What Drives Positive Youth Development? Assessing Intentional Self-Regulation as a Central Adolescent Asset

Abstract

The “5 Cs of PYD Model” is an influential theory involving the role of Intentional Self-Regulation (ISR) in positive youth development (PYD). The model, which was developed with U.S. populations, has recently been used in other cultural contexts and, as such, the measurement invariance of the ISR-PYD model across cultural contexts must be assessed. This study examined whether the ISR-PYD model is conceptually similar across two cultures, the U.S. and Iceland. Participants were 2,473 U.S. and 539 Icelandic adolescents (65% female, 15.00 years; 46% female, 14.30 years, respectively). Results suggested weak invariance and partial strong invariance for PYD and partial weak invariance across groups and time for ISR. Bifactor modeling showed that ISR was more strongly associated with global PYD than individual Cs. These findings demonstrate construct validity of the ISR-PYD model, suggesting that ISR is a central asset in general healthy functioning among youth in different Western cultural contexts.

Keywords

Intentional self-regulation positive youth development adolescence cultural contexts validity

Theorists and researchers have traditionally portrayed adolescence as a period of stress and psychological upheaval, leading to an explicit emphasis on interventions that prevent adolescents from exhibiting maladaptive developmental outcomes (Lerner, Lerner, Bowers, & Geldhof, 2015). However, in the last two decades, practitioners and researchers have increasingly reframed their work to understand young people’s strengths and healthy development (i.e., Positive Youth Development [PYD]; Catalano, Berglund, Ryan, Lonczak, & Hawkins, 2004). Although many key models of PYD have arisen, the Five Cs Model of PYD promoted by Lerner, Lerner, and colleagues has received considerable empirical support and influenced research and practice associated with the second decade of life (Lerner, Lerner, & Benson, 2011; Lerner et al., 2015).

One key hypothesis derived from the Five Cs Model is that young people’s mindful, volitional control over their actions (i.e., intentional self-regulation; ISR) represents a necessary precursor to adaptive development across contexts and domains of functioning (Gestsdottir & Lerner, 2008). This hypothesis has received substantial empirical support, with goal-directed ISR emerging as a central strength that promotes PYD (e.g., Gestsdottir & Lerner, 2007, 2008; Lerner et al., 2005, 2015). However, although the PYD model has been used to study youth in various cultures outside of the U.S., including Europe (see e.g., Erentaitė & Raižienė, 2015; Holsen, Geldhof, Larsen, & Aardal, 2016; Wen, Su, Li, & Lin, 2015), evidence supporting the association between ISR and PYD has mostly come from a single study of youth living in the U.S, the 4-H Study of Positive Youth Development. Considering the growing interest in PYD and self-regulation in relation to adolescence, it is important that the generalizability of the ISR-PYD model across groups and contexts be assessed. As Little (1997) explained: “Cross-cultural research can be a powerful method for understanding the nature and generalizability of psychological phenomena” (p. 52).

In this article we advocate for ISR as a universal predictor of adaptive functioning among youth in all cultures. Specifically, we predicted that ISR is equivalently associated with PYD across same-aged youth in two cultural settings in North American and Europe; the U.S. and Iceland. Furthermore, previous studies have typically treated PYD as either a single factor or as five unique constructs (“Cs”). In the present study, we predicted that ISR is a central asset in thriving and, as such, should be equally or more strongly related to a global PYD index than to the residual variance associated with the individual Cs after controlling for global PYD. The findings of this study may contribute to the field’s emerging understanding of ISR and its role in the healthy development of all young people.

The Five Cs Model of Positive Youth Development

In contrast to the deficit-oriented perspectives that dominated the field of adolescent development through much of the 20th century, models of PYD emphasize that all young people have strengths and can thrive by aligning those strengths with the opportunities and supports afforded by their contexts (Lerner et al., 2015). There are several influential models that have been used to frame descriptive and explanatory research that have helped create a better understanding about the positive development of youth. Some of the best known conceptions of youth’s strengths and resilience include William Damon’s (2008) study of purpose, the work of Peter Benson on individual and ecological developmental assets (Benson, 2003), Ann Masten’s (2004) work on resilience, Margaret Beale Spencer’s (2006) phenomenological variant of ecological systems theory (PVEST), and Catalano and Hawkins’ (1996) research on protective factors (for further reviews see Catalano et al., 2004; Lerner et al., 2015).

However, the Five Cs Model is one of the most comprehensive approaches to PYD and has been elaborated in the context of decade-long longitudinal study of PYD conducted by Lerner, Lerner, and colleagues, the 4-H Study of PYD. This research seeks to identify the individual and ecological relations that may promote thriving and diminish risk and problem behaviors. According to the model, PYD is comprised of five Cs – Competence, Confidence, Character, Connection, and Caring (Lerner et al., 2005). Competence represents a young person’s ability to competently navigate the obstacles and possible goals presented by his or her unique environment. Confidence reflects the self-efficacy and well-being that accompany knowing one can competently navigate the context and can intentionally impact one’s own developmental pathway. Character represents the degree that a young person behaves according to internalized moral standards and, therefore, reflects an intersection of prior experiences, the lessons a young person has learned from them, and the degree that previously learned lessons match the social expectations of a young person’s current environment. Caring indicates how well an individual displays a developmentally and contextually appropriate level of concern for the well-being of others. Connection indexes the diversity and strength of the relationships between an individual and the interpersonal and institutional features of his or her context.

The Five Cs Model emphasizes that youth strengths must be aligned with assets in their contexts to promote PYD. As such, positive development means that individuals must structure their own development in ways that promote optimal person-context alignment. ISR has thus become a central asset believed to promote PYD, a hypothesis consistent with the literature examining the importance of agency and self-regulation for adaptive human functioning across the life span (e.g., Baltes, 1997; Brandtstädter, 2006).

Self-Regulation in Adolescence

Literature on the role of self-regulation in adolescent functioning has grown immensely over the last two decades, but suffers from a paucity of theories that describe the nature and role of self-regulation during adolescence (McClelland, Geldhof, Cameron, & Wanless, 2015). This gap has led to numerous conceptions and measures of adolescent self-regulation, oftentimes focusing on a narrow range of self-regulation skills that may not fully capture all aspects of self-regulation. For example, advanced cognitive and psychological processes that allow mindful, future-oriented self-regulation, such as ISR, are expected to grow and acquire increased significance in adolescence, as it allows the young person to set, prioritize, and obtain long-term goals and promote self-development (e.g., Brandtstädter, 2006; Demetriou, 2000; Gestsdottir & Lerner, 2008; Gollwitzer, 1999). Nevertheless, adolescent research frequently uses measures of emotion regulation or executive functioning to index self-regulation that, although related to the pursuit of long-term goals, primarily describe how people react to situations that arise at a given moment (e.g., Duckworth & Seligman, 2005; Lopes, Salovey, Côté, & Beers, 2005).

In the absence of theories of self-regulation intended to be applied specifically to the adolescent period, researchers within the 4-H Study of PYD assessed ISR using the Selection, Optimization, and Compensation (SOC) model. Devised by Baltes, Baltes, Freund, and colleagues (Baltes, 1997; Freund & Baltes, 2002; Wiese, Freund, & Baltes, 2000), the SOC model was developed with adult populations but is intended to be applicable across the life span. It describes three self-regulatory strategies: Selection pertains to how people select, prioritize, and commit themselves to specific goals. Optimization strategies help a person identify and utilize resources that assist goal pursuit and attainment. Compensation similarly involves goal-directed actions, but aims to avoid or minimize the negative impact of losing goal-relevant means. SOC attributes allow people to formulate future-directed plans and to evaluate and modify courses of action directed toward a given outcome, thereby giving each person the power to shape his or her own developmental trajectory (Baltes, 1997; Bandura, 2006; Brandtstädter, 2006).

In an attempt to clarify the role of ISR during adolescence, several studies based on the 4-H Study data set have assessed the association between SOC and positive development. These studies have demonstrated that the three SOC processes do not fully differentiate into separate factors during adolescence but, instead, are comprised of a single factor containing a subset of items drawn from the three subscales (see e.g., Geldhof et al., 2015; Gestsdottir & Lerner, 2007). Furthermore, a study of youth from four Western cultures provided broad support for the adolescent-specific factor structure, showing that a single SOC composite fit well among 14-year-old youth in the U.S., Canada, Germany, and Iceland. These findings suggest that SOC processes are not fully developed in adolescence and may differ qualitatively from adult self-regulation (for further discussion, see Gestsdottir et al., 2015). Nevertheless, several publications based on the 4-H data have demonstrated that SOC has positive links to indicators of healthy functioning and negative associations with problematic behaviors across adolescence in the U.S. (e.g., Gestsdottir, Bowers, von Eye, Napolitano, & Lerner, 2010; Gestsdottir & Lerner, 2007). However, the 4-H study’s findings need to be validated across cultural contexts to speak the universality of the 5Cs model of PYD and to the role of ISR in healthy adolescent development.

The Current Study

The present study examined two questions relating to the role of ISR in the healthy development of youth. First, we tested measurement invariance to examine whether the ISR-PYD model is conceptually similar across the two cultures under investigation, the U.S. and Iceland. Previous studies have typically treated PYD as either a single factor or as five separate Cs. Treating PYD as a second-order construct assumes that participants have set levels of PYD and that PYD itself causes participants to display specific levels of each C. We used bifactor modeling to better understand the structure of PYD. A bifactor model suggests that each item has two sources of true-score variance: global PYD and C-specific variance (e.g., Reise, 2012). Therefore, we used bifactor confirmatory factor analysis to test our first hypothesis, which predicted measurement invariance for PYD and ISR across cultures and across two waves of data collection. Factorial invariance across the two samples would suggest that youth interpreted the constructs similarly in both cultures (Wu, Li, & Zumbo, 2007).

Theories of self-regulation and PYD suggest that ISR should promote general healthy development in all contexts and regardless of domain of functioning. To examine the criterion validity of the ISR-PYD model, we examined associations between ISR and PYD across the two cultural contexts. In a second hypothesis, we expected that ISR, as indexed by SOC, was a predictor of overall PYD approximately one year later in both samples. Thus, we hypothesized that ISR would relate to global PYD rather than to the residual Cs. Together, support of the hypotheses set forth in this study could provide evidence for the internal and external validity of SOC and PYD and, as well, for the generalizability of Lerner and colleagues’ hypothesis that ISR is a central adolescent strength across cultures.

Method

Participants

United States

We analyzed data from 2,473 adolescents (64% girls) who participated in Waves 5 and 6 of the 4-H Study (approximately corresponding to U.S. Grades 9 and 10), referred to as Time 1 and Time 2 in this study. Data for Time 1 were collected in 2006 and for Time 2 in 2007. Participants were recruited by contacting parents and children through their schools and extracurricular programs. The mean age of participants was 15.00 years (SD = 0.74) at Time 1. With respect to race/ethnicity, the sample was 6.05% Asian; 1.91% Black; 7.27% Latino; 71.82% White; and 14.4% other (including Native American, Multiethnic/multiracial, or “other”). Additional details on the 4-H Study sample are provided by Bowers and colleagues (2015).

Iceland

We compared data from the 4-H Study to data from 539 youth when they were in Grades 9 and 10 in Iceland, referred to as Time 1 and Time 2. Data for Time 1 were collected in 2012 and data for Time 2 in 2013. Mean age of the participants at Time 1 was 14.3, SD = 0.3; 46% girls. A total of 94% of the sample was of Icelandic origin (meaning that both parents were of Icelandic heritage). To obtain the data, we randomly selected 20 out of the 54 medium to large-sized schools in the SW region of the country, of which fifteen agreed to participate. To ensure diverse responses at the school level, two classrooms in each school were selected at random. These 30 classrooms had a total of 625 students for which 561 parents (90%) gave written consent to participate. Among students whose parents provided consent, 539 youth participated (96%).

Procedure

U.S.

A detailed protocol was used to ensure that data collection was administered uniformly and to ensure the return of all study materials (see Bowers et al., 2015, for a discussion of the method of the 4-H Study). For the present study, youth completed a survey online unless they requested a paper survey. The two waves of data collection, referred to as Time 1 and Time 2, were approximately 12 months apart.

Iceland

Participants completed a paper-and-pencil survey during a school visit by trained research staff during a single 40-minute session. Standardized instructions were used to ensure that data collection was implemented uniformly. Students who were absent during the school visit were contacted with the help of school staff by e-mail, mail, or phone, and asked to complete and return the survey by mail. Data collections at Time 1 and Time 2 were approximately 12 months apart.

Measures

Positive Youth Development

We operationalized PYD using the short-form (SF) measure of the Five Cs of PYD (PYD-SF) discussed by Geldhof and colleagues (2014). The PYD-SF measure includes 34 items that capture the five subcomponents of PYD; Competence using six items, two items representing Academic Competence, Social Competence, and Physical Competence, respectively. All Competence items were scored on a four-point Likert-type scale. The PYD-SF similarly measures Confidence using six items, two representing Self Worth, Positive Identity, and Physical Appearance, respectively. A sample item is: “Some kids feel like they are just as smart as other kids their age but other kids aren’t so sure and wonder if they are as smart”.

The Self-Worth and Physical Appearance items followed the same response format as the Confidence items, but the Positive Identity items were scored using a five-point Likert-type scale. A sample item of the Confidence scale is: “Some kids are happy with themselves most of the time but other kids are often not happy with themselves”. Character in the PYD-SF includes eight items, two representing Social Conscience, Values Diversity, Conduct Behavior, and Personal Values. All items except Conduct Behavior were scored on a five-point Likert-type scale. The Conduct Behavior items used a four-point scale. A sample item from the Character scale is: (How important is ... the following to you in your life) “Helping to make the world a better place to live in”. Items representing Caring and Connection were scaled on a five-point Likert-type scale. Caring was indicated by six items and a sample item is: “When I see someone being taken advantage of, I want to help them”, whereas the Connection scale contained eight items that represented connection to participants’ families, neighborhoods, schools, and peers, including this sample item: “I have lots of good conversations with my parents”.

The English version of the PYD-SF measure was translated into Icelandic by a certified translator and a researcher fluent in English and Icelandic and then back-translated to English. The two English versions were compared to each other, and to the Icelandic version, and minor changes were made to the wording of the Icelandic translation.

Because the PYF-SF may be best modeled using a bifactor model, common estimates of internal consistency (e.g., Cronbach’s α) are less useful for understanding item reliability than are standardized factor loadings. Internal consistency indicates the percentage of factor-relevant variance relative to total variance in a unit-weighted composite of all items, often assuming a single underlying dimension. The PYD-SF is not strictly unidimensional, making the interpretation of any internal consistency estimate difficult at best. Instead, we provide standardized factor loadings as part of our results. Note that each loading indicates only part of each item’s reliable variance and that loadings onto both PYD and the residual Cs must be jointly considered when assessing item quality.

Self-Regulation

All participants completed a 9-item subset of Freund and Baltes’ (2002) Selection, Optimization, and Compensation (SOC) questionnaire that has been shown to be unidimensional in both the U.S. and Iceland (Gestsdottir et al., 2015). In the original version of the measure, which was used with the U.S. sample, each item consisted of two statements, one describing behavior reflecting selection, optimization, or compensation and the other describing a non-SOC related behavior. A sample item from the Selection scale is: “I always focus on the one most important goal at a given time”; a sample item from the optimization scale is: “When I start something that is important to me but has little chance at success, I make a particular effort” and a sample item from the compensation scale is: “When something doesn’t work as well as usual, I look at how others do it” (all sample items include statements that reflect the use of a SOC strategy). Participants were asked to decide which of the statements is most similar to the way they would behave and, as such, all items were scored as a binary representation of whether or not a participant selected the SOC-related statement.

In Iceland, the English version was translated into Icelandic by a researcher fluent in English and Icelandic. This version was compared to the English version, and minor changes were made to wording. The 4-H Study was the first study to use the SOC measure to index ISR in early adolescence and to explore its relation to PYD in a U.S. sample. As the study progressed, findings suggested that SOC may be captured better, or at least equally well, by using only SOC-relevant statements and a Likert-type scale instead of a binary-scale, a finding that has was replicated with an Icelandic sample (Geldhof, Little, & Hawley, 2012; Stefansson, Gestsdottir, & Skulason, 2014). As such, an Icelandic version of the SOC measure included only SOC-relevant statements and asked to respond to each statement on a five-point Likert-type scale. Somewhat similar to the case of the PYD-SF, the categorical nature of the SOC items in the U.S. sample makes any estimate of those items’ reliability difficult to interpret. We instead include standardized factor loadings in the results to represent overall item reliability.

Analyses

We examined the longitudinal properties of the SOC scale between the U.S. and Icelandic samples using longitudinal confirmatory factor analysis (CFA). All models that included the SOC scale were estimated using robust weighted least squares in Mplus, with the SOC items treated as categorical in the US sample only. All other items were treated as continuous.

We modeled SOC as a unidimensional factor but, as noted above, we modelled PYD using a variant of the bifactor structure. Each PYD item loaded simultaneously on a group- and time-specific PYD factor as well as loading onto one of five group- and time-specific residual C constructs (representing Competence, Confidence, Character, Caring, and Connection after controlling for overall PYD). We hypothesized that the residual Cs represented meaningful constructs and were therefore allowed to freely covary with each other but not with the global PYD construct. In addition, each indicator at Time 1 freely covaried with the parallel indicator at Time 2 to account for shared specific variances.

We established factorial invariance separately for SOC and PYD, using phantom indicators (Geldhof, Pornprasertmanit, Schoemann, & Little, 2013) to establish configural invariance of the SOC measure across groups. We tested configural and weak invariance for each scale. For SOC, weak invariance involved equating the factor loadings for the binary indicators in the U.S. sample with the factor loadings for the Likert-type indicators in the Icelandic sample. We then examined strong factorial invariance of the PYD items, but strong invariance was not examined for SOC due to scaling differences as well as the limitations associated with binary data described by Millsap and Yun-Tien (2004). Strong invariance of SOC across time was tested for the Icelandic sample, however. As suggested by Cheung and Rensvold (2002; see also Little, 2013), we used a change in CFI criterion (i.e., Δ CFI > –0.01) when examining each level of invariance.

We then fit a two-group longitudinal CFA that included both PYD and SOC. Longitudinal CFA was chosen over a longitudinal structural equation model because latent regressions would violate the assumption of orthogonality between PYD and the residual Cs. This CFA allowed us to examine whether and how the PYD constructs correlated with SOC.

Results

Factorial Invariance of PYD and SOC across Cultures

The initial longitudinal CFA model for the PYD measure suggested one character item should load onto the Competence factor in the US sample. The resulting model produced acceptable fit (χ² (df = 4034) = 7489.87, p < 0.001; RMSEA = 0.02, [90% C.I. 0.02, 0.03]; CFI = 0.95, TLI = 0.94). Further analysis then supported weak invariance(Δ CFI = 0.005) but only partial strong invariance (Δ CFI = 0.006). For the partial strong invariance model, intercepts for two indicators of the Competence scale, one indicator from the Confidence scale, and one indicator from the Caring scale differed across groups but not across time within either group. Thus, aside from the items with non-invariant intercepts, there was no evidence that participants systematically scored higher or lower on individual items depending on their group membership. Standardized factor loadings from the partial strong invariance model are provided in Table 1. Raw-metric intercepts from this model are providedin Table 2.

Table 1

Standardized Factor Loadings from the Partial Strong Invariance PYD Model

	U.S.				Iceland
	Time 1		Time 2		Time 1		Time 2
	PYD	Cs	PYD	Cs	PYD	Cs	PYD	Cs
Competence:
Academic 1	0.25	0.41	0.23	0.41	0.28	0.42	0.30	0.46
Academic 2	0.17	0.55	0.15	0.56	0.19	0.59	0.19	0.60
Social 1	0.45	0.30	0.41	0.30	0.49	0.31	0.51	0.32
Social 2	0.62	0.16	0.60	0.17	0.67	0.16	0.70	0.17
Physical 1	0.61	0.12	0.59	0.12	0.72	0.13	0.74	0.13
Physical 2	0.58	0.10	0.57	0.11	0.73	0.12	0.73	0.12
Conduct Behavior 2	–0.19	–0.24
Confidence:
Self-Worth 1	0.49	0.45	0.48	0.48	0.64	0.47	0.65	0.47
Self-Worth 2	0.51	0.45	0.53	0.52	0.67	0.48	0.69	0.49
Positive Identity 1	0.36	0.48	0.34	0.49	0.45	0.47	0.46	0.48
Positive Identity 2	0.24	0.46	0.23	0.47	0.33	0.50	0.30	0.46
Appearance 1	0.63	0.25	0.64	0.28	0.81	0.26	0.83	0.26
Appearance 2	0.65	0.30	0.67	0.33	0.83	0.31	0.85	0.31
Character:
Social Conscience 1	0.59	0.38	0.59	0.38	0.75	0.36	0.73	0.37
Social Conscience 2	0.62	0.34	0.61	0.34	0.77	0.32	0.78	0.34
Values Diversity 1	0.32	0.23	0.31	0.22	0.37	0.20	0.37	0.21
Values Diversity 2	0.41	0.27	0.39	0.25	0.47	0.23	0.48	0.25
Conduct Behavior 1	–0.02	0.50	–0.02	0.56	–0.03	0.63	–0.03	0.63
Conduct Behavior 2	0.03	0.47	0.03	0.52	0.04	0.53	0.04	0.53
Personal Values 1	0.49	0.41	0.50	0.41	0.60	0.37	0.63	0.41
Personal Values 2	0.44	0.42	0.45	0.42	0.54	0.39	0.57	0.42
Caring:
Item 1	0.59	0.32	0.60	0.33	0.64	0.43	0.70	0.39
Item 2	0.58	0.27	0.56	0.27	0.68	0.40	0.75	0.36
Item 3	0.59	0.24	0.56	0.23	0.55	0.28	0.66	0.28
Item 4	0.82	0.29	0.80	0.29	0.81	0.36	0.88	0.32
Item 5	0.74	0.26	0.75	0.26	0.74	0.32	0.82	0.29
Item 6	0.80	0.28	0.79	0.29	0.78	0.35	0.85	0.31
Connection:
Family 1	0.36	0.43	0.30	0.41	0.41	0.46	0.42	0.48
Family 2	0.38	0.47	0.32	0.45	0.44	0.51	0.44	0.53
Neighborhood 1	0.55	0.39	0.50	0.40	0.57	0.38	0.59	0.39
Neighborhood 2	0.57	0.37	0.50	0.37	0.56	0.35	0.60	0.38
School 1	0.35	0.49	0.32	0.51	0.36	0.47	0.38	0.50
School 2	0.33	0.45	0.29	0.45	0.34	0.43	0.35	0.46
Peers 1	0.22	0.33	0.20	0.35	0.27	0.39	0.29	0.43
Peers 2	0.28	0.35	0.26	0.37	0.33	0.39	0.35	0.42

Table 2

Intercepts from the Partial Strong Invariance PYD Model

	U.S.		Iceland
	Time 1	Time 2	Time 1	Time 2
Competence:
Academic 1	2.97	2.97	2.97	2.97
Academic 2	3.18	3.18	3.18	3.18
Social 1	3.10	3.10	3.10	3.10
Social 2	2.64	2.64	2.64	2.64
Physical 1	2.27	2.27	2.78	2.78
Physical 2	2.21	2.21	2.67	2.67
Confidence:
Self-Worth 1	3.22	3.22	2.81	2.81
Self-Worth 2	3.16	3.16	3.16	3.16
Positive Identity 1	4.25	4.25	4.25	4.25
Positive Identity 2	4.15	4.15	4.15	4.15
Appearance 1	2.80	2.80	2.80	2.80
Appearance 2	2.82	2.82	2.82	2.82
Character:
Social Conscience 1	3.95	3.95	3.95	3.95
Social Conscience 2	3.75	3.75	3.75	3.75
Values Diversity 1	3.21	3.21	3.21	3.21
Values Diversity 2	3.71	3.71	3.71	3.71
Conduct Behavior 1	3.13	3.13	3.13	3.13
Conduct Behavior 2	2.85	2.85	2.85	2.85
Personal Values 1	4.13	4.13	4.13	4.13
Personal Values 2	4.08	4.08	4.08	4.08
Caring:
Item 1	4.38	4.38	4.38	4.38
Item 2	3.86	3.86	4.33	4.33
Item 3	3.89	3.89	3.89	3.89
Item 4	4.38	4.38	4.38	4.38
Item 5	4.38	4.38	4.38	4.38
Item 6	4.37	4.37	4.37	4.37
Connection:
Family 1	4.12	4.12	4.12	4.12
Family 2	4.14	4.14	4.14	4.14
Neighborhood 1	3.37	3.37	3.37	3.37
Neighborhood 2	3.30	3.30	3.30	3.30
School 1	3.80	3.80	3.80	3.80
School 2	3.87	3.87	3.87	3.87
Peers 1	4.33	4.33	4.33	4.33
Peers 2	4.35	4.35	4.35	4.35

For the SOC scale, the configural invariance model suggested several correlated residuals within and across time (four total) in the U.S. group. After freeing these covariances, the configural invariance model displayed good fit (χ² (df = 240) = 594.94, p < 0.001; RMSEA = 0.03, [90% C.I. 0.03, 0.03]; CFI = 0.92, TLI = 0.89). The low TLI likely reflects an excessive degrees of freedom due to the presence of phantom indicators, as a the model fit well when fit only to the U.S. (χ² (df = 123) = 253.920, p < 0.001; RMSEA = 0.02, [90% C.I. 0.02, 0.03]; CFI = 0.96, TLI = 0.95) and only to the Icelandic (χ² (df = 125) = 390.771, p < 0.001; RMSEA = 0.06, [90% C.I. 0.06, 0.07]; CFI = 0.93, TLI = 0.92) datasets. Partial weak invariance was then supported across groups and time (Δ CFI = 0.001), with two factor loadings differing in the U.S. sample at Time 2 only. Standardized factor loadings from the partial weak invariance model are provided in Table 3. A final model then supported invariance of the SOC intercepts across time (i.e., strong invariance) among the Icelandic sample only (Δ CFI < 0.001).

Table 3

Standardized Factor Loadings from the Partial Weak Invariance SOC Model

	U.S.		Iceland
	Time 1	Time 2	Time 1	Time 2
SOC3	0.59	0.63	0.65	0.70
SOC5	0.57	0.31	0.65	0.71
SOC7	0.60	0.20	0.68	0.76
SOC8	0.65	0.69	0.74	0.83
SOC10	0.62	0.66	0.74	0.83
SOC13	0.53	0.57	0.61	0.68
SOC15	0.52	0.56	0.60	0.70
SOC17	0.44	0.48	0.39	0.48
SOC18	0.33	0.36	0.28	0.33

Assessing the SOC-PYD Model across Cultures

The full CFA model was quite large, and, to facilitate model convergence, we simplified the PYD model by equating latent means, variance, and covariances across groups and time in the partial strong invariance PYD model (results not shown; the likelihood ratio tests associated with these constraints were nonsignificant at α= 0.001, a Type-I error rate chosen to accommodate the very high power of our model). These equality constraints were then imposed on the full CFA that modeled PYD and SOC in both groups at both time points. The resulting model showed acceptable fit (χ² (df = 6970) = 9372.82, p < 0.001; RMSEA = 0.02, [90% C.I. 0.01, 0.02]; CFI = 1.00, TLI = 0.92).

Latent criterion correlations from this final model are presented in Table 4. As this table shows, the residual Competence factor consistently correlated with ISR, whereas the other criterion correlations between SOC and other residual Cs were more sporadic. The relatively weak correlations between SOC and the residual Cs were overshadowed by the substantially stronger correlations between ISR and overall PYD, however. Supporting the hypothesis derived from the PYD perspective, ISR was much more strongly linked to overall adaptive functioning than to the residual Cs.

Table 4

Criterion Correlations between PYD and SOC

	U.S. SOC		Icelandic SOC
	Time 1	Time 2	Time 1	Time 2
Time 1
Competence	0.22^**	0.10	0.27^***	0.29^***
Confidence	–0.03	0.04	0.05	0.10
Character	0.15^**	0.13	0.21^***	0.15^**
Caring	0.01	0.07	0.18^**	0.05
Connection	0.08	0.00	0.07	0.01
PYD	0.63^***	0.49^***	0.64^***	0.54^***
Time 2
Competence	0.38^***	0.15^***	0.22^***	0.30^***
Confidence	0.11	0.08	0.00	0.09
Character	–0.01	0.11^*	0.10	0.08
Caring	0.04	0.06	0.12^*	0.03
Connection	0.00	0.03	0.06	0.04
PYD	0.63^***	0.62^***	0.55^***	0.61^***

To sum, the weak invariance of the PYD measure, and the partial weak invariance across groups on the SOC measure, suggests that the strength of the relation between each item on the measure and the underlying construct was similar across groups and that ISR and PYD may be measured in a similar way across the two cultures. The analyses also show that ISR has strong links to overall PYD, as compared to individual Cc, suggesting that ISR is central to general healthy functioning regardless of culture.

Discussion

Two changes within the study of adolescent development have been especially salient during the last twenty years. First, the dominant view of adolescence as a period of risk taking, conflict, and difficulties has been rejected (Benson, 2003; Lerner et al., 2015). An alternative conception of adolescence, the PYD perspective, focuses on the strengths believed to exist among all youth and on how youth thriving can be promoted (Benson, 2003; Larson, 2000; Lerner et al., 2015). Second, there has been a growth in research that has identified self-regulation as a central skill in adolescent functioning. In particular, research has supported the importance of ISR. In the last few years, there have been studies that have demonstrated the validity of various models of positive development across cultural contexts, such as the 5Cs model and developmental assets model (see e.g., Conway, Heary, & Hogan, 2015; Holsen et al., 2016; Scales, Roehlkepartain, Wallace, Inselman, Stephenson, & Rodriguez, 2015). However, to our knowledge, there has not been research that has assessed the validity of the role of ISR in the positive role of youth across cultural contexts. The current study aimed to assess the construct comparability of ISR-PYD model across cultures by testing two hypotheses regarding the validity of ISR as a central asset promoting general healthy functioning among youth living in different cultural contexts.

Using longitudinal CFA, our results revealed weak invariance of the PYD measure, suggesting that the factor loading pattern of the PYD model, or the strength of the relation between each item and its underlying construct, was invariant across cultures. For the same measure, our results suggest only partial strong invariance, and some differences were observed on the intercepts for four indicators between youth in the U.S. and Iceland. Nevertheless, taken together, these results suggest that PYD has reasonably equivalent measurement properties across cultures and time, and that cross-cultural comparisons using the PYD measure are meaningful. Similarly, analysis of the SOC measure supported partial weak invariance across groups. These findings provide empirical evidence for our first hypothesis, that the ISR-PYD model demonstrates construct validity when used with youth in these two Western cultures, the U.S. and Iceland.

In regard to our second hypothesis, our findings suggest a strong link between ISR and a global measure of PYD taken a year after the ISR measure in both cultures. Weaker associations were observed between ISR and the residual Cs. Several theorists (e.g., Bandura, 2006; Demetriou, 2000) have explained how ISR skills depend partly on self-evaluations, such as one’s beliefs about one’s self-efficacy or self-esteem, which are constructs that overlap with Competence in the Five Cs model of PYD (Lerner et al., 2015). The more consistent relation to Competence, as compared to the other Cs, may be explained by the fact that Competence may have a conceptual overlap with ISR. Interestingly, this stronger relation of ISR to Competence as compared to the other Cs has not been apparent in previous research that has treated the Five Cs as separate constructs, rather than using a bifactor approach (see e.g., Gestsdottir et al., 2010). However, our findings also clearly show that the relations to individual Cs are overshadowed by the much stronger links between ISR and overall adaptive functioning, supporting our second hypothesis that ISR is a central asset in the general healthy functioning of youth in different cultural contexts.

Limitations and Future Directions

The current study provided evidence to support both of our hypotheses, suggesting that the structure of the Five Cs of PYD can be replicated in samples outside the U.S. and that ISR promotes thriving (PYD) across cultures. This finding paves the way for further use of the ISR-PYD model with populations in Europe and beyond. The generalizability of these findings is constrained by the limited nature of the data analyzed, however. First, data from the U.S. were only compared to data from one other Western country. Additional replication is needed across several cultural contexts, for instance to determine whether the structure of PYD replicates in other European countries, and/or among youth living in more collectivistic cultures. Indeed, the present findings may not even replicate in subpopulations of youth living within the United States or Iceland, for instance those living in severely disadvantaged communities. Finally, although previous studies based on the current data have typically revealed similar findings for boys and girls in relation to PYD and ISR (see Lerner et al., 2015, for a review), an examination of possible gender differences in relation to the constructs under investigation is warranted.

Our findings are also limited to one specific model of PYD (the Five Cs Model) and one operationalization of ISR (the SOC model). Both PYD and ISR can be operationalized in different ways (e.g., see Geldhof, Little, & Colombo, 2010; Lerner et al., 2015), and the hypothesis that ISR promotes thriving across cultures must be replicated using multiple measures of each. Furthermore, the hypothesis that ISR promotes thriving is not constrained to adolescence, and we suggest further replication of these findings across from diverse contexts and age periods.

The present findings also have technical limitations, the most obvious of which is the fact that the SOC questionnaire was administered using a binary response format in the U.S. but a Likert-type response format in Iceland. Previous research has found few differences between Likert-type and forced-choice binary measures of SOC (e.g., Geldhof et al., 2015), and the convergent findings across the two cultures suggests that measurement differences were not a problem in the present analyses. Despite this similarity, equivalent response metrics would have been helpful for establishing the equivalence of the associations between SOC and PYD. Future research might therefore attempt to measure the associations between SOC and the Five Cs using multiple measures of each construct.

Despite a rapid growth of research that has identified self-regulation as an important aspect of adolescent functioning, our understanding of this topic has suffered from a lack of an empirically supported theory of the nature and role of ISR in adolescence. Our findings provide evidence of the construct comparability of SOC and PYD across cultures and support the generalizability of ISR as a central adolescent strength that can promote healthy functioning.

References

Baltes

P. B.

(1997). On the incomplete architecture of human ontogeny - Selection, optimization, and compensation as foundation of developmental theory. American Psychologist, 52, 366–380. doi: 10.1037//0003-066x.52.4.366

Bandura

(2006). Toward a psychology of human agency. Perspectives on Psychological Science, 1, 164–180. doi: 10.1111/j.1745-6916.2006.00011.x

Benson

P. L.

(2003). Developmental assets and asset-building community: Conceptual and empirical foundations. In Lerner

R. M.

& Benson

P. L.

(Eds.), Developmental assets and asset building communities: Implications for research, policy, and practice, (pp. 19–43). Norwell, MA: Kluwer. doi: 10.1007/978-1-4615-0091-9_2

Bowers

E. P.

, Geldhof

G. J.

, Johnson

S. K.

, Hilliard

L. J.

, Hershberg

R. M.

, Lerner

J. V.

, & Lerner

R. M.

(Eds.). (2015). Promoting positive youth development: Lessons learned from the 4-H Study,. New York: Springer. doi: 10.1007/978-3-319-17166-1

Brandstädter

(2006). Action perspectives on human development. In Damon

& Lerner

R. M.

(Eds.), Handbook of child psychology: Theoretical models of human development, (6. ed., Vol. 1, pp. 516–568). New York, NY: Wiley. doi: 10.1002/9780470147658.chpsy0110

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. Annals of the American Academy of Political and Social Science, 591, 98–124. doi: 10.1177/0002716203260102

Catalano

R. F.

, & Hawkins

J. D.

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

J. D.

(Ed.), Delinquency and crime: Current theories, (pp. 149–197). New York, NY: Cambridge University Press.

Cheung

G. W.

, & Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255. doi: 10.1207/s15328007sem0902_5

Conway

R. J.

, Heary

, & Hogan

M. J.

(2015). An evaluation of the measurement properties of the Five Cs Model of Positive Youth Development. Frontiers in Psychology, 6, 1941. doi: 10.3389/fpsyg.2015.01941

10.

Damon

(2008). The path to purpose: How young people find their calling in life,. New York: The Free Press.

11.

Demetriou

(2000). Organization and development of self-understanding and self-regulation. In Boekaerts

, Pintrich

P. R.

& Zeider

(Eds.), Handbook of self-regulation, (pp. 209–251). San Diego, CA: Academic Press. doi: 10.1016/B978-012109890-2/50036-6

12.

Duckworth

A. L.

, & Seligman

M. E. P.

(2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16, 939–944. doi: 10.1111/j.1467-9280.2005.01641.x

13.

Erentaitė

, & Raižienė

(2015). Lithuanian version of measure of positive youth development based on the Five Cs model. European Journal of Developmental Psychology, 12, 701–717. doi: 10.1080/17405629.2015.1091772

14.

Freund

A. M.

, & Baltes

P. B.

(2002). Life-management strategies of selection, optimization, and compensation: Measurement by self-report and construct validity. Journal of Personality and Social Psychology, 82, 642–662. doi: 10.1037//0022-3514.82.4.642

15.

Geldhof

G. J.

, Bowers

E. P.

, Boyd

M. J.

, Mueller

M. K.

, Napolitano

C. M.

, Schmid

K. L.

, … Lerner

R. M.

(2014). Creation of short and very short measures of the five Cs of positive youth development. Journal of Research on Adolescence, 24, 163–176. doi: 10.1111/jora.12039

16.

Geldhof

G. J.

, Gestsdottir

, Stefansson

, Johnson

S. K.

, Bowers

E. P.

, & Lerner

R. M.

(2015). Selection, optimization, and compensation: The structure, reliability, and validity of forced-choice versus Likert-type measures in a sample of late adolescents. International Journal of Behavioral Development, 39, 171–185. doi: 10.1177/0165025414560447

17.

Geldhof

G. J.

, Little

T. D.

, & Colombo

(2010). Self-regulation across the life span. In Lerner

R. M.

(Ed.), The handbook of life-span development, (Vol. 2, pp. 116–158). New Jersey, NJ: John Wiley & Sons, Inc. doi: 10.1002/9780470880166.hlsd002005

18.

Geldhof

G. J.

, Little

T. D.

, & Hawley

P. H.

(2012). Two measures of self-regulation for young adults and late adolescents in the academic and social domains. International Journal of Behavioral Development, 36, 476–488. doi: 10.1177/0165025412462153

19.

Geldhof

G. J.

, Pornprasertmanit

, Schoemann

A. M.

, & Little

T. D.

(2013). Orthogonalizing through residual centering: Extended applications and caveats. Educational and Psychological Measurement, 73, 27–46. doi: 10.1177/0013164412445473

20.

Gestsdottir

, Bowers

, von Eye

, Napolitano

C. M.

, & Lerner

R. M.

(2010). Intentional self regulation in middle adolescence: The emerging role of loss-based selection in positive youth development. Journal of Youth and Adolescence, 39, 764–782. doi: 10.1007/s10964-010-9537-2

21.

Gestsdottir

, Geldhof

G. J.

, Paus

, Freund

A. M.

, Adalbjarnardottir

, Lerner

J. V.

, & Lerner

R. M.

(2015). Self-regulation among youth in four Western cultures: Is there an adolescence-specific structure of the Selection-Optimization-Compensation (SOC) model?International Journal of Behavioral Development, 39, 346–358. doi: 10.1177/0165025414542712

22.

Gestsdottir

, & Lerner

R. M.

(2007). Intentional self-regulation and positive youth development in early adolescence: Findings from the 4-H study of positive youth development. Developmental Psychology, 43, 508–521. doi: 10.1037/0012-1649.43.2.508

23.

Gestsdottir

, & Lerner

R. M.

(2008). Positive development in adolescence: The development and role of intentional self-regulation. Human Development, 51, 202–224. doi: 10.1159/000135757

24.

Gollwitzer

P. M.

(1999). Implementation intentions - Strong effects of simple plans. American Psychologist, 54, 493–503. doi: 10.1037/0003-066x.54.7.493

25.

Holsen

, Geldhof

G. J.

, Larsen

, & Aardal

(2016). The Five Cs of Positive Youth Development in Norway: Assessment and associations with positive and negative outcomes. International Journal of Behavioral Development, Online: doi: 10.1177/0165025416645668

26.

Larson

R. W.

(2000). Toward a psychology of positive youth development. American Psychologist, 55, 170–183. doi: 10.1037//0003-066x.55.1.170

27.

Lerner

R. M.

, Lerner

J. V.

, Almerigi

J. B.

, Theokas

, Phelps

, Gestsdottir

, ... von Eye

(2005). Positive youth development, participation in community youth development programs, and community contributions of fifth-grade adolescents: Findings from the first wave of the 4-H study of positive youth development. Journal of Early Adolescence, 25, 17–71. doi: 10.1177/0272431604272461

28.

Lerner

R. M.

, Lerner

J. V.

, & Benson

J. B.

(2011). Positive youth development: Research and applications for promoting thriving in adolescence. Advances in Child Development and Behavior, 41, 1–17. doi: 10.1016/B978-0-12-386492-5.00001-4

29.

Lerner

R. M.

, Lerner

J. V.

, Bowers

, & Geldhof

G. J.

(2015). Positive youth development and relational developmental systems. In Overton

W. F.

& Molenaar

P. C.

(Eds.), Theory and Method. Volume 1 of the Handbook of Child Psychology and Developmental Science (7th ed.). Editor-in-chief: Lerner

R. M.

. (pp. 607–651). Hoboken, NJ: Wiley. doi: 10.1002/9781118963418.childpsy116

30.

Little

T. D.

(1997). Means and Covariance Structures (MACS) Analyses of cross-cultural data: Practical and theoretical issues. Multivariate Behavioral Research, 32, 53–76. doi: 10.1207/s15327906mbr3201_3

31.

Little

T. D.

(2013). Longitudinal structural equation modeling. New York: Guilford Press.

32.

Lopes

P. N.

, Salovey

, Cote

, & Beers

(2005). Emotion regulation abilities and the quality of social interaction. Emotio, 5, 113–118. doi: 10.1037/1528-3542.5.1.113

33.

Masten

A. S.

(2004). Regulatory processes, risk, and resilience in adolescent development. Annals of the New York Academy of Science, 102, 310–319. doi: 10.1196/annals.1308.036

34.

McClelland

, Geldhof

G. J.

, Cameron

, & Wanless

(2015). Development and Self-regulation. In Overton

W. F.

& Molenaar

P. C.

(Eds.), Volume 1 of the Handbook of Child Psychology and Developmental Scienc (7th ed). Editor-in-chief: Lerner

R. M.

. (pp. 523–565). Hoboken, NJ: Wiley. doi: 10.1002/9781118963418.childpsy114

35.

Millsap

R. E.

, & Yun-Tein

(2004). Assessing factorial invariance in ordered-categorical measures. Multivariate Behavioral Researc, 39, 479–515. doi: 10.1207/s15327906mbr3903_4

36.

Reise

S. P.

(2012). The rediscovery of bifactor measurement models. Multivariate Behavioral Researc, 47, 667–696. doi: 10.1080/00273171.2012.715555

37.

Scales

P. C.

, Roehlkepartain

E. C.

, Wallace

, Inselman

, Stephenson

, & Rodriguez

(2015). Assessing youth well-being in global emergency settings: Early results from the Emergency Developmental Assets Profile. Journal of Adolescenc, 45, 98–102. doi: 10.1016/j.adolescence.2015.09.002

38.

Spencer

M. B.

(2006). Phenomenology and ecological systems theory. Development of diverse groups. In Lerner

R. M.

(Ed.), Theoretical models of human development (pp. 829–893). Vol. 1 in Damon

& Lerner

R. M.

(Eds.-in-Chief), Handbook of child psycholog (6th ed.). Hoboken, NJ: Wiley. doi: 10.1002/9780470147658.chpsy0115

39.

Stefansson

K. K.

, Gestsdottir

, & Skulason

(2014). ðróun og mat á rétttmæti mælitækis á meðvitaðri sjálfstjórnun ungmenna [The development and validation of the SOC scale for youth in Iceland]. Sálfræðiritð [The Icelandic Psychological Journal], 19, 41–55.

40.

Wen

, Su

, Li

, & Lin

(2015). Positive youth development in rural China: The role of parental migration. Social Science & Medicin, 132, 261–269. doi: 10.1016/j.socscimed.2014.07.051

41.

Wiese

B. S.

, Freund

A. M.

, & Baltes

P. B.

(2000). Selection, optimization, and compensation: An action-related approach to work and partnershi. Journal of Vocational Behavior, 57, 273–300. doi: 10.1006/jvbe.2000.1752

42.

A. D.

, Li

, & Zumbo

B. D.

(2007). Decoding the meaning of factorial invariance and updating the practice of multi-group confirmatory factor analysis: A demonstration with TIMSS data. Practical Assessment, Research & Evaluatio, 12, 1–26.