Development and Validation of a Positive Youth Development Measure

Abstract

Objective:

Positive Youth Development (PYD) is a resilience-based framework that accentuates positive traits in youth. PYD constructs are characteristics that are hypothesized to promote healthy development. However, measures to assess PYD constructs are lacking.

Methods:

An instrument was created to assess PYD. One hundred and forty youth between the age of 7 and 18 (M = 11.6) completed the measure over five time points. Data were then analyzed using psychometric analysis techniques.

Results:

Findings suggest that the instrument is a reliable measure of PYD. The scale shows good reliability, with an overall Cronbach’s α of 92.

Conclusions:

The Bridge-PYD may be a useful tool for program evaluation and the assessment of theoretical constructs in the PYD models.

Keywords

children and youth positive youth development risk and resilience assessment tools

Antisocial and delinquent behaviors are serious social concerns in the United States. In 2009, approximately 1.5 million delinquency cases were handled in juvenile court, a 30% increase since 1985 (Office of Juvenile Justice and Delinquency Prevention [OJJDP], 2012). Of all the delinquency cases in 2009, 52% were held for youth under the age of 16; minority youth represent 36% of these cases. There have also been increases in the juvenile court in the period between 1985 and 2009, with drug offense cases rising by 117%, public order cases by 108%, and person offense cases by 99% (OJJDP, 2012).

Unfortunately, young people living in low-income communities are at increased risk for delinquency and other problem behaviors (Coley & Medeiros, 2007; Fleisher, 1996; Stewart, 2003; Vogel, Bradley, Raikes, Boller, & Shears, 2006). Low-income youth, compared to middle- and high-income youth, have higher rates of early delinquency leading to criminal behavior that continues into young adulthood (Mason et al., 2010). Neighborhood characteristics play an important role in adolescent delinquency and well-being. For example, neighborhood poverty has been linked to delinquent behavior problems among young males (Leventhal & Brooks-Gunn, 2000), and low neighborhood cohesion is associated with hyperactivity, anxiety, aggression, indirect aggression, and property offenses among young people (Kohen, Leventhal, Dahinten, & McIntosh, 2008).

With 63% of all juvenile violent crime occurring on weekdays during after-school hours (OJJDP, 2012) and the preponderance of crime transpiring in low-income communities, after-school programs based on Positive Youth Development (PYD) principles have been suggested as promising approaches for improving positive outcomes in at-risk children and youth residing in disadvantaged neighborhoods (Gottfredson, Cross, & Soule, 2007; Jenson, Alter, Nicotera, Anthony, & Forrest-Bank, 2013). This article describes the steps used to develop and test a new instrument aimed at assessing the 5 Cs of PYD using a sample of after-school participants residing in urban public housing neighborhoods.

Risk, Resilience, and PYD

Community and school-based interventions for at-risk children and youth are increasingly based on a social development perspective that aims to reduce risk, increase protection, and promote resilience (Catalano & Hawkins, 1996; Hawkins, Kosterman, Catalano, Hill, & Abbott, 2008; Jenson et al., 2013). There have been considerable advances in preventing and treating child and adolescent problem behavior in the past two decades (Jenson & Bender, 2014). However, practitioners have recently acknowledged limitations associated with risk and resilience frameworks, particularly attributed to the overemphasis on risk and deficits in young people. Therefore, interventions that augment resources to foster individual strengths and promote positive social bonds are being supported as successful adaptations to risk and resilience intervention models (Jenson et al., 2013),

The PYD approach has been instrumental in endorsing interventions for children and youth that are based on young people’s strengths and assets (Catalano, Berglund, Ryan, Lonczak, & Hawkins, 2004). The premise of PYD is that youth possess important resources within themselves. Thus, PYD intervention strategies seek to generate positive behavior by helping young people recognize and apply these resources to their daily lives (Lerner et al., 2005). PYD postulates that internal resources and assets bolster positive values, responsibility, and connectedness in children and youth. PYD approaches typically target individual and social characteristics or behaviors that are hypothesized to lead to healthy child and adolescent development (Catalano et al., 2004; Jenson et al., 2013).

PYD grew out of the grassroots efforts of youth workers who were interested in promoting programs and policies to support healthy child development (Lerner et al., 2005; Lerner, Dowling, & Anderson, 2003). Today, the on-the-ground proponents of PYD are social workers, psychologists, and other professionals who advocate for policy change and funding for interventions and community-based services aimed at promoting children’s strengths and increasing resilience. Child advocates such as Geoffrey Canada in Harlem and representatives from groups like the Annie E. Casey Foundation are but a few of the national spokespersons for PYD (Annie E. Casey Foundation, 2013; Tough, 2009).

PYD is an extension of applied developmental science, a strategy that builds on decades of research while extending easily from the ivory tower to communities and neighborhoods. Lerner, Fisher, and Weinberg (2000) explain that applied developmental science has roots in traditions that include (1) philosophy of science (Pepper, 1942); (2) developmental psychology (Cairns, 1998); (3) developmental theory (Lerner, 1998); (4) life span theory (Baltes, Staudinger, & Lindenberger, 1999); and (5) bioecological approaches (Bronfenbrenner, 1979, 1986). Borrowing from these influences, PYD suggests that changes in human behavior occur in a bidirectional context involving interactions between individuals and their environments. Further, the model stipulates that developing strengths and competencies as opposed to treating deficits is the most effective way to preventing problems and enhancing resilience (Lerner et al., 2005).

A framework developed by Lerner and colleagues is recognized by practitioners, policy makers, and researchers as one of the most influential and widely used PYD models (Lerner et al., 2005). In Lerner’s model, the key tenants of PYD are defined as 5 Cs of healthy development (Lerner et al., 2005). The “5 Cs” represent the constructs of caring/compassion, confidence, character, connection, and competence. Caring and compassion connotes a sense of sympathy and empathy for others, the ability to see outside one’s self. Confidence is an internal sense of overall positive self-worth and self-efficacy, one’s global self-regard as opposed to the domain-specific beliefs. Character reflects a respect for societal and cultural norms, possession of standards for correct behaviors, as sense of right and wrong and integrity. Connection refers to positive bonds with people and institutions that are reflected in bidirectional exchanges between the individual and his or her peers, family, school, and community in which both parties contribute to the relationship. Competence refers to having a positive view of one’s actions in specific areas, including social, academic, cognitive, and vocational settings. Social competence refers to school performance as shown, in part, by school grades, attendance, and test scores. Health competence involves using nutrition, exercise, and rest to keep oneself fit. Vocational competence involves work habits and explorations of career choices. Effective entrepreneurial skills may be one instance of vocational competence.

Elements of PYD are included in many community-based prevention and treatment programs for at-risk children and youth (Jenson et al., 2013). Catalano, Berglund, Ryan, Lonczak, and Hawkins (1998; 2004) conducted systematic reviews of PYD programs that were based on the model of 5Cs. Twenty-five programs met the criteria for an adequate evaluation design and operationally defined PYD components. The programs ranged from single interventions such as the Big Brothers/Big Sisters mentoring program (Tierney, Grossman, & Resch, 1995) to comprehensive programs such as the Adolescent Transitions Project (Andrews, Soberman, & Dishion, 1995) that utilizes youth and parent components as an overall strategy to promote healthy youth development. Findings revealed increases in commitment to school and improvements in interpersonal skills in 19 programs and significant reductions in substance use and high-risk sexual activity in 24 programs (Catalano et al., 2004).

Despite the positive outcomes associated with PYD-informed programs, there has been relatively little empirical work aimed at measuring the 5Cs. Although multiple indicators are available to assess risk and protective factors associated with delinquency and other problem behaviors (Arthur, Hawkins, Pollard, Catalano, & Baglioni, 2002; Bowen, 2006; Bowen & Richman, 1995), these tools typically predict future risk and seldom separate indicators of PYD constructs. The lack of a reliable measure of PYD constructs represents a significant gap in work being conducted to promote positive development in young people.

In the current study, we describe the developmental process used to create the Bridge-PYD, an assessment instrument used to evaluate the 5Cs of caring/compassion, confidence, character, connection, and competence. Psychometric properties for the instrument derived from the administration of the instrument to 140 children and youth living in public housing neighborhoods are presented. Recommendations for measuring PYD constructs with the Bridge-PYD are noted.

Method

Participants

An after-school program located in four urban public housing neighborhoods was the site used to administer and test the Bridge-PYD. The program site, the Bridge Project, serves at-risk children and youth in four public housing communities. Interventions at the sites include structured tutoring, mentoring, social and emotional skill development, and technology training. The program’s interventions are guided by the 5Cs of PYD developed by Lerner and colleagues (2005). A detailed description of the specified programs, objectives, and goals of the Bridge Project are available (see Anthony, Alter, & Jenson, 2009; Bender et al., 2011; Jenson et al., 2013).

Participants for the Bridge project were recruited and asked to consent voluntarily to complete the Bridge-PYD. One hundred and forty subjects completed the instrument over five different data points (N = 140). Age ranged from 7 to 18 years; the mean age was 11.6 years (standard deviation = 3.4). Gender was almost evenly split; 49% of the participants were female (n = 68) and 51% were male (n = 72). Of the participants, 40% (n = 56) were Hispanic, 30% (n = 42) were African American, 10% (n = 14) were Asian, 5% (n = 7) were Caucasian, and 15% (n = 21) of the participants self-reported belonging to an “other” racial or ethnic group.

Instrument Development

The Bridge-PYD was created using a scale development approach articulated by DeVellis (2007). This process included six stages: (a) a review of current risk and resilience measures; (b) development of items based on literature and expert review; (c) pilot testing and analysis of survey items using Item Response Theory (IRT); (d) factor analysis to determine which individual items may best represent each of the 5Cs subscales and analysis of scale reliabilities; (e) pre- and posttesting to determine whether the measure was sensitive to change; and (f) a confirmatory factor analysis (CFA) contributing to an overall model of PYD.

To begin the process, a literature review was conducted of all available scales for the constructs included in the 5Cs. To be considered in this review, measures had to be applicable for use with the after-school population of children and youth between the age of 7 and 18. Second, instruments needed to use self-report administrative strategies rather than be reliant on teacher or parent assessments. Third, to avoid problems with testing fatigue, completion time for selected measures was limited to 30 min. Finally, available measures are needed to specifically include items that were directly relevant to one or more of the 5Cs.

Several individual scales of risk and resilience among youth were found. There were two measures that addressed all 5Cs in a single scale. The first was by Lerner and colleagues (2005), but this scale was lengthy and had only been validated with children over the age of 12. The second was by Arnold, Nott, and Meinhold (2012). Similar to the Lerner instrument, this measure was both lengthy and not validated for use with children under age 11. A shorter version of the measure was developed by Arnold et al. (PYDI-S). Although the shorter version addressed the concern of test fatigue, validation studies of this scale appeared to measure constructs other than the 5Cs. Although the scales and surrounding literature provided useful guidance, none of these scales was deemed to be appropriate for use with children and adolescent populations participating in community-based programs like the Bridge Project.

Since there was no single measure of all of the 5Cs that appeared to be appropriate, a subsequent search was conducted to locate individual scales that measured each of the constructs independently. Despite an exhaustive search, only a scale for confidence was identified. This scale was the Piers Harris Self-Concept Scale (Piers, Harris, & Herzberg, 2007), which measures confidence through self-report, is validated for age 7–18, and can be completed in less than 15 min. There were other scales located that measured similar concepts, but they did not specifically indicate they measured one of the 5Cs constructs. These instruments included the School Success Profile (Bowen & Richman, 1995), the Elementary School Success Profile (Bowen, 2006), Communities that Care Survey (Arthur et al., 2002), the Morgan-Jinks Self-Efficacy Scale (Jinks & Morgan, 1999), the Carolina Child Checklist (McGowan, Nash, & Fraser, 2002), Early School Rating System (Caldwell & Planta, 1991), Social Skills Rating System (Gresham & Elliot, 1990), the School Situations Questionnaire (Altpeper & Breen, 1989), the Index of Empathy (Bryant, 1982), the Child Behavior Checklist (Achenbach, 1991), and the Behavior Assessment System for Children (Reynolds & Kamphaus, 2002). Although items from these measures assessed related constructs, our review of each of these instruments confirmed the lack of a singular, complete, simple measures of the 5Cs for children and youth. Therefore, the staff members and researchers of the Bridge Project initiated a process to develop and select individual items to assess the 5Cs of competence, confidence, connection, character, and caring/compassion in five scales, which could then be combined to create a single measure of PYD.

Item selection

As noted in Table 1, the team generated operational definitions of each construct in relation to intervention items and possible question items. These elements formed the basis of the Bridge-PYD instrument.

Table 1.

Working Definitions of Each of the Constructs and Potential Items or Questions.

Construct	Definition	PYD Informed Programming	Potential Items
Competence	Participants will demonstrate skills independently in the area of social, cognitive, academic health, and vocation	Tutoring, homework help, mentoring, reading/literacy, computers	I enjoy school I am good at math I like reading I get good grades
Confidence	Participants will develop self-worth, positive self-estem, and a sense of ownership	Tutoring, mentoring, social skills, art, yoga	Feeling proud Being good at certain activities Being popular or being liked by peers
Connection	Participants will develop meaningful relationships with Bridge and the larger community	Mentoring, tutoring, social skills groups, community service activities	I have adults I can talk to My friends support me I know how to make friends I like the staff at Bridge My teacher and I get along
Character	Participants will demonstrate leadership and civic engagement with a focus on positive behaviors and future goals	Tutoring, mentoring, social skills groups	Being able to define right from wrong; ability to overcome failure and plan for the future: I fight with other kids I get into trouble I follow the rules I can set goals If I do not do well at something, I will try again
Caring and Compassion	Participants will look outside themselves and develop the notion of human rights and respect for the environment and all living things	Mentoring, social skills	I help out neighbors I help out adults I help out other kids I stand up for my friends I make fun of other kids

Items were created and adapted to develop subscales for each of the 5Cs. This process involved searching existing instruments and the empirical literature. Input from the staff of Bridge project and from experts in the areas of high-risk youth and resilience was also used in the item development and selection process. Several items were developed and then loosely organized into one of the five subscales, one for each of the 5Cs. The complete instrument included 9 items to measure competence focused on how the participants felt about their school performance, academic abilities, and potential for school success; 10 items that indicated the youths’ level of confidence; 10 items to measure connection that evaluated how the participants viewed their relationships with positive adults, including parents and teachers; 11 items to measure character, including reports of positive behaviors, adherence to rules, and ability to set goals and plan for the future; and 7 items for caring and compassion that evaluated friendship, empathy, and caring about others. The items were scaled from 1 to 4, with 1 = strongly disagree, 2 = disagree, 3 = agree and 4 = strongly agree. Items for each scale were reviewed by after-school staff and substantive experts to ensure criterion validity as well as by the education program directors at the Bridge Project to ensure appropriate literacy levels.

Data Analyses

Several different data analysis techniques were used to evaluate the findings. First, the data were analyzed to reduce and condense individual scale items using IRT. IRT is a mathematical model that is based on the idea that the probability of answering an item in a certain way is the function of a latent trait or variable (Van der Linden & Hambleton, 1997). In this case, the individual items were analyzed to determine whether or not they “fit” the underlying 5Cs constructs. Once it was determined which items contributed to the latent PYD construct, the items were organized into subscales using exploratory factor analysis (EFA). This is a technique designed to explore the underlying structure, based on the patterns in the theory (DeVellis, 2007).

These techniques led to an overarching measure of PYD made up of the five subscales. This single measure was then administered and analyzed using CFA. By using a measure of “goodness of fit,” the CFA assesses whether the presenting data are a “fit” to the theory-based model (Kline, 2011). Finally, to determine whether the Bridge-PYD is sensitive to change over time, pre- and posttest measures were administered and analyzed using paired sample t-tests. The paired sample t-test is a straightforward and simple way to determine changes within a particular individual over two time points (Meyers, Gamst, & Guarino, 2006).

Findings

IRT

The first iteration of the Bridge-PYD was administered to 135 participants at the Bridge Project. Data were analyzed using Winsteps IRT (Bond & Fox, 2007). Items were considered to be a good fit if the overall scale and individual items had infit and outfit scores close to 1 with a SD close to 0. According to IRT, the closer the infit and outfit scores are to 1, the better the fit (Bond & Fox, 2007). Initial results of all 48 items together revealed an adequate fit, with an overall infit of 1.19 (zstd = .0) and outfit of 1.14 (zstd = .0).

Although the initial measures of overall fit were acceptable, there were several individual items that appeared to be item misfits, with infit or outfit scores higher than 3. The misfitting items fell into two separate and nonrelated constructs that included feelings about the staff and empathy items. Specifically, the first grouping of potentially misfitting items were there is someone I can talk to at the program; I feel safe at the program; and program staff helps me. It was determined that these questions were related to program satisfaction, rather than to a young person’s sense of connection, so these items were eliminated. There were also several poor fitting items that were originally created to measure the construct caring/compassion, but upon analysis, these items did not fit with the rest of the scale. Examples of misfitting items included I get upset if I see someone getting hurt, I’m happy to watch someone open presents even if I don’t get one myself, and If I see someone laughing, I laugh too. Discussion with staff about these questions revealed that participants likely experienced difficulty in understanding these items. Subsequently, misfitting items were removed and replaced with statements that focused on friendship, rather than empathy, to better reflect caring and compassion. Examples of these items include my friends and I are there for each other and it makes me sad to see a kid who has no one to play with.

Overall, 8 items were removed due to misfits, leaving 40 questions in the full instrument. Once the misfitting items were eliminated, all items had individual infit/outfit scores lower than 1.4. The final overall infit was 1.06 (zstd = .0) and outfit was 1.07 (zstd = .0). The person separation reliability was .90. This statistic is a measure that indicates how well the items can be used to distinguish individual responses from the latent variable. The closer the number is to 1, the better the person separation reliability fit (Bond & Fox, 2007).

Exploratory Factor Analyses and Scale Reliability

In the second step of data analyses, an EFA was conducted to determine whether individual items clustered on the PYD constructs of character, connection, caring and compassion, and competence. All 40 items were included, and since many of the underlying constructs are somewhat similar, a Varimax rotation was applied allowing items to correlate with each other (Meyers, Gamst, & Guarino, 2006). This resulted in five components with eigenvalues of two or higher that were extracted and rotated. There were three additional factors with eigenvalues greater than one, but since these only minimally contributed to the overall variance, they were eliminated and five factors were used. These five factors accounted for 64.6% of the total variance explained. Although there were five factors, not all of the items loaded as expected. Similar to the IRT findings, there was a difference in connection with adults and connection with other children. The items thought to be related to connection were separated. The items that focused on relationships with adults became the Connection subscale, while items addressing relationships with friends and other youth were used to create the Caring and Compassion subscale. Although some of the items cross-loaded onto more than one scale, the majority of the remaining items loaded on to the scales as expected. Table 2 provides results of the EFA.

Table 2.

Exploratory Factor Analysis and Factor Loadings.

	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5
Item	Character	Caring	Connection	Confidence	Competence
I could get the best grades if I tried	.390	−.024	.084	.301	.432
I think assignments are easy	.031	.077	−.010	.153	.615
I am one of the best students	.285	.141	.334	.433	.441
My teacher thinks I am smart	.116	.160	.453	.344	.465
I am good at reading	−.002	.176	.277	.429	.475
I am good at math	.275	.018	.288	−.200	.458
I am smart	.111	.468	.079	.269	.475
I work hard in school	.111	.157	.181	.072	.370
I am good at sports	.238	.172	.238	.688	−.089
I am a happy person	.197	.172	.626	.400	.111
I like who I am	.341	.372	.126	.707	.248
I feel left out	.344	−.150	.094	.679	.266
I am an important part of my class	.110	−.342	.048	.913	.045
People make fun of me	.634	−.424	.129	.600	.007
People say I have good ideas	.238	−.023	.238	.688	−.089
I am easy to get along with	.197	−.168	.600	.626	.111
I am good at music or art	.341	.337	.126	.707	.248
I feel proud of myself	.344	−.043	.094	.679	.266
Adults are nice to me	.135	.247	.784	−.052	.059
Adults tell me to do a good job	.314	.223	.517	.266	−.086
I can ask adults for help	.203	.288	.800	.089	−.075
I get along with my teacher	.392	−.093	.687	.289	.007
My teacher listens to me	.409	−.003	.561	.532	.010
If I don’t understand, my teacher will help me	.137	.132	.579	.312	.102
I fight with other kids	.400	.368	−.049	−.049	−.460
I am mean to other kids	.646	.611	.190	−.288	−.284
I help other kids	.684	.161	.367	.102	.076
I help adults	.648	.210	.096	.073	.203
I follow the rules	.529	.483	.198	−.020	.045
My friends follow the rules	.436	.427	.230	.336	.090
My friends listen to adults	.439	.095	.482	.011	−.041
It is important to plan for the future	.549	.210	.182	.138	.082
If I do not do well at something, I will try again	.413	.097	.368	−.311	.195
If I set a goal, I will complete it	.538	.038	.157	.098	.023
If I am upset, my friends help me	.238	.688	.238	.114	−.089
My friends listen to me	.197	.626	.400	.174	.111
My friends are happy for me	.341	.707	.126	.139	.248
My friends and I have fun together	.344	.679	.094	−.038	.266
My friends are on my side	.110	.913	.048	.134	.045
It makes me sad to see a kid with no one to play with	.600	.634	.129	.111	.007
If a kid does not have something, I will share	.238	.688	.238	.114	−.089

Using the subscales from the EFA, Cronbach’s α reliabilities were conducted for each of the subscales and the instrument as a whole (Myers et al., 2006). The final administration of the 5Cs subscales demonstrated a good reliability. The overall PYD, which included all five subscales, revealed a reliability of .92. Subscale reliabilities were within acceptable ranges, with an α of .80 for the competence scale; .75 for the confidence scale; .84 for character; .81 for connection, and .88 for the caring and compassion scale.

CFA

The final step of the scale development process was accomplished using CFA. To determine goodness of fit, five different fit indices were used. Each of the fit indices is sensitive to the different aspects of the data, including sample size and complexity of the model (Byrne, 1994). The goodness-of-fit index (GFI), the comparative fit index (CFI), and the Tucker–Lewis index (TLI) were used to determine how much better the presenting model was than the independence (null) model. For the GFI, CFI, and TLI, scores of .90 or higher are considered acceptable, and a score of .95 is considered to be a good fit (Byrne, 1994; Hu & Bentler, 1995; Kline, 2011). These particular indices can be sensitive to sample sizes of less than 200, so to account for the smaller sample size in this study, a root mean square error of approximation (RMSEA) was also evaluated to confirm goodness of fit. In the RMSEA, a score of less than .08 is considered acceptable, while a score of less than .05 is considered to be a good fit (Kline, 2011).

Before determining the fit of the PYD as a whole, item fit for each subscale was analyzed before evaluating the higher order factor (Marsh & Hocever, 1988). Each of the subscales demonstrated a good fit based on these fit indices. Table 3 provides results for the fit of the subscales. In addition to the overall fit, the data were analyzed to determine how much each of the items contributed to the underlying construct. To do this, the factor loadings were analyzed, with the first item in each scale fixed at 1. Each of the other items was then allowed to vary. This allows for comparison among the different items. Using this model, each of the factor loading estimates for the individual items were statistically significant at p < .001. Table 4 provides the factor loading results.

Table 3.

Confirmatory Factor Analysis Fit Statistics.

Scale	Chi-Square	GFI	TLI	CFI	RMSEA
Competence	18.9, p = .396	.967	.991	.994	.020
Confidence	7.35, p = .118	.982	.973	.993	.081
Connection	28.12, p = .255	.955	.981	.987	.037
Character	49.04, p = .016	.931	.933	.955	.071
Caring	11.9, p = .290	.974	.984	.992	.039
Positive Youth Development	2.24, p = .326	.996	.997	.999	.022

Note. CFI = comparative fit index; GFI = goodness of fit index; RMSEA = root mean square error of approximation; TLI = Tucker–Lewis index.

Table 4.

Confirmatory Factor Analysis of Individual Item and Subscale Factor Loadings.

Item	Factor Loadings
Item	Standardized	Unstandardized
Competence
I could get the best grades if I tried	.685	1
I think assignments are easy	.223	.58
I am one of the best students	.488	1.10
My teacher thinks I am smart	.308	.64
I am good at reading	.459	.75
I am good at math	.600	1.17
I am smart	.674	1.22
I work hard in school	.475	.95
Confidence
I am good at sports	.412	1
I am a happy person	.578	1.6
I like who I am	.527	1.4
I feel left out	.225	.836
I am an important part of my class	.654	1.8
People make fun of me	.062	.102
People say I have good ideas	.663	1.4
I am easy to get along with	.580	1.5
I am good at music or art	.370	.833
I feel proud of myself	.655	1.5
Connection
Adults are nice to me	.632	1
I can ask adults for help	.673	1.10
I get along with my teacher	.526	.77
My teacher listens to me	.755	.91
My teacher cares about my school work	.589	.89
If I don’t understand, my teacher will help me	.540	.86
I fight with other kids	.530	.703
Character
I am mean to other kids	.337	1
I help other kids	.479	.82
I help adults	.861	1.4
I follow the rules	.549	.88
My friends follow the rules	.678	1.14
My friends listen to adults	.448	.72
It is important to plan for the future	.404	.72
If I do not do well at something, I will try again	.484	.41
If I set a goal, I will complete it	.570	.80
Caring and compassion
If I am upset, my friends help me	.752	1
My friends listen to me	.759	.98
My friends are happy for me	.748	1.19
My friends and I have fun together	.487	1.03
My friends are on my side	.721	1.41
It makes me sad to see a kid with no one to play with	.421	.52
If a kid does not have something, like lunch or school supplies, I will share	.333	.62
Positive Youth Development
Competence	.416	1.91
Confidence	.695	3.56
Connection	.612	3.26
Character	.813	5.26
Caring and compassion	.573	2.22

A final test was conducted to determine whether this instrument represents a unidimensional measure of PYD (see Figure 1). Because of the small number of participants and correlations among individual item residuals, there was not enough statistical power to determine whether the scale represented a second-order factor of PYD. Instead, the subscales were treated as individual constructs that makeup the PYD model (Marsh & Hocever, 1988). To create these items for analysis, each of the individual items within the subscales was weighted using its respective factor loading scores. These scores were then added together to create a total score for each subscale. These weighted scores from each scale served as the factor, and the PYD construct served as the latent variable. Using this process, the results revealed acceptable fit statistics. This indicates that the five subscales come together to form a unidimensional measure of PYD. The fit statistics for the unidimensional PYD model are as follows: GFI = .996; TLI = .997; CFI = .999; and RMSEA = .022. The chi-square test was not significant (χ² = 2.24, p = .326), indicating no difference between the presenting data and the theoretical model.

Figure 1.

Measurement model for 5 Cs PYD-bridge.

Change Scores

The Bridge-PYD Instrument (see Appendix) was administered to participants during the fall (pretest) and spring (posttest) semesters of the school year. Differences between pretest and posttest scores on each PYD construct were analyzed to assess whether the newly developed scales were sensitive to change. In the first academic year of the study, 99 participants completed both pre- and posttests. Paired t-tests were conducted to analyze the change in scores on the Bridge-PYD. Participants demonstrated a significant and positive change in the overall PYD scores (t = −2.13, p < .05). In addition, the participant scores on the subscales showed statistically significant increases in confidence (t = −2.54, p < .05). The other scales did not show statistically significant change.

Changes in pretest and posttest scores for another 95 participants were analyzed during the second year of the study. Participants demonstrated a significant and positive change in the overall PYD scores (t = 2.63, p < .01). There were also statistically significant increases in the participants’ scores on the subscales of individual scales in competence (t = 2.15, p < .05), connection (t = 1.91, p < .05), and caring and compassion (t = 2.48, p < .05).

Discussion and Applications to Social Work

The purpose of this study was to develop an instrument to measure PYD constructs among at-risk children and youth involved in community-based programs. Findings suggest that the Bridge-PYD is a relatively easy-to-administer and valid instrument that is appropriate for a wide age range of children and youth. The Bridge-PYD measure adds to what is known about assessing levels of PYD constructs in several important ways.

CFA findings support the assertion that the 5Cs are important constructs in an overall theory of PYD. The results from the model fit indices support the underlying model. Specifically, each of the items contributes to a specific construct, and the five subscales converge to explain a higher order construct of PYD. Our findings also begin to shed light on the degree to which individual constructs (subscales) contribute to the broader concept of PYD. For example, in the final CFA analyses, character contributed most to the PYD model. Thus, it may be important for community-based programs like the Bridge Project to enhance and implement interventions that focus on building character in young people. Findings from the connection scale revealed that the strongest indicator of whether a youth felt connected was whether they could ask adults for help. Therefore, creating an atmosphere in which connections are fostered by encouraging youth to ask for help when needed may be particularly relevant in community-based interventions for at-risk youth.

Study findings also suggest that there may be benefits to using Bridge-PYD to inform program evaluation efforts. Results reveal that subscales are capable of demonstrating change in the 5Cs over time; all five subscales revealed improvements in PYD constructs between pretest and posttest. Reviews of PYD literature suggest that the lack of measures for the 5Cs is a serious limitation for community-based programs evaluation (Catalano et al., 2004). The Bridge-PYD should be tested in additional settings and its utility as a measure of PYD should be monitored.

Finally, the Bridge-PYD has many practical applications. The instrument was validated using a diverse age range of youth and is, therefore, applicable to students in elementary, middle, and high school. This characteristic offers great utility for practitioners and administrator as program staff and researchers do not have to use multiple surveys for different age ranges and the PYD-Bridge can be completed in approximately 15 min. Findings from this study reveal the promise of a single instrument to assess the 5Cs of PYD for at-risk children and youth in elementary, middle, and high school. Additional tests of the instrument should be conducted with larger samples to provide new evidence of the practical and theoretical importance of basing community programs on PYD constructs.

There are several limitations associated with the current analyses that should be noted. Structural equation modeling (SEM) and CFA approaches are intended to be the large-scale analyses. Each of the samples used for analyses had approximately 140 youth, which is acceptable for SEM but is still a small sample size for psychometric analyses (Kline, 2011; Myers et al., 2006). To support the current findings, additional analyses of the Bridge-PYD should be conducted with larger samples. In addition, CFA is a technique that is commonly referred to as being “sample specific.” Thus, there is no guarantee that our study findings will yield the same results with a new or different sample of at-risk children and youth. In order to increase confidence in results, the steps used in the current analysis should be replicated with different youth and in different service settings. Finally, participants in the current analyses were all recruited from the same community-based program. Children and youth with different demographic and risk profiles should be included in future studies of the Bridge-PYD.

Footnotes

Appendix

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Achenbach

T. M.

(1991). Manual for the child behavior checklist/4-18 and 1991 profiles. Burlington, VT: University of Vermont Department of Psychiatry.

Altpeper

T. S.

Breen

M. J.

(1989). The Home Situations Questionaire (HSQ) and the School Situations Questionaire (SSQ): Normative data and an evaluation of psychometric properties. Journal of Pyschoeducational Assessment, 7, 312–319.

Andrews

D. W.

Soberman

L. H.

Dishion

T. J.

(1995). Adolescent transitions program for high-risk teens and their parents: Toward a school-based intervention. Education and Treatment of Children, 18, 478–498.

Annie

Casey Foundation. (2013). Initiatives. Retrieved from http://www.aecf.org/MajorInitiatives.aspx

Anthony

E. K.

Alter

C. F.

Jenson

J. M.

(2009). Advances in preventing childhood and adolescent problem behavior. Research on Social Work Practice, 20, 701–713.

Arnold

M. E.

Nott

B. D.

Meinhold

J. L.

(2012). The positive youth development inventory. Corvalis, OR: Oregon State University.

Arthur

M. W.

Hawkins

J. D.

Pollard

J. A.

Catalano

R. F.

Baglioni

A. J.

(2002). Measuring risk and protective factors for substance use, delinquency and other adolescent problem behaviors: The communities that care youth survey. Evaluation Review, 26, 575–601.

Baltes

P. B.

Staudinger

U. M.

Lindenberger

(1999). Lifespan psychology: Theory and application to intellectual functioning. In Spence

J. T.

Darley

J. M.

Foss

D. J.

(Eds.), Annual review of psychology (Vol. 50, pp. 471–507). Palo Alto, CA: Annual Reviews.

Bender

Brisson

Jenson

Forrest-Bank

Lopez

Yoder

(2011). Challenges and strategies for conducting program-based research in after-school settings. Child and Adolescent Social Work Journal, 28, 319–334.

10.

Bond

T. G.

Fox

C. M.

(2007). Applying the rasch model: Fundamental measurement in the human sciences (2nd ed). New York, NY: Routledge.

11.

Bronfenbrenner

(1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.

12.

Bronfenbrenner

(1986). Ecology of the family as a context to human development: Research perspectives. Development Psychology, 22, 723–742.

13.

Bryant

B. K.

(1982). An index of empathy for children and adolescents. Child Development, 53, 413–425.

14.

Byrne

B. M.

(1994). Structural equation modeling with EQS and EQS/Windows. Thousand Oaks, CA: Sage.

15.

Bowen

G. L.

Richman

J. M.

(1995). School success profile. Chapel Hill, NC: The University of North Carolina at Chapel Hill, School of Social Work.

16.

Bowen

N. K.

(2006). Psychometric properties of the elementary school success profile for children. Social Work Research, 301, 51–63.

17.

Cairns

R. B.

(1998). The making of developmental psychology. In Lerner

R. M.

(Ed.), Damon

(Series Ed.), Handbook of child psychology: Vol 1. Theoretical models of human development (5th ed., pp. 419–448). New York, NY: John Wiley.

18.

Caldwell

Planta

(1991). A measure of young children’s problem and competence behaviors: The Early School Behavior Scale. Journal of Psychoeducational Assessment, 9, 32–44.

19.

Catalano

R. F.

Hawkins

J. D.

(1996). The social development model: A theory of antisocial behavior. In Hawkins

J. D.

(Ed.), Delinquency and crime. London, England: Cambridge University.

20.

Catalano

R. F.

Berglund

M. L.

Ryan

J. A.

Lonczak

H. S.

Hawkins

J.D.

(1998). Positive youth development in the United States: Research findings on evaluations of positive youth development programs. Washington, DC: US Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation and National Institute for Child Health and Human Development.

21.

Catalano

R. F.

Berglund

M. L.

Ryan

J. A. M.

Lonczak

H. S.

Hawkins

J. D.

(2004). Positive youth development in the United States: Research findings on evaluations of positive youth development programs. The Annals of the American Academy of Political and Social Science, 591, 98–124.

22.

Coley

R. L.

Medeiros

(2007). Reciprocal longitudinal relations between nonresidential father involvement and adolescent delinquency. Child Development, 78, 132–147.

23.

DeVellis

R. F.

(2007). Scale development: Theory and applications (2nd ed). Thousand Oaks, CA: Sage.

24.

Fleisher

B. M.

(1996). The effect of income on delinquency. The American Economic Review, 56, 118–137.

25.

Gottfredson

D. C.

Cross

Soule

D. A.

(2007). Distinguishing characteristics of effective and ineffective after-school programs to prevent delinquency and victimization. Criminology & Public Policy, 6, 289–318.

26.

Gresham

F. M.

Elliott

S. N.

(1990). The social skills rating system. Circle Pines, MN: American Guidance Service.

27.

Hawkins

J. D.

Kosterman

Catalano

R. F.

Hill

K. G.

Abbott

R. D.

(2008). Effects of social development intervention in childhood 15 years later. Pediatrics & Adolescent Medicine, 162, 1133–1141.

28.

L. T.

Bentler

P. M.

(1995). Evaluating model fit. In Hoyle

R. H.

(Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 76–99). Thousand Oaks, CA: Sage.

29.

Jenson

J. M.

Alter

C. F.

Nicotera

Anthony

E. K.

Forrest-Bank

S. S.

(2013). Risk, resilience, and positive youth development: Developing effective community programs for high-risk youth. Lessons from the Denver Bridge Project. (2013). New York, NY: Oxford University Press.

30.

Jenson

J. M.

Bender

K. A

.(2014). Preventing childhood and adolescent problem behavior. New York, NY: Oxford University Press.

31.

Jinks

Morgan

(1999). Children’s perceived academic self-efficacy: An Inventory Scale. Clearing House, 72, 224–231.

32.

Kline

R. B.

(2011). Principles and practice of structural equation modeling (3rd ed). New York, NY: Guilford Press.

33.

Kohen

D. E.

Leventhal

Dahinten

V. S.

McIntosh

C. N.

(2008). Neighborhood disadvantage: Pathways of effects for young children. Child Development, 79, 156–169.

34.

Lerner

R. M.

(1998). Theories of human development: Contemporary perspectives. In Lerner

R M.

(Ed.), Handbook of child psychology: Vol. 1. Theoretical models of human development. Hoboken, NJ: John Wiley.

35.

Lerner

R. M.

Dowling

E. M.

Anderson

P. M.

(2003). Positive youth development: Thriving as the basis of personhood and civil society. Applied Developmental Science, 7, 172–180.

36.

Lerner

R. M.

Fisher

C. B.

Weinberg

R. A.

(2000). Toward a science for and of the people: Promoting civil society through the application of developmental science. Child Development, 71, 11–20.

37.

Lerner

R. M.

Lerner

J. V.

Almerigi

J. B.

Theokas

Phelps

Gestsdottir

Simpson

(2005). Positive youth development, participation in community youth development programs and community contributions of fifth grade adolescents: Findings from the first wave of the 4H study of positive youth development. Journal of Early Adolescence, 25, 17–71.

38.

Leventhal

Brooks-Gunn

(2000). The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126, 309–337.

39.

Marsh

H. W.

Hocevar

(1988). A new, more powerful approach to multitrait-multimethod analyses: Application of second-order confirmatory factor analysis. Journal of Applied Psychology, 73, 107–117

40.

Mason

W. A.

Hitch

J. E.

Kosterman

McCarthy

C. A.

Herrenkohl

T. I.

Hawkins

J. D.

(2010). Growth in adolescent delinquency and alcohol use in relation to young adult crime, alcohol use disorders, and risky sex: A comparison of youth from low-versus middle-income backgrounds. Journal of Child Psychology & Psychiatry, 51, 377–385.

41.

Mcgowan

M. J.

Nash

J. K.

Fraser

M. W.

(2002). The Carolina checklist of risk and protective factors for aggression. Research in Social Work Practice, 12, 253–261.

42.

Meyers

L. S.

Gamst

Guarino

A. J.

(2006). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage.

43.

Office of Juvenile Justice and Delinquency Prevention: U. S. Department of Justice. (2012). Statistical briefing book. Retrieved May 29, 2012, from www.ojjdp.gov/ojstatbb

44.

Piers

E. V.

Harris

D. B.

Herzberg

D. S.

(2007). Piers-Harris 2: Piers-Harris children’s self-concept scale (2nd ed). Los Angeles, CA: Western Psychological Services.

45.

Pepper

S. C.

(1942). World hypotheses. Berkeley, CA: University of California Press.

46.

Reynolds

C. R.

Kamphaus

R. W.

(2002). The clinician’s guide to the Behavior Assessment System for Children (BASC) . New York, NY: Guilford Press.

47.

Stewart

S. D.

(2003). Nonresident parenting and adolescent adjustment: The quality of nonresident father-child interaction. Journal of Family Issues, 24, 217–244.

48.

Tierney

J. P.

Grossman

J. B.

Resch

N. L.

(1995). Making a difference: An impact study of big brothers big sisters. Philadelphia, PA: Public/Private Ventures.

49.

Tough

(2009). Whatever it takes: Geoffrey Canada’s quest to change Harlem and America. New York, NY: Houghton Mifflin Company.

50.

Van der Linden

W. J.

Hambleton

R. K.

(Eds.). (1997). Handbook of modern item response theory. New York, NY: Springer.

51.

Vogel

C. A.

Bradley

R. H.

Raikes

H. H.

Boller

Shears

J. K.

(2006). Relation between father connectedness and child outcomes. Parenting: Science and Practice, 2, 189–209.