); the Erikson Psycho-Social Stage Index (EPSI):
The PE subscale consists of six items, the scores of which are averaged Identity Conflict Resolution (IDCR); and the Behavior Problem Index
to generate a subscale score. Sample items included “When I do these (BPI): Internalizing (INT) and Externalizing (EXT) problem behaviors.
activities, I feel like it’s what I was meant to do.” Items are rated on
a 7-point Likert scale from 1 (strongly disagree) to 7 (strongly agree)
Table 1. Descriptive Statistics for Continuous Outcome Variables
(Waterman, 1995). Waterman (2004) found significant correlations
between PE scores and identity status and identity style, providing Outcomes M SD Skewness Kurtosis
evidence of concurrent validity. Waterman (2005) reported an alpha IDCR at BL 3.725 0.566 -0.306 -0.426
reliability coefficient of .77 for the PE subscale; in this study it was .91. IDCR at FU1 3.739 0.617 -0.456 0.091
Positive outcome: Personally expressive life goals. The Personally PE at BL 5.531 1.295 -1.055 1.014
Expressive Activities Questionnaire – Qualitative Extension (PEAQ- PE at FU1 5.305 1.469 -0.736 -0.182
QE) (Rinaldi et al., 2012) adds an open-ended response component INT at BL 1.680 0.446 0.462 -0.533
to the PEAQ to provide a method for eliciting the narrative/linguistic INT at FU1 1.571 0.392 0.642 0.066
expressions of meaning and significance of participants’ most INT at FU2 1.601 0.403 0.475 -0.437
important life goals. Specifically, participants were asked to identify EXT at BL 1.771 0.401 0.275 -0.185
up to three life goals and then to identify their most important life EXT at FU1 1.709 0.412 0.272 -0.671
goal. They were then asked to provide an open-ended description EXT at FU2 1.679 0.376 0.470 0.160
of its meaning and significance. More specifically, participants were Note. IDCR = Identity Conflict Resolution; BL = Baseline; FU1 = Follow-up 1;
asked “What does this life goal mean to you?” and “Why is this PE = Personal Expressiveness; INT = Internalizing; EXT = Externalizing; FU2 = Follow-up.
27 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

Table 2 presents the number and proportion of responses assigned the simple main effects are included in the model. Participants’ age
to level 1 and level 2 subcategories of the Life Goal variable. Baseline and scores of outcomes at BL were included as covariates.
evaluation indicated that 92% of descriptions of participants’ most The fit of Model 1 was evaluated with Mplus 5.0 (Muthen &
important life goals had the non-personally expressive property and, Muthen, 1998-2007) using the sample covariance matrix as input and
more specifically, that the majority of responses (75%) had the self- a robust weighted least squares solution. The model is statistically
satisfying property. This proportion was similar for both males and overidentified. Data analysis used the robust WLSMV estimator
females and for both Hispanic and African American participants available in Mplus in order to adjust the parameter estimates,
(ranging from 75% to 76% within each subgroup), as well as among standard errors, and fit indices for the categorical nature of the Life
age subgroups (ranging from 71% to 83%) The variation in proportions Goal variable. Computation of standard errors and the chi-square test
by age was not significant (χ2 = 1.515, df = 4, p = .82). of model fit took into account non-independence of observations due
to cluster sampling, specifically to account for possible counseling
Table 2. Descriptive Statistics for Life Goal group clustering effects.
Cases with incomplete covariate data were not included in the
Level 1 Subcategories n %
model analysis, resulting in a data set of 212 cases. These data were
Personally Expressive Life Goal at BL 18 8%
assessed for missingness and appear to meet the specified criteria.
Personally Expressive Life Goal at FU1 18 8%
Although a high percentage of data for the FU2 evaluation was missing
Non-Personally Expressive Life Goal at BL 206 92%
(42% missing for INT and EXT at FU2 and 9% or less missing data for
Non-Personally Expressive Life Goal at FU1 211 92%
FU1 variables), Little’s MCAR test was found to be non-significant (χ2
Level 2 Subcategories n %
= 26.980, df = 24, p = .31). In addition, dummy variables were created
Personally Expressive through Self at BL 17 8%
for missing data and correlated with gender, ethnicity, age, and school
Personally Expressive through Self at FU1 15 7%
location, as well as all other variables included in the model. Missing
Personally Expressive through Others at BL 1 0.4%
data were not strongly correlated (all correlations were less than .17)
Personally Expressive through Others at FU1 3 1%
with any of these variables.
Benefit of Others at BL 12 5%
Following recommendations of Bollen and Long (1993), a variety of
Benefit of Others at FU1 6 3%
global fit indices were used, including indices of absolute fit, indices of
Self-Satisfying at BL 168 75%
relative fit, and indices of fit with a penalty function for lack of parsimony.
Self-Satisfying at FU1 174 76%
First, the chi-square and its probability value (p-value) were examined.
Prove to Others at BL 3 1%
The overall chi square test of model fit of Model 1 was statistically
Prove to Others at FU1 2 0.9%
non-significant, χ2(7) = 6.991, p = .43. The RMSEA was .00, WRMR
Mixed Prove to Others/Self Satisfying at BL 13 6%
was .459, and the CFI was 1.00. Thus, fit indices were consistent with
Mixed Prove to Others/Self Satisfying at FU1 9 4%
good model fit. The model in Figure 3 presents statistically significant
Mixed Self Satisfying/Benefit of Others at BL 10 4%
paths in bold, while Tables 3 and 5 present the parameter estimates
Mixed Self Satisfying/Benefit of Others at FU1 20 9%
for the major analyses.
Note. BL = Baseline; FU1 = Follow-up 1.
.79 .71
Outliers, Missing Data, and Non-Normality
PE n -.141* Life Goal
at FU1 at FU1
Prior to the study’s main analysis, data for the continuous r
q -.153*
variables were evaluated for outliers and normality. Outliers were t
h
evaluated by examining leverage statistics for each individual; an -.417* -.209† -.612*
outlier was defined as an individual with a leverage score four b i s INT
e3
c at FU2
times greater than the mean leverage. No outliers were found. e4
AA: -.215* l -.205*
Kurtosis and skewness were within acceptable ranges (see Table 1). a H: .167 IDCR
CLP m -.093† .074*
at FU1
e j
Main Analyses k -.164*
-.171* EXT
o at FU2 e5
d -.178*
Model 1 (Figure 1), the hypothesized Outcome Mediation Cascade -.114†
(OMC) model, was evaluated using a Structural Equation Modeling p
(SEM) approach for modeling hypothesized causal pathways of
EXT EXT
intervention change. Model specification was guided by Figure 2. at FU1 at FU1
Figure 3 and Tables 3 and 5 present the results (outcome, moderation,
.69 .70
and mediation) for Model 1, including relations between the
.078*
hypothesized mediator (IDCR), the two positive identity outcome
variables (PE and Life Goal), and the two problem behavior outcome
variables (INT and EXT) for the full sample. Figure 3. OMC Model Results.
Interaction effects for the moderated relationships were tested
using product terms (Jaccard & Turrisi, 2003). Two-valued dummy Hypothesized Direct and Moderated Intervention Effects
variables for two exogenous interpersonal contextual covariates,
Gender (G) and Ethnicity (E), were included in the analysis of Following the logic of Rausch, Maxwell, and Kelly (2003), the
outcome as measured at pretest, as were all possible interaction scores of the baseline measures (PE, IDCR, Life Goal, INT, EXT) were
terms (i.e., CLP*G, CLP*E, G*E, CLP*G*E), with CLP designated as used for the analysis of covariance of a quasi-experimental outcome
the focal independent variable. Non-significant interactions were design with two waves of assessment (BL, FU1) to evaluate whether
dropped from the final model, leaving one significant interaction participation in the CLP was associated with change in IDCR, PE,
term (CLP*E). To reduce clutter and improve visual clarity, the model Life Goal, INT, and EXT relative to the comparison control condition.
in Figure 3 excludes the paths associated with E and CLP*E; however, Specifically, CLP was defined as a two-valued dummy variable (scored
 Positive Youth Development 28

1 or 0) for the two intervention conditions (CLP vs. comparison & Sheets, 2002) by significance of the paths associated with the
control). By design, difference in this variable was hypothesized to be causal chains between distal variables and change in hypothesized
related to differential outcome (change in IDCR, PE, Life Goal, INT, EXT) mediating variables (including positive outcomes hypothesized to
at FU1 (IDCR, PE, Life Goal, INT, EXT) controlling for BL (IDCR, PE, Life mediate change in problem outcomes) and, in turn, the association
Goal, INT, EXT). The hypothesized differences were evaluated using between change in mediating variables and change in outcome
covariate-adjusted change in which the measure of the outcome at variables. In addition, the baseline measures of the hypothesized
BL and the outcome at the FU1 are strategically used as covariates to mediators were modeled as covariates for the outcomes variables
define different features of change (Rausch et al., 2003). (and thus held constant) in order to model the relationship between
As can be seen from Figure 3 and Table 3, the pattern of findings change in mediators and change in outcomes. All hypothesized
provided evidence of a significant relationship between CLP and mediated relationships were modeled in this manner; these paths
change in PE (path b; b = .417, p < .01, 95% CI = 0.145 to 0.689). Exa- were not included in Figure 3 to improve visual clarity.
mination of group means indicates that PE scores for the intervention CLP had a statistically significant relationship with change in IDCR,
group decreased less than PE scores for the comparison control group the hypothesized mediator. Specifically, the path coefficient for the
(see Table 4 for mean change). For Life Goal, a personally expressive relationship between CLP and change in IDCR was statistically sig-
life goal (i.e., life goals with the personally expressive property) versus nificant, but the relationship was moderated by E, indicating a mod-
a non-personally expressive life goal (i.e., life goals without the per- erated specificity of effect. The 2-way CLP x E interaction term had
sonally expressive property) was also represented by a two-valued a statistically significant coefficient when predicting change in IDCR
dummy variable (scored 1 = assignment to the Personally Expressive (b = .368, p < .01, 95% CI = 0.102 to 0.633), indicating that intervention
subcategory, 0 = assignment to the Non-personally Expressive sub- change in IDCR differed between ethnic groups. Among Hispanic par-
category). Results indicated that the direct relationship between CLP ticipants, the path coefficient for the relationship between CLP and
and change in Life Goal (path c) was not significant. As can be seen change in IDCR was not significant (path a; b = .167, p = .15, 95% CI
from Figure 3 and Table 5, the pattern of findings also did not provide = -0.060 to 0.394). Among African American participants, the path
evidence of significant relationships between CLP and changes in the coefficient was significant (path a; b = -.215, p < .001, 95% CI = -0.309
untargeted problem outcome variables (INT and EXT; paths d and e). to -0.120), indicating that the intervention group decreased in IDCR
relative to the comparison control group.

Table 3. Summary of Major Path Analyses for Mediator and Positive Outcomes
Findings provided support for IDCR as a plausible moderated me-
Outcomes Predictors Path coefficient p-value 95% CI diator of the relationship between CLP and change in PE and between
CLP*E .368 (.626) .007 0.102 to 0.633 CLP and qualitative change in Life Goal. Findings indicated that BL to
CLP – Hispanic FU1 change in IDCR was marginally significantly associated with con-
.167 (.284) .150 -0.060 to 0.394
(path a)
IDCR2 temporaneous change in PE (path h; b = .209, p < .08, 95% CI = -0.022
CLP – African
-.215 (-.366) < .001 -0.309 to -0.120 to 0.440). The confidence interval for the indirect effect of the CLP*E
American (path a)
product term was computed using the PRODCLIN program (MacKin-
Age -.037 (-.070) .284 -0.104 to 0.030
non, Fritz, Williams, & Lockwood, 2007) and provided further support
CLP (path b) .417 (.282) .003 0.145 to 0.689
for moderated mediation (product = .077, Sobel SE = .052, 95% CI =
IDCR2 (path h) .209 (.083) .077 -0.022 to 0.440
PE2 -0.324 to -0.026).
Life Goal1 .302 (.206) .603 -0.839 to 1.444
Age .029 (.022) .695 -0.116 to 0.173
CLP (path c) -.237 (-.200) .841 -2.554 to 2.079 Table 5. Summary of Major Path Analyses of Effects on Problem Outcomes
IDCR2 (path i) -.612 (-.302) .003 -1.009 to -0.214
Outcomes Predictors Path coefficient P-value 95% CI
Life Goal2 PE2 (path n) .141 (.176) .010 0.033 to 0.249
CLP (path d) .001 (.004) .976 -0.083 to 0.086
PE1 -.078 (-.081) .771 -0.606 to 0.449
INT2 IDCR2 (path j) -.164 (-.245) < .001 -0.238 to -0.090
Age .301 (.285) .058 -0.010 to 0.613
Age .003 (.008) .911 -0.049 to 0.054
Note. Standardized coefficients shown in parentheses. CLP = Changing Lives Pro-
gram; E = Ethnicity; IDCR = Identity Conflict Resolution; PE = Personal Expressive- CLP (path e) .041 (.098) .443 -0.064 to 0.146
ness. EXT2 IDCR2 (path k) -.171 (-.241) < .001 -0.236 to -0.105
Age .030 (.083) .158 -0.012 to 0.073
Table 4. Mean Change in Continuous Outcomes
Note. Standardized path coefficients shown in parentheses. INT = Internalizing;
M CLP = Changing Lives Program; IDCR = Identity Conflict Resolution; PE =
Outcomes BL FU1 FU2 Personal Expressiveness; EXT = Externalizing.

CLP 3.668 3.683

IDCR
Ctrl 3.799 3.815 The relationship between IDCR and Life Goal was modeled
CLP 5.586 5.421 using the latent response formulation for categorical outcomes as
PE
Ctrl 5.459 5.140 implemented in Mplus 5.0 with the WLSMV estimator (Muthen &
CLP 1.756 1.601 1.645 Muthen, 1998-2007). The probit regression coefficients generated
INT
Ctrl 1.587 1.529 1.540 by the analysis were interpreted in terms of probability units (i.e.,
CLP 1.855 1.753 1.697 probits) and reflect the relationship between a one unit change in a
EXT
Ctrl 1.671 1.649 1.654 mediator and the probability that Life Goal = 1 (see Agresti, 2007). This
approach models the observed categorical response variable as the
Note. Means are not adjusted for covariates. IDCR = Identity Conflict Resolution;
CLP = Changing Lives Program; PE = Personal Expressiveness; INT = Internalizing; EXT realization of a latent continuous response variable. Specifically, the
= Externalizing observed ordinal value changes when a probit threshold is exceeded
on the latent continuous variable (Muthen & Muthen, 1998-2007). In
Hypothesized Mediation of Positive Outcomes the case of the dichotomous Life Goal variable, there is a single probit
threshold for the latent continuous variable such that Life Goal = 1
Hypothesized mediation (see Figure 2) was evaluated with the (Personally Expressive) when this threshold is exceeded and Life Goal
joint significance test (MacKinnon, Lockwood, Hoffman, West, = 0 (Non-personally Expressive) when the threshold is not exceeded.
29 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

Findings indicated that change in IDCR was significantly associated was .218. This reflects primarily that, among participants with a self-
with the probit of contemporaneous change in Life Goal (path i; satisfying life goal at BL, CLP participants’ responses at FU1 were
b = -.612, p < .01, 95% CI = -1.009 to -0.214). The indirect effect of CLP*E more likely to have the personally expressive property than expected
was significant (product = -.225, Sobel SE = .112, 95% CI = -0.579 to and that comparison control participants’ responses at FU1 were
-0.058). The relationship between PE and Life Goal was modeled in less likely to have the personally expressive property than expected
the same manner; change in PE was significantly associated with the (see Table 6). Of the 66 comparison control group participants with
probit of contemporaneous change in Life Goal (path n; b = .141, p = a self-satisfying life goal at BL, zero described life goals at FU1 with
.01, 95% CI = 0.033 to 0.249). the personally expressive property. Of the 98 CLP participants with a
self-satisfying life goal at BL, seven described life goals at FU1 with
Hypothesized Mediation of Problem Outcomes the personally expressive property.

Findings provided support for IDCR as a moderated mediator of Discussion

the relationship between CLP and changes in INT and EXT. Findings
indicated that BL to FU1 change in IDCR was significantly associated This paper contributes to the evaluation of the theoretical and
with contemporaneous change in INT (path j; b = -.164, p < .001, 95% empirical utility of integrating developmental and intervention
CI = -0.238 to -0.090) and EXT (path k; b = -.171, p < .001, 95% CI = science in the service of promoting positive youth development
-0.236 to -0.105). In addition, BL to FU1 change in IDCR was signifi- (PYD). Rooted in concepts and constructs drawn from applied
cantly associated with change in INT at FU2 (path l; b = -.205, p = developmental science (Lerner, Fisher, & Weinberg, 2000), the aims
.001, 95% CI = -0.328 to -0.083), and the indirect effect of CLP*E was of developmental intervention science include the development
significant (product = -.075, Sobel SE = .036, 95% CI = -0.180 to -0.018). and evaluation of evidence-based intervention strategies that
Change in IDCR was marginally significantly associated with change draw on developmental and intervention science in promoting
in EXT at FU2 (path m; b = -.093, p < .06, 95% CI = -0.187 to 0.002), and positive development and in reducing problem behaviors. Using
the indirect effect CLP*E was significant (product = -.034, Sobel SE = this approach, a study was conducted to evaluate the hypothesized
.022, 95% CI = -0.135 to -0.013). cascade effects of a PYD psychosocial intervention that specifically
targeted positive identity development in at-risk adolescents. The
investigation used an Outcome Mediation Cascade (OMC) model
Hypothesized Cascade Effects
that expanded the outcome mediation model described in the
With regard to hypothesized cascade effects, that is, intervention treatment literature (Silverman, Kurtines, Jaccard, & Pina, 2009) by
effects in untargeted problem outcomes mediated by change in tar- adding features of the developmental cascade model described in the
geted positive outcomes, analyses provided support for three plausi- developmental psychopathology literature (Cicchetti & Cohen, 2006;
ble causal chains that represent intervention cascade effects mediat- Masten et al., 2005). Intervention ↔ developmental cascades were
ed by qualitative change in life goals. The findings provided support conceptualized as intervention gains in positive outcomes that “spill
over” to generate reductions in problem outcomes. These analyses
for Life Goal as a mediator of the relationship between CLP and INT
contributed to advancing our knowledge (both theoretical and
as part of the pathways CLP → IDCR → Life Goal → INT (paths a, i, r),
empirical) of how interventions work.
CLP → PE → Life Goal → INT (paths b, n, r), and CLP → IDCR → PE →
Life Goal → INT (paths a, h, n, r). Specifically, change in Life Goal was
significantly related to change in INT (path r; b = -.153, p = .001, 95% Does It Work? Positive and Problem Outcomes
CI = -0.241 to -0.066).
With respect to hypothesized positive and problem outcomes, the
evaluation of the OMC model’s fit to the data yielded results that were
Qualitative Change in Life Goal: Fine-grained Analyses
promising at many different levels. The findings provided empirical
support for the hypothesis that participation in the intervention has
Due to the large proportion (75%) of participants with a self-
significant effects on both of the study’s positive outcome variables.
satisfying life goal at baseline, finer-grained analyses of the
First, the evaluation of the CLP indicated that participation in the
relationship between CLP and Life Goal at FU1 (path c) were pursued.
intervention was associated with a differential change consistent with
Limited information tests of conditional independence for CLP and
the hypothesized aims of the intervention in participants’ feelings
Life Goal at FU1 indicated that this relationship was not significant
of personal expressiveness, a measure of a core emotion-focused
for participants with a non-personally expressive life goal at BL, χ2(1)
component of positive identity, relative to comparison control group
= .487, p = .49; p = .56, Fisher’s exact test, 95% CI = 0.397 to 7.253, or
participants. Feelings of personal expressiveness (PE) reflect the
for those with a personally expressive life goal at BL (p = 1.00, Fisher’s
degree of fit between the pursuit of life goals and participants’ unique
exact test, 95% CI = 0.092 to 10.881).
potentials.
Second, the full information investigation of the OMC evaluation
Table 6. Frequencies of Change in Self Satisfying Life Goals model did not provide evidence of a direct intervention effect on
Life Goal at FU1 the probability of a personally expressive life goal, a marker of
Non-Personally Personally an increasingly integrated and complex self-structure. Limited
Condition Total information tests of conditional independence among participants
Expressive Expressive
Ctrl 66 0 66 with a personally expressive life goal at BL and among participants
Changing Lives Program 91 7 98 with a non-personally expressive life goal at BL also did not provide
157 7 164 evidence of a direct intervention effect. However, a finer-grained
limited information test of conditional independence, focusing
specifically on differential intervention effect on the presence/absence
However, CLP was significantly associated with Life Goal at FU1 of a personally expressive life goal at FU1 among participants with a
among participants with a self-satisfying life goal at BL, χ2(1) = 4.925, self-satisfying life goal at BL (a group that included the majority, 75%,
p = .03; p = .04, Fisher’s exact test, 95% CI = 1.001 to ∞). As indexed by a of the sample) did provide evidence for a direct differential effect
fourfold point correlation coefficient, the strength of the relationship between the intervention condition and the control. Thus, although
 Positive Youth Development 30

both of the intervention and comparison control conditions had zero and sameness. However, the conclusions that can be drawn from
personally expressive life goals at BL (because BL designations were the current findings are limited. Future research should consider
held constant, i.e., all self-satisfying life goals are non-personally pursuing this issue, for example, by investigating person ↔ context
expressive), seven of 98 CLP participants had a personally expressive relations that result in identity conflict or identity deficit, because
life goal at FU1 while the 66 comparison control group participants this line of inquiry has the potential to refine current knowledge of
still had zero personally expressive life goals at FU1. In addition, effective intentional person ↔ context intervention processes.
full information investigation of the OMC model indicated that an
increase in feelings of personal expressiveness was associated with Does Positive Cascading Change Reduce Problem Behaviors?
an increase in the probability of a personally expressive life goal. Intervention ↔ Developmental Cascades
With respect to intervention effects on problem outcomes, the
results did not provide evidence of direct intervention effects on This study operationalized the concept of an intervention ↔
untargeted internalizing or externalizing problem behaviors. The developmental cascade as the statistical mediation of the relationship
findings, however, did suggest several mediated and cascading between participation in the CLP intervention and differential change
causal pathways through which the intervention had an effect on in problem behavior outcomes (internalizing and externalizing) by
problem outcomes, described below. In addition, results suggested differential change in positive identity outcomes (feelings of personal
that internalizing and externalizing problem behaviors were linked expressiveness and the probability of a personally expressive life goal).
over time. A BL to FU1 increase in externalizing problem behaviors The analyses identified several causal pathways through which change
predicted a FU1 to FU2 increase in internalizing problem behaviors, in feelings of personal expressiveness and the presence/absence of a
and a BL to FU1 increase in internalizing problem behaviors predicted personally expressive life goal mediated the effects of the intervention
a FU1 to FU2 decrease in externalizing problem behaviors. on progressive change in internalizing problem behaviors.
Findings provided support for significant differential quantitative
How Does It Work? Mediation of Positive and Problem change in feelings of personal expressiveness and theoretically
Outcomes meaningful qualitative (categorical) change in participants’ most
important life goals as mediators of progressive change in internalizing
The pattern of results of this study’s mediation analyses yielded problem behaviors through several causal pathways. The qualitative
some interesting and suggestive potential answers to the theoretical change in participants’ life goals from non-personally expressive to
question of how the intervention works to promote change in personally expressive from BL to FU1 was associated with a decrease
positive and problem outcomes. Engagement in identity crisis was in internalizing problem behaviors from FU1 to FU2. Thus, findings
conceptualized as a mediator of change in both positive and problem provided support for change in feelings of personal expressiveness
outcomes among adolescents in the early stages of the identity process. and qualitative change in participants’ life goals as mediators of
A measure of the resolution of conflicts indicative of the identity crisis progressive change in problem behaviors via the following pathways:
was used to assess change in this hypothesized mediator; a decrease CLP → PE → Life Goal → INT (path b, n, r), CLP → IDCR → PE → Life Goal
in identity conflict resolution was conceptualized to reflect increased → INT (path a, h, n, r), and CLP → IDCR → Life Goal → INT (path a, i, r).
identity processing as a prelude to identity exploration. These findings make an initial contribution to closing the gap in our
The pattern of results provided support for the hypothesis that identity knowledge of the relationship between intervention change in positive
conflict resolution mediates intervention effects on positive and problem outcomes and intervention change in problem outcomes. They provided
outcomes. Specifically, results indicated that participation in the CLP preliminary evidence consistent with the hypothesis that in addition
was associated with a BL to FU1 decrease in identity conflict resolution to having an effect on targeted positive outcomes, PYD interventions
among African American participants, relative to the comparison control are likely to have progressive cascading effects on untargeted problem
group, but not among Hispanic participants. Results further indicated outcomes, that is, mediated intervention effects on problem outcomes
that BL to FU1 change in identity conflict resolution was associated with that operate through effects on positive outcomes. The pattern of
BL to FU1 change in feelings of personal expressiveness, the probability findings was also consistent with the hypothesis that the causal effects
of a personally expressive life goal, and internalizing and externalizing that are generated for both positive and problem outcomes are likely
problem behaviors. In addition, BL to FU1 change in identify conflict to follow complex pathways that interact in multifaceted ways with
resolution was associated with FU1 to FU2 change in internalizing and moderator and mediator variables rather than solely flowing directly
externalizing problem behaviors. Thus, the findings provided support for from intervention to positive outcomes.
identity conflict resolution as a moderated mediator of contemporaneous
changes in feelings of personal expressiveness (CLP → IDCR → PE; path Directions for Future Research: Refining
a, h), the probability of a personally expressive life goal (CLP → IDCR → the OMC Evaluation Model
Life Goal; path a, i), internalizing problem behaviors (CLP → IDCR → INT;
path a, j) and externalizing problem behaviors (CLP → IDCR → EXT; path The findings raise a number of questions regarding the relationship
a, k) and progressive changes in internalizing problem behaviors (CLP → between intervention change in positive outcomes and intervention
IDCR → INT; path a, l) and externalizing problem behaviors (CLP → IDCR change in problematic outcomes. For example, the finding that an
→ EXT; path a, m). increase in identity conflict predicted increases in problem behaviors
In the context of our working hypothesis that both PYD and and an increase in the probability of positive qualitative change in
prevention perspectives “work” by promoting positive change in life goals suggests a seemingly straightforward question: should PYD
person ↔ context relations (Schwartz et al., 2007), the finding that psychosocial interventions seek to promote identity exploration even
change in identity conflict resolution mediated all of this study’s if the identity conflict that prompts exploration predicts problem
positive and problem outcomes is particularly intriguing. The behaviors in the short-term? As noted, identity exploration (or crisis)
possibility that changes in psychosocial dimensions indicate changes has been described as the “work” of the identity process (Grotevant,
in the person ↔ context relations targeted by PYD and prevention 1987) and has been consistently found to co-occur with problematic
interventions is in line with Theokas et al.’s (2005) findings outcomes (Kidwell et al., 1995; Luyckx, Goossens, & Soenes, 2006;
highlighting the lack of pure discrimination between positive identity Schwartz, Mason, Pantin, & Szapocznik, 2008).
and connections with social contexts, as well as Erikson’s (1963) The findings from this study do not offer a satisfying answer to
description of identity as an inward and outward sense of continuity the question. However, the findings suggest that integrating the
31 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

analysis of quantitative data with the analysis of qualitative data behaviors may all reflect mediated moderation. On the other hand,
expanded the range of possible answers beyond the rather narrow the effect of the interaction between CLP and ethnicity on identity
conceptualization of identity exploration as either maladaptive or conflict resolution and the effects of identity conflict resolution on
adaptive. These findings instead invite speculation about the personal the probability of a personally expressive life goal, internalizing
meaning and significance of the identity choices faced by the problem behaviors, and externalizing problem behaviors could reflect
adolescent, or even of behaviors researchers describe as internalizing moderated mediation because in this scenario mediated moderation
and externalizing problems. and moderated mediation are analytically equivalent.
With respect to the analytic goals of an investigation of
Limitations and Future Directions hypothesized intervention ↔ developmental cascade effects, the
most pertinent research question appears to be whether or not the
Missing Data and Participant Attrition hypothesized cascading dynamic depends upon potential contextual
moderators (such as ethnicity), i.e., a question of moderated
At the time of this study, research regarding participant attrition in mediation. This study, however, did not examine moderation of
the CLP is unavailable, though attrition is known to be considerable. the effect of identity conflict resolution on the probability of a
The alternative high schools from which the sample was drawn do not personally expressive life goal, internalizing problem behaviors, and
record their own attrition data (Albrecht, 2007); however, Miami-Dade externalizing problem behaviors.
County Public Schools (M-DCPS) reports county-wide graduation rates Future research regarding moderation of intervention cascade
of approximately 59% and an official drop-out rate of 4.5%. On-time effects should draw on Edwards and Lambert’s (2007) model to
graduation rates are reported at 45.3% (Toppo, 2006). None of these investigate contextual moderation by examining second stage
figures account for students who are not officially withdrawn. Because moderation (moderation of the relationship between the mediator/
little specific information is available regarding these students, it is positive outcome and the outcome/problem behavior) in addition
possible that selection effects are potentially biasing results. Future to direct effect moderation (moderation of the relationship between
CLP research should begin to address this issue. participation in the intervention and the outcome/problem behavior)
The missing data found in the analyses may be due to the and first stage moderation (moderation of the relationship between
nature of the population the intervention targets, i.e., a population participation in the intervention and the mediator/positive outcome).
composed of multi-problem adolescents in alternative high
schools, many of whom are coming of age in disempowering What do Problem Behaviors Mean to the Adolescent?
contexts. This population faces risk factors that contribute to poor
attendance among other psychosocial problems that result in While refinement of the analyses has the potential to enhance
highly variable school/class attendance and high attrition rates. the present findings, perhaps more clarity may be gained by
However, participant attrition is one source of missing data among refining the research questions themselves. Among the potential
many potential sources of missing data. A better understanding of new research questions raised by this study is whether the use of
the role that attrition plays in generating observed missing data qualitative methods designed to capture the structural organization
patterns, as well as a more nuanced understanding of the factors of the subjective meaning and significance of internalizing and
underlying participant attrition, will allow future researchers to externalizing behaviors would contribute to the evolution of more
proactively model missing data mechanisms. A number of sources empirically supported theory informed models of both short and
(e.g., Collins, Schafer, & Kam, 2001; Graham, 2009; Little, 1995) long-term developmental change (identity).
have suggested modeling auxiliary variables useful in predicting Siegel and Scovill (2000) argue that problem behaviors are viewed
missing values, and Mplus (Muthen & Muthen, 1998-2007) includes most productively in their double aspect, or as a double symptom,
a feature for doing so. Future studies may need to model these with a positive aspect that reflects adolescents’ attempts to find
auxiliary variables in order to investigate the potential effects of an appropriate pathway to satisfy their developmental needs and
attrition in psychosocial interventions that target identity. accomplish developmental tasks successfully. While the negative
aspect of problem behaviors is well documented, the hypothesized
Mediated Moderation or Moderated Mediation? positive aspect has less often been the object of investigation. Similar
to Luyckx et al.’s (2008) identification of adaptive and maladaptive
One possible refinement to the investigation of the OMC evaluation identity exploration, this argument appears to reflect the recognition
model is to undertake a more comprehensive analysis of moderation of subjectively meaningful qualitative differences in problematic
of intervention cascade effects by relevant contextual factors. The behaviors. As Siegal and Scovill point out, different behaviors have
moderation analyses conducted by this study represent an initial different meanings for different adolescents. Moreover, “Adolescents,
attempt to identify relevant contextual moderators of intervention like adults do things for a reason. Behaviors that adults label
effects. The finding that a number of the cascade pathways identified ‘dysfunctional’ or ‘inappropriate’ may be the same behaviors that
by the analyses were moderated by ethnicity suggests the need help the adolescent survive in his environment” (p. 785). Siegal and
for further research regarding the nature of this moderation. One Scovill cite extant research that has suggested that the perceived
question for future research is whether this moderation is better benefits of these behaviors play a central role in the behaviors.
conceptualized as mediated moderation or as moderated mediation. An examination of theoretically meaningful qualitative
As a number of researchers (e.g., Edwards & Lambert, 2007; structural organizational change in the subjective meaning and
Iacobucci, 2008; Muller, Judd, & Yzerbcyt, 2005) have pointed out, significance of internalizing and externalizing behaviors requires
there is overlap in the strategies for evaluating mediated moderation the collection of new data. The instrument for collecting this type
and moderated mediation. Mediated moderation refers to an of data may resemble the qualitative extension of the Personally
interaction between an independent variable and a moderator variable Expressive Activities Questionnaire used in this study.
that has an effect on a mediator variable that, in turn, has an effect on
an outcome variable. Thus, the effect of the interaction between CLP Study Design
and ethnicity on identity conflict resolution and the effects of identity
conflict resolution on the probability of a personally expressive life Although the findings of this study indicate contemporaneous
goal, internalizing problem behaviors, and externalizing problem effects on positive identity via identity distress, the limitations of this
 Positive Youth Development 32

study’s design may have masked the possibility of even greater delayed to provide empirical evidence for intervention ↔ developmental
or lagged effects. Specifically, the development of a positive sense of cascades, specifically effects on positive identity outcomes that
self and identity is likely to continue well after participation in the “spill over” to generate change in problem behaviors. This cascading
CLP. Inclusion of future time points may reveal greater intervention dynamic was investigated by analyzing mediation of intervention
effects (i.e., direct effects) on the development of a positive sense of effects on problem behaviors by positive identity outcomes. The
self and identity. Additionally, a note of caution should be taken in intervention had significant effects on both positive identity outcomes,
regards to the generalizability of these findings. The current study and both positive identity outcomes mediated intervention effects on
was conducted at several alternative high schools in Miami that problem behaviors. In addition, change in the hypothesized mediator
unlike more traditional educational institutions, is predominantly a was a significant mediator of change in all positive and problem
school that focuses on at-risk populations. Moreover, it is important outcomes. The relationship between the intervention and change in
to note that the study took place with a diverse student population the mediator was moderated by ethnicity. Future research is needed
that may face cultural identity issues. For these adolescents who may to evaluate moderation of intervention ↔ developmental cascades.
be faced with deciding how to adapt their cultural identities and A second contribution of this study was to demonstrate an
maintain a personal identity, identity-focused interventions may be integrated analysis of quantitative dimensional change and
more effective and imperative. Future studies should also examine the qualitative structural organizational change using structural equation
effectiveness of the CLP in encouraging cultural identity development. modeling techniques. Previous studies (Arango et al., 2008; Kortsch
et al., 2008; Kurtines, Montgomery, Arango, & Kortsch, 2004) have
Statistical Power and Sample Size Considerations used small samples to investigate qualitative intervention change
in the narrative/linguistic expressions of the subjective experience
To determine an appropriate sample size, structural equation of self and identity. However, the present study was the first to
modeling requires that in addition to statistical power, issues of conduct a quantitative and qualitative evaluation model then use a
the stability of the covariance matrix and the use of asymptotic path model and to test the fit of the data. This approach allowed the
theory be taken into account. In terms of power, it is difficult to simultaneous modeling of multiple pathways of intervention change,
evaluate the power associated with specific path coefficients in including direct, mediated, and moderated effects, as well as cascade
complex SEM models because of the large number of assumptions effects. The analysis used a continuous latent response formulation
about population parameters that must be made. In the case as implemented in Mplus (Muthen & Muthen, 1998-2007) to model
of this study, with a sample of 212 used for the full information categorical outcomes within an SEM framework.
analyses, there was an effect on the statistical power available for A third contribution of the study was to test specific hypothesized
analyses of the full model, especially with respect to analyses of relationships (direct, mediated, moderated) among theoretically
dichotomous outcomes that require large sample sizes to capture meaningful qualitative structural organizational change in the
the relationship, if any, between the independent and dependent subjective meaning and significance of participants’ most important
variables, thereby increasing the likelihood of the occurrence life goals, quantitative dimensional positive and problematic
of type II errors, i.e., the inability to detect small effects. Limited outcomes, and participation in the CLP. Findings suggested (1) that
information strategies were used to further probe these relations; an increase in identity conflict resolution was associated with a
however, these strategies are problematic in that they do not use all decreased probability of a personally expressive life goal; (2) that an
available information from the hypothesized model. Future studies increase in feelings of personal expressiveness generated by life goal
may want to include larger sample sizes. pursuit was associated with an increased probability of a personally
expressive life goal; (3) that the change from a non-personally
expressive life goal to a personally expressive life goal was associated
Practical Implications: A Ready at Hand Intervention
with a progressive decrease in internalizing problem behaviors; and
(4) that participation in the CLP was associated with change from
Efforts to identify effective positive development strategies are
a non-personally expressive life goal to a personally expressive life
ongoing; however, the current study has important implications.
goal among participants with a self-satisfying life goal at baseline
Recent studies have highlighted the development of a positive sense
of self and identity as a potential intervention target, capable of evaluation.
facilitating mental health and reducing or preventing various negative The pattern of these findings likely reflects the complex and
psychosocial outcomes (i.e., internalizing symptoms, externalizing contingent nature of the systematic, successive, and adaptive
problems, and health risk behaviors). The CLP, demonstrates the need developmental processes involved in constructing and re-
for interventions to target identity as a main intervention goal, and constructing a positive sense of self and identity during initial stages
moreover works as an easily transportable and ready at hand program of formation. The pattern of these findings further highlights the need
that can be used at any institution for fostering positive development. for an expanded toolbox of the models and methods, particularly
The pattern of findings revealed in this study provides support for those capable of richly reflecting rather than reducing the life
the model explored in the investigation of the CLP. These findings course experiences of youth and rendering explicit and intelligible
suggest that identity focused psychosocial interventions can provide the theoretical meaning of these experiences. This study represents
opportunities for minority youth growing up in disempowering an initial attempt to contribute to the array of tools available in
community contexts to discover their personal potentials and create developmental intervention research and to identify potential
their own solutions to life challenges. Providing opportunities for directions for future use and further refinement of these tools.
marginalized youth to expand and enhance the aspects of their lives
that are meaningful to them can empower these youths to define for Conflict of Interest
themselves the direction of their own positive development (Eichas,
Montgomery, Meca, & Kurtines, 2017). The authors of this article declare no conflict of interest.

Conclusions Note

This study represents the first evaluation of hypothesized 1

Roberto L. Rinaldi received his PhD in developmental
intervention cascade effects. A main contribution of this study was psychology from Florida International University’s Developmental
33 K. Eichas et al. / Psychosocial Intervention (2018) 27(1) 22-34

Science Doctoral Program. His research interests are positive Eichas, K., Montgomery, M. J., Meca, A., & Kurtines, W. M. (2017).
youth development interventions with adolescents and young Empowering marginalized youth: A self-transformative intervention
for promoting positive youth development. Child Development, 88,
adults and the integration of qualitative and quantitative research 1115-1124. https://doi.org/10.1111/cdev.12866
methods. Erikson, E. H. (1963). Childhood and society. New York, NY: Norton.
Celeste Farr received her Ph.D., in health communication Erikson, E. H. (1968). Identity: Youth and crisis. New York, NY: Norton.
Erikson, E. H. (1985). The life cycle completed. New York, NY: Norton.
from Michigan State University. Her research interests are in the Erikson, E. H., & Erikson, K. T. (1957). The confirmation of the delinquent. In
improvement of health of minority communities. She has taught S. Schlein (Ed.), A way of looking at things: Selected papers from 1930
research methods at both the undergraduate and graduate levels to 1980 Erik H. Erikson. New York, NY: Norton.
Ferrer-Wreder, L., Montgomery, M. J., & Lorente, C. C. (2003). Identity
and frequently uses mixed methods in her work. promotion, adolescence. In T. Gullotta & M. Bloom (Series Ed.) & G. R.
Adams (Vol. Ed.), The encyclopedia of primary prevention and health
promotion: Adolescent volume (pp. 600-607). New York, NY: Kluwer
References Academic/Plenum.
Ferrer-Wreder, L., Palchuk, A., Poyrazli, S., Small, M. L., & Domitrovich, C. E.
Achenbach, T., & Edelbrock, C. (1981). Behavioral problems and (2008). Identity and adolescent adjustment. Identity: An International
competencies reported by parents of normal and disturbed children Journal of Theory and Research, 8, 95-105.
aged four through sixteen. Monographs of the Society for Research in Graham, J. W. (2009). Missing data analysis: Making it work in the real
Child Development, 46(1), 1-82. world. Annual Review of Psychology, 60, 549-576.
Adams, G. R., & Marshall, S. K. (1996). A developmental social psychology of Greenberger, E., Knerr, C., Knerr, B., and Brown, J. B. (1974). The Measurement
identity: Understanding the person-in-context. Journal of Adolescence, and Structure of Psychosocial Maturity (Report No. 170). Center for
19, 429-442. Social Organization of Schools. Johns Hopkins University. Baltimore.
Adams, G. R., Munro, B., Doherty-Poirer, M., Munro, G., Petersen, A. R., & Grotevant, H. (1987). Toward a process model of identity formation. Journal
Edwards, J. (2001). Diffuse-avoidant, normative, and informational of Adolescent Research, 2, 203-222.
identity styles: Using identity theory to predict maladjustment. Grunbaum, J., Kann, L., Kinchen, S. A., Ross, J. G., Gowda, V. R., Collins, J. L.,
Identity, 1, 307-320. & Kolbe, L. J. (1999). Youth risk behavior surveillance: National high
Agresti, A. (2007). An introduction to categorical data analysis (2nd Ed.). school risk behavior survey. Morbidity and Mortality Weekly Report,
Hoboken, NJ: John Wiley. 48(55), 1-52.
Albrecht, R. E. (2007). A model of self-transformative identity development Hernandez, L., Montgomery, M. J., & Kurtines, W. M. (2006). Identity
in troubled adolescent youth (unpublished doctoral dissertation). distress and adjustment problems in at-risk adolescents. Identity: An
Florida International University, Miami. International Journal of Theory and Research, 6, 27-33.
Arango, L. L., Kurtines, W. M., Montgomery, M. J., & Ritchie, R. (2008). A Iacobucci, D. (2008). Mediation analysis. Thousand Oaks, CA: Sage.
multi-stage longitudinal comparative design stage II evaluation of the Jaccard, J., & Turrisi, R. (2003). Interaction effects in multiple regression.
Changing Lives Program: The Life Course Interview (RDA-LCI). Journal Thousand Oaks, CA: Sage.
of Adolescent Research, 23, 310-341. Kelly, G. A. (1955). The psychology of personal constructs. New York, NY: Norton.
Archer, S. (Ed.) (1994). Interventions for adolescent identity development. Kerpelman, J. L., Pittman, J. F., & Lamke, L. K. (1997). Toward a microprocess
Thousand Oaks, CA: Sage. perspective on adolescent identity development: An identity control
Berman, A., Schwartz, S. J., Kurtines, W. M., & Berman, S. L. (2001). The theory approach. Journal of Adolescent Research, 12, 325-346.
process of exploration in identity formation: The role of style and Kidwell, J., Dunham, R. M., Bacho, R. A., Pastorino, E., & Portes, P. R.
competence. Journal of Adolescence, 24, 513-528. (1995). Adolescent identity exploration: A test of Erikson’s theory of
Berzonsky, M. D. (1989). Identity style: Conceptualization and transitional crisis. Adolescence, 30(120), 785-793.
measurement. Journal of Adolescent Research, 4, 267-281. Kortsch, G., Kurtines, W. M., & Montgomery, M. J. (2008). A multi-stage
Berzonsky, M. D. (1990). Self-construction over the lifespan: A process longitudinal comparative (MLC) design stage II evaluation of the Changing
perspective on identity formation. In G. J. Neimeyer & R. A. Neimeyer Lives Program (CLP): The Possible Selves Questionnaire—Qualitative
(Eds.), Advances in personal construct theory (vol. 1, pp. 155-186). Extensions (PSQ-QE). Journal of Adolescent Research, 23, 342-358.
Greenwich, CT: JAI. Kroger, J. (1997). Gender and identity: The intersection of structure,
Berzonsky, M. D. (1993). A constructivist view of identity formation: People content, and context. Sex Roles, 36, 747-770.
as postpositivist self-theorists. In J. Kroger (Ed.), Discussions on ego Kurtines, W. M., Montgomery, M. J., Arango, L. L., & Kortsch, G. (2004). Does
identity (pp. 169-203). Hillsdale, NJ: Lawrence Erlbaum. intervention change anything? New Directions in promoting positive
Berzonsky, M. D. (2004). Identity processing style, self-construction, and youth development. European Journal of Developmental Psychology,
personal epistemic assumptions: A social cognitive. European Journal 1, 383-397.
of Development Psychology, 1, 303-315. Kurtines, W. M., Montgomery, M. J., Eichas, K, Ritchie, R., Garcia, A.,
Berzonsky, M. D., & Kuk, L. S. (2000). Identity status, identity processing Albrecht, R., … Lorente, C. C. (2008). Promoting positive development
style, and the transition to university. Journal of Adolescent Research, in troubled youth: A Developmental Intervention Science outreach
15, 81-98. research approach. Identity: An International Journal of Theory and
Berzonsky, M. D., & Sullivan, C. (1992). Social–cognitive aspects of identity Research, 8, 125-138.
style: Need for cognition, experiential openness, and introspection. Lerner, R. M., Fisher, C. B., & Weinberg, R. A. (2000). Toward a science for
Journal of Adolescent Research, 7, 140-155. and of the people: Promoting civil society through the application of
Bollen, K., & Long, S. (1993). Testing structural equation models. Newbury Developmental Science. Child Development, 71, 11-20.
Park, CA: Sage. Little, R. J. A. (1995). Modeling the drop-out mechanism in repeated-
Bosma, H. A., & Kunnen, E. S. (2001). Determinants and mechanisms in ego measures studies. Journal of the American Statistical Association,
identity development: A review and synthesis. Developmental Review, 90(431), 1112-1121.
21, 39-66. Luyckx, K., Goossens, L., & Soenens, B. (2006). A developmental contextual
Breakwell, G. M. (1988). Strategies adopted when identity is threatened. perspective on identity construction in emerging adulthood: Change
Revue Internationale de Psychologie Sociale, 1, 189-203. dynamics in commitment formation and commitment evaluation.
Cicchetti, D., & Cohen, D. J. (2006). Developmental Psychopathology (2nd Developmental Psychology, 42, 366-380.
Ed., Vol. 3). New York, NY: John Wiley and Sons. Luyckx, K., Schwartz, S. J., Berzonsky, M. D., Soenens, B., Vansteenkiste,
Coatsworth, J. D., Sharp, E. H., Palen, L., Darling, N., Cumsille, P., & Marta, M., Smits, I., & Goossens, L. (2008). Capturing ruminative exploration:
E. (2005). Exploring adolescent self-defining leisure activities and Extending the four-dimensional model of identity formation in late
identity experiences across three countries. International Journal of adolescence. Journal of Research in Personality, 42, 58-82.
Behavioral Development, 29, 361-370. MacKinnon, D. P., Fritz, M. S., Williams, J., & Lockwood, C. M. (2007).
Collins, L. M., Schafer, J. L., & Kam, C. (2001). A comparison of inclusive and Distribution of the product confidence limits for the indirect effect:
restrictive strategies in modern missing data procedures. Psychological Program PRODCLIN. Behavior Research Methods, 39, 384-389.
Methods, 6, 330-351. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets,
Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. San Francisco, V. (2002). A comparison of methods to test mediation and other
CA: Jossey-Bass. intervening variable effects. Psychological Methods, 7, 83-104.
Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. Mainieri, T. (2006). The panel study of income dynamics child development
New York, NY: Basic Books. supplement: User guide for CDS-II. Ann Arbor, MI: The Institute for
Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation Social Research.
and mediation: A general analytical framework using moderated path Marcia, J. E. (1966). Development and validation of ego-identity status.
analysis. Psychological Methods, 12, 1-22. Journal of Personality and Social Psychology, 3, 551-558.
Eichas, K., Albrecht, R. E., Garcia, A. J., Ritchie, R. A., Varela, A., Garcia, A., Marcia, J. E. (1980). Identity in adolescence. In J. Adelson (Ed.), Handbook of
… Kurtines, W. M. (2010). Mediators of positive youth development adolescent psychology (pp. 159-187). New York, NY: Wiley.
intervention change: Promoting change in positive and problem Marcia, J. E. (1988). Common processes underlying ego identity, cognitive/
outcomes? Child and Youth Care Forum, 39, 21-237. moral development, and individuation. In D. K. Lapsley & F. C. Power
 Positive Youth Development 34

(Eds.), Self, ego, and identity: Integrative approaches (pp. 211-266). sexual behavior in Hispanic immigrant early adolescents. Identity,
New York, NY: Springer-Verlag. 8,107-124.
Maslow, A. (1968). The farther reaches of human nature. New York, NY: Penguin. Schwartz, S., J. Pantin, H., Coatsworth, J. D., & Szapocznik, J. (2007).
Masten, A., Roisman, G. I., Long, J. D., Burt, K. B., Obradovic, J. & Riley, J. Addressing the challenges and opportunities for today’s youth: Toward
R. (2005). Developmental cascades: Linking academic achievement, an integrative model and its implications for research and intervention.
externalizing and internalizing symptoms over 20 years. Developmental The Journal of Primary Prevention, 28, 117-144.
Psychology, 41, 733-746. Seligman, M., & Csikszentmihalyi, M. (2000). Positive psychology: An
McWhinnie, C., Abela, J. R. Z., Hilmy, N., & Ferrer, I. (2008). Positive youth introduction. American Psychologist, 55, 5-14.
development programs: An alternative approach to the prevention of Siegel, A. W., & Scovill, L. C. (2000). Problem behavior: The double symptom
depression in children and adolescents. In J. R. Z. Abela & B. L. Hankin of adolescence. Development & Psychopathology, 12, 763-793.
(Eds.), Handbook of depression in children and adolescents (pp. 354- Sharp, E. H., Coatsworth, J. D., Darling, N., Cumsille, P., & Ranieri, S.
373). New York, NY: Routledge. (2007). Gender differences in self-defining activities and identity
Montgomery, M. J. (2005). Psychosocial intimacy and identity: From early experiences of adolescents and emerging adults. Journal of
adolescence to emerging adulthood. Journal of Adolescent Research, 20, Adolescence, 30 , 251-269.
346-374. Silverman, W. K., Kurtines, W. M., Jaccard, J., & Pina, A. A. (2009).
Montgomery, M. J., Hernandez, L., & Ferrer-Wreder, L. (2008). Identity Directionality of change in youth anxiety treatment involving parents:
development and intervention studies: The right time for a marriage? An initial examination. Journal of Consulting and Clinical Psychology,
Identity: An International Journal of Theory and Research, 8, 173-182. 77, 474-485.
Montgomery, M. J., Kurtines, W. M., Ferrer-Wreder, L., Berman, S. L., Lorente, Soenens, B., Berzonsky, M. D., Vansteenkiste, M., Beyers, W., & Goossens,
C. C., Briones, E., … Eichas, K. (2008). A Developmental Intervention L. (2005). Identity styles and causality orientations: In search of the
Science (DIS) outreach research approach to promoting youth motivational underpinnings of the identity exploration process.
development: Theoretical, methodological, and meta-theoretical European Journal of Personality, 19, 427-442.
challenges. Journal of Adolescent Research, 23, 268-290. Theokas, C., Almerigi, J. B., Lerner, R. M., Dowling, E. M., Benson, P. L., Scales,
Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is P. C., & von Eye, A. (2005). Conceptualizing and modeling individual
mediated and mediation is moderated. Journal of Personality and asset components of thriving in early adolescence. Journal of Early
Social Psychology, 89, 852-863. Adolescence, 25, 113-143.
Muthen, L. K., & Muthen, B. O. (1998-2007). Mplus User’s Guide (5th ed.). Toppo, G. (2006). Big city schools struggle with graduation rates. USA
Los Angeles, CA: Muthen & Muthen. Today [Electronic version]. Retrieved from http://www.usatoday.com/
Palen, L., & Coatsworth, J. D. (2007). Activity-based identity experiences news/education/2006-06-20-dropout-rates_x.htm
and their relations to problem behavior and psychological well-being Waterman, A. S. (1984). Discovery or creation? Journal of Early Adolescence,
in adolescence. Journal of Adolescence, 30, 721-737. 4, 329-341.
Peterson, J. L., & Zill, N. (1986). Marital disruption, parent-child Waterman, A. S. (1990). Personal expressiveness: Philosophical and
relationships, and behavioral problems in children. Journal of Marriage psychological foundations. Journal of Mind and Behavior, 11, 47-74.
and the Family, 48, 295-307. Waterman, A. S. (1993). Finding something to do or someone to be: A
Rausch, J. R., Maxwell, S. E., & Kelly, K. (2003) Analytical methods for eudaimonist perspective on identity formation. In L. Kroger (ed.),
questions pertaining to randomized pretest, posttest, follow-up design. Discussions on Ego Identity (p. 147-167). Hillsdale, NJ: Lawrence
Journal of Clinical Child and Adolescent Psychology, 32, 467-86. Erlbaum Associates Inc.
Rinaldi, R. L., Eichas, K., Meca, A., Albrecht, R. E., Ritchie, R. A., Garcia, Waterman, A. S. (1995). Eudaimonic theory: Self-realization and the
A., & Kurtines, W. M. (2012). The development and evaluation of an collective good. In W. M. Kurtines & J. L. Gewirtz (Eds.), Moral
RDA qualitative extension for the Personally Expressive Activities development: An introduction (p. 255-278). Boston, MA: Allyn &
Questionnaire (PEAQ-QE). Identity: An International Journal of Theory Bacon.
and Research, 14, 312-331. Waterman, A. S. (1999). Identity, the identity statuses, and identity status
Rosenthal, D.A., Gurney, R. M., & Moore, S. M. (1981). From trust to intimacy: development: A contemporary statement. Developmental Review, 19,
A new inventory for examining Erikson’s stages of psychosocial 591-621.
development. Journal of Youth and Adolescence, 10, 525-537. Waterman, A. S. (2004). Finding someone to be: Studies on the role of
Schwartz, S. J. (2001). The evolution of Eriksonian and Neo-Eriksonian intrinsic motivation in identity formation. Identity: An International
identity theory and research: A review and integration. Identity: An Journal of Theory and Research, 4, 209-228.
International Journal of Theory and Research, 1, 7-58. Waterman, A. S. (2005). When effort is enjoyed: Two studies of intrinsic
Schwartz, S. J. (2002). In search of mechanisms of change in identity motivation for personally salient activities. Motivation and Emotion,
development: Integrating the constructivist and discovery perspectives 29, 165-188.
in identity. Identity: An International Journal of Theory and Research, Waterman, S., Schwartz, S., Goldbacher, E., Green, H., Miller, C., & Philip, S.
2, 317-339. (2003). Predicting the subjective experience of intrinsic motivation:
Schwartz, S. J. (2006). Predicting identity consolidation from self- The roles of self-determination, the balance of challenges and skills,
construction, eudaimonistic self-discovery, and agentic personality. and self-realization values. Personality and Social Psychology Bulletin,
Journal of Adolescence, 29, 277-293. 29, 10, 1-12.
Schwartz, S. J., Mason, C. A., Pantin, H., & Szapocznik, J. (2008). Effects Windle, M. (2003). Alcohol use among adolescents and young adults.
of family functioning and identity confusion on substance use and Alcohol Health and Research World, 27, 79-85.