Indicators of positive youth development can be maladaptive: The example case of caring

Journal of Adolescence 71 (2019) 1–9

Contents lists available at ScienceDirect

Journal of Adolescence journal homepage: www.elsevier.com/locate/adolescence

Indicators of positive youth development can be maladaptive: The example case of caring

T

G. John Geldhofa,∗, Torill Larsenb, Helga Urkeb, Ingrid Holsenb, Hillary Lewisa, Corine P. Tylera a b

Human Development and Family Studies, Oregon State University, USA Health Promotion and Development, University of Bergen, Norway

ABS TRA CT

Introduction: Research on positive youth development typically takes a more-is-better approach in which higher scores on measures of positive youth development indicate better outcomes. We question the validity of this assumption and describe how an imbalance among the Five Cs of Positive Youth Development can lead to a situation where “more” is actually “less.” We then provide an empirical illustration using crosssectional data. Methods: We examined conditional associations between indicators of thriving and the C of caring using survey data obtained from a sample of 2,386 students in Norway (mean age 16.60 years [SD = 1.98]; 44.8% girls). Multilevel models assessed whether associations between caring and the thriving indicators (here defined as greater mental well-being and lower anxiety and depressive symptoms) differed when versus when not controlling for overall positive youth development. Results: When positive youth development was not included as a covariate, caring was positively associated with mental well being and anxiety but not significantly associated with depressive symptoms. After adding positive youth development to the model, caring became less strongly associated with mental well-being, more strongly associated with anxiety, and significantly (positively) associated with depressive symptoms. Conclusions: Our results indicate that, after controlling for a set level of positive youth development, caring does not predict adaptive outcomes in a monotonic more-is-better way. These findings illustrate an instance where developmental regulations benefit the context at the cost of the individual and therefore speak to a growing acknowledgement in the literature—young people may suffer if they “care too much.”

1. Introduction The positive youth development perspective has gained widespread traction since its inception in the 1990s and early 2000s (e.g., by Catalano, Berglund, Ryan, Lonczak, & Hawkins, 2002; Eccles & Gootman, 2002; Lerner et al., 2005). The field now largely acknowledges the importance of preventing and remediating negative developmental phenomena while simultaneously promoting skills young people need to thrive. As strength-based efforts become increasingly pervasive, however, researchers and practitioners must remember that the characteristics and actions used to operationalize positive youth development are not inherently adaptive in and of themselves. Indicators of positive youth development promote thriving only when they exist as part of mutually beneficial relations between individuals and their unique contexts. This article highlights the conditional association between indicators of positive youth development and adaptive development. We begin by discussing adaptive development through the lens of relational developmental systems metatheory and describe one model of positive youth development that has been especially linked to that perspective—the Five Cs of Positive Youth Development (Lerner et al., 2005). Of the Five Cs (competence, confidence, character, caring, and connection), existing empirical evidence points, ∗

Corresponding author. HDFS, 470 Waldo Hall, Oregon State University, Corvallis, OR, 97331, USA. E-mail addresses: [email protected] (G.J. Geldhof), [email protected] (T. Larsen), [email protected] (H. Urke), [email protected] (I. Holsen), [email protected] (H. Lewis), [email protected] (C.P. Tyler). https://doi.org/10.1016/j.adolescence.2018.11.008 Received 26 June 2018; Received in revised form 28 November 2018; Accepted 29 November 2018 0140-1971/ © 2018 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

in particular, to ambivalent associations in which caring has been positively associated with both adaptive and maladaptive outcomes (e.g., anxiety; Geldhof, Bowers, Mueller, et al., 2014; Holsen, Geldhof, Larsen, & Aardal, 2017). We therefore use that C as an example case, testing whether caring can be positively associated with greater internalizing symptoms under some circumstances. 1.1. Relational developmental systems approaches to adaptive development Relational developmental systems perspectives (RDSPs) reflect a broad paradigm committed to understanding development as a holistic, relational process (e.g., Overton, 2010; 2015). Neither characteristics of the individual nor characteristics of the context can influence development independently because both person and context are merely components of a larger, fused system (see also, Overton’s, 2015, Identity of Opposites). Despite this widely acknowledged fusion, however, parsimony dictates that it is easier—and oftentimes more useful—to model1 the person-context system as containing distinct “person” and “context” components (see also Overton's, 2015, Opposites of Identity). We can therefore heuristically describe development of the person-context system as occurring through mutually influential (i.e., bidirectional) person ←→ context relations. Described in this way, it becomes clear that both person and context co-regulate development. Lerner (1998; 2015; Lerner et al., 2006) therefore uses the term “developmental regulations” to describe mutually influential person ← → context associations. 1.1.1. Adaptive developmental regulations Developmental regulations are the unit of analysis when approaching development from a RDSP (Geldhof, Bowers, Johnson, et al., 2014), meaning they underlie both adaptive and maladaptive outcomes. Researchers must therefore find some way to distinguish between developmental regulations they would like to promote (i.e., that are adaptive) and those they would like to prevent (i.e., that are maladaptive). Lerner, Dowling, and Anderson (2003) clarified that adaptive developmental regulations cannot selectively benefit either the person or the context at the expense of the other, meaning adaptive developmental regulations can only include person ← → context relations that are mutually beneficial (i.e., from which both person and context benefit). Adaptive developmental regulations indicate cohesion between the unique characteristics of individuals and their contexts and thus reflect true thriving. 1.1.2. Non-adaptive developmental regulations In this article, we extend the existing literature on adaptive developmental regulations by additionally defining several kinds of non-adaptive developmental regulations. For example, neutral developmental regulations neither harm nor benefit the person or the context to a discernible degree (e.g., choosing which breakfast cereal to eat tomorrow) and may be relatively unimportant for adaptive development. Negative developmental regulations, on the other hand, include processes that actively harm the person and/or the context and may be especially important to understand as part of a holistic positive youth development perspective. The most readily apparent form of negative developmental regulations, maladaptive developmental regulations, harm both person and context and can occur for a variety of reasons. For instance, individuals may be naive to the effects of their actions (e.g., failing to appreciate the full impact of cigarette smoking) or may even engage in maladaptive developmental regulations under the misguided assumption that they are doing good (e.g., as may have occurred when tobacco companies advertised their cigarettes as “healthy”). Parasitic developmental regulations, then, benefit the individual at the expense of the context. Individual instances of parasitic developmental regulations may not have direct negative implications for the individual (e.g., instrumental aggression against a peer of lower social status), although their consistent occurrence can quickly escalate into a maladaptive pattern of behavior (e.g., becoming known as a bully). Last, some negative developmental regulations will benefit the context at the expense of the individual. For instance, sacrificial developmental regulations entail context-level changes that directly harm an individual but benefit the context as a whole (e.g., passing a policy that benefits the majority but harms a subset of the population). Martyring developmental regulations, then, reflect an individual's agentic actions that benefit the context at the expense of the self (e.g., altruistic self-sacrificing). Thus, the distinction between sacrificial and martyring developmental regulations lies in whether they are initiated primarily at the contextual level or by the individual harmed by them. Although developmental regulations undergird development across the life span, periods of high relative plasticity (e.g., adolescence) provide ideal ‘ontogenetic laboratories’ for testing the tenets of RDSPs in general and assumptions about adaptive developmental regulations in particular (Lerner, Lerner, & Benson, 2011; Lerner, Freund, De Stefanis, & Habermas, 2001). In many ways, models of positive youth development are therefore youth-specific models of adaptive developmental regulations. The 4-H Study of Positive Youth Development (Lerner et al., 2005) was even explicitly designed to test whether aligning the strengths of youth with opportunities and resources available in their contexts can promote adaptive developmental regulations as operationalized by the Five Cs of Positive Youth Development. Because of the congruency between our metatheoretical perspective and the Five Cs Model, we focus next on describing that model and how each of the Five Cs can reflect both adaptive and non-adaptive developmental regulations. 1.2. The Five Cs Model The primary hypothesis underlying the Five Cs Model is that positive youth development can be described by five broad 1 We encourage readers to keep in mind that models are, as a necessity, approximations that sacrifice precision for utility (e.g., as discussed by Box, 1976).

2

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

constructs—competence, confidence, character, caring, and connection (Eccles & Gootman, 2002; Lerner et al., 2005). In other words, when youth exhibit all five of the Cs, they are likely experiencing adaptive developmental regulations. Development of the Cs should therefore lead young people to also exhibit the ‘sixth’ C of contribution. The Five Cs reflect prominent terms used by practitioners, parents, and young people when describing the characteristics of “thriving youth” (King et al., 2005) and therefore have practical validity in addition to a strong theoretical basis in a RDSP. Despite the outwardly straightforward nature of the Five Cs, it is important to note that each C describes more than just a personal asset. Reflecting the holistic, relational nature of RDSPs, each C describes an intersection between young people and their contexts. Competence represents an ability to navigate the context effectively in order to achieve strived-for goals. Confidence arises when the navigation of one's context results in feelings of personal agency and self-worth. Character represents the internalization of moral standards through repeated person ← → context relations as well as the behavioral manifestations of that moral code. Caring indicates developmentally and contextually appropriate levels of concern for others, and connection requires that the individual be embedded in, and supported by, a reliable and diverse social network. 1.2.1. Modelling the Five Cs The Five Cs Model has received copious empirical support (e.g., Heck & Subramaniam, 2009; Lerner, Lerner, Bowers, & Geldhof, 2015), especially from data collected as part of the 4-H Study. For instance, Lerner et al. (2005) used data from the first wave of the 4H Study to confirm the reasonableness of aggregating measures of the Five Cs into corresponding latent constructs and of imposing a higher-order factor that represented overall positive youth development (see also Jeličić, Bobek, Phelps, Lerner, & Lerner, 2007). Further research then extended these findings across early and middle adolescence (e.g., Bowers et al., 2010; Phelps et al., 2009). Rather than specify positive youth development as a universal composite of the Five Cs, more recent research has hypothesized that indicators of positive youth development might differentially represent overall adaptive functioning versus the individual Cs (Geldhof, Bowers, Mueller, et al., 2014). Leveraging the strengths of bifactor confirmatory factor analysis, this work argued for a more nuanced understanding of youth development. For example, relevant findings suggest that a young person's self-rated physical attractiveness indicates self-rated confidence but is a much weaker indicator of overall positive youth development. Athleticism (i.e., physical competence) similarly indicates competence but not overall positive youth development. In short, the aforementioned work highlighted the fact that indicators of the Five Cs may not consistently indicate overall thriving. 1.2.2. The Five Cs are not universally adaptive Studies that utilized bifactor models provided initial evidence that indicators of positive youth development are not strictly unidimensional. Measures of the Five Cs can meaningfully be decomposed into variance related to adaptive developmental regulations (i.e., global positive youth development) and C-specific variance not related to overall positive youth development. In such models, each component may be uniquely correlated with criterion variables and therefore be uniquely associated with developmental strengths and outcomes. For instance, Gestsdottir, Geldhof, Lerner, and Lerner (2017) showed that self-regulation skills are strongly related to global positive youth development, moderately correlated with the residual C of competence, and essentially uncorrelated with the other residual Cs. Geldhof, Bowers, Mueller, et al. (2014) similarly found that risk taking is more strongly associated (albeit negatively) with global positive youth development than with the residual Cs. Although overall positive youth development is associated with lower depression and/or anxiety (Erentaitė & Raižienė, 2015; Geldhof, Bowers, Mueller, et al., 2014; Holsen et al., 2017), the residual C of caring can correlate positively with these outcomes (Geldhof, Bowers, Mueller, et al., 2014; Holsen et al., 2017; Truskauskaitė-Kunevičienė et al., 2014). Separating the Five Cs into multiple components highlights the context-dependent association between these indicators and actual thriving. Using the above typology of adaptive versus neutral and negative developmental regulations, individuals with elevated levels of sociopathy or who display Machiavellian tendencies may also exhibit excess levels of competence and confidence (which Geldhof, Bowers, Mueller, et al., 2014, called the efficacious Cs). Thus, higher levels of these Cs, after accounting for global positive youth development, might place individuals at risk for parasitic developmental regulations. Conversely, excess levels of the socioemotional Cs (i.e., character and caring; Geldhof, Bowers, Mueller, et al., 2014) may increase the likelihood that an individual will engage in martyring developmental regulations. Thus, to paraphrase a common saying, nice people sometimes finish last. 1.2.3. Caring as an exemplary case As noted above, existing data especially support the notion that excessive Caring can be maladaptive. When discussing caring that extends beyond positive youth development, Geldhof, Bowers, Mueller, et al. (2014) speculated that the positive association between caring and maladaptive outcomes may indicate, “emotional hypersensitivity, or an anxiety-producing over concern for (or about) others' thoughts and feelings” (p. 944). Holsen et al. (2017) similarly speculated that, “Some young people high on empathy and sympathy might have a tendency toward manifesting higher anxiety and depressive feelings,” (p. 567). In addition to empirical studies derived from the positive youth development literature, broader research on empathy and sympathy suggests higher levels of caring can sometimes be maladaptive. Although empathizing with others' positive and negative emotions is associated with many positive social, emotional, and moral outcomes, recent research suggests doing so might come at a cost for some adolescents (Piko & Pinczés, 2014; Smith, 2015; Van Lissa, Hawk, Koot, Branje, & Meeus, 2017), and that gender may moderate this overall association (Piko & Pinczés, 2014; Smith, 2015; Smith & Rose, 2011). Indeed, some researchers have speculated that caring and empathy for others may play a role in the expression of these internalizing problems (Smith, 2015; Tone & Tully, 2014; Van Lissa et al., 2017). There has accordingly been a shift in the research literature away from independently cultivating positive behaviors (e.g., empathy, sympathy) and discouraging maladaptive behaviors (e.g., aggression) to instead adopt a more 3

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

holistic view that recognizes positive and negative outcome trade-offs. In tandem with this shift, the idea has emerged that there may be certain risks associated with “caring too much” for adolescents in their relationships with people close to them (Piko & Pinczés, 2014; Smith, 2015; Van Lissa et al., 2017). Thus, treating positive and negative outcomes as either independent or as opposite ends of a single continuum is an over-simplification. Internalizing symptoms may be related to a complex intersection between empathy, intraindividual (e.g., genetically influenced biological predispositions), and interindividual (e.g., inadequate parenting, negative home environments) factors (Tone & Tully, 2014). Across these studies, we can see a clear pattern suggesting the potential risk of internalizing problems when an adolescent “cares too much.” 1.2.4. Summarizing the goals of the present paper The overarching goal of this paper is to highlight the possibility that individual measures of adaptive development can reflect, in certain contexts, negative developmental regulations. We focus on the Five Cs as a common model of thriving that exemplifies RDSP definitions of adaptive (and negative) developmental regulations and focus more specifically on the C of caring in our empirical illustration. Whereas previous research derived from the positive youth development perspective has reported positive associations between caring and internalizing symptoms as a secondary (and typically not hypothesized) finding, we instead make this association the focus of our analyses. We not only attempt to replicate the positive association between caring and internalizing symptoms seen when controlling for overall positive youth development, but we also examine unconditional associations in order to more fully understand the potential for maladaptive outcomes. The cited positive youth development research also often relies on one specific model—bifactor confirmatory factor analysis—to control for global positive youth development. Although sophisticated modelling approaches like bifactor CFAs are critical for pushing the field forward, it is also important to establish methodological robustness by replicating key findings using alternative analyses. Doing so ensures that key findings are not unduly influenced by the nuances of a specific method. As such, the present paper also extends previous positive youth development research by examining these associations using a multilevel regression framework. 2. Methods 2.1. Participants and procedures Data for the present study come from the second wave (T1) of the COMPLETE study (Larsen et al., 2018), an evaluation of a threeyear randomized controlled trial of the Dream School Program (DSP) (Adults for Children [Voksne for Barn], 2017) and the Mental Health Support Team (MHST) (Larsen et al., 2018). Both interventions aim to promote positive youth development and mental health wellbeing among youth in upper secondary schools in Norway. In Norway, the first ten-year school track is mandatory for all children aged 6–16 years. All students in Norway are also guaranteed acceptance to a three-year upper secondary public school where they can choose between vocational education programs or general studies programs. The majority of students start upper secondary school at age 15 or 16. The total sample for the survey consisted of 2386 (response rate 77.5%) 1st grade students (mean age 16.60 years (SD = 1.98); 44.8% girls and 55.2% boys) from 17 upper secondary schools. Participants answered a confidential web-based survey in their respective classrooms that took approximately 30 min to complete. Two representatives from the research team were present at each school, and the Norwegian Centre for Research Data (NSD) approved this study. Larsen et al. (2018) provide additional project details. 2.2. Measures 2.2.1. Outcome measures Anxiety and depressive symptoms were measured by the short form of the Symptom Check List (SCL-5;Tambs and Moum, 1993). Participants responded to five items on how they felt during the past two weeks, with response options ranging from 1 = not bothered to 4 = very bothered. Three items indicated anxiety (α = 0.86) and two items indicated depressive symptoms (α = 0.78). Mental wellbeing was measured by the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS; Tennant et al, 2007; Clarke et al, 2011), a 14-item scale in which participants respond to statements about feelings and thoughts during the past two weeks. Response options ranged from 1 = none of the time to 5 = all of the time. Reliability of mental well-being was α = 0.96 in the present sample. 2.2.2. Target predictor Caring was measured using seven items reflecting sympathy and empathy. Six items came from the short measure of the Five Cs of Positive Youth Development (Geldhof, Bowers, Mueller, et al., 2014; items originally from the Eisenberg Sympathy Scale [Eisenberg et al., 1996] and the Empathic Concern Subscale of the Interpersonal Reactivity Index [Davis, 1983]) plus one additional item (“It bothers me when bad things happen to good people”). The participants responded to how much these statements “describe you,” with response options ranging from 1 = not well to 4 = very well. Reliability of the seven-item composite was α = 0.91 in the present sample. 2.2.3. Covariates In order to test the association between caring and the selected outcome measures, after controlling for global positive youth development, we selected three subscales from the competence and confidence assessments of a short measure of the Five Cs (i.e., 4

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

academic competence, self-worth, and positive identity; Geldhof, Bowers, Mueller, et al., 2014). The selected items (academic competence, self-worth, and positive identity) consistently indicated global positive youth development in the scale's original validation, but the residual competence and confidence factors (i.e, after controlling for global positive youth development) were only weakly correlated with the residual caring factor. Thus, controlling for these items provided an opportunity to control for global positive youth development without relying on a bifactor CFA. Academic competence was measured by two items from the short measure of the Five Cs of Positive Youth Development (Geldhof, Bowers, Mueller, et al., 2014; original items from; Harter, 1988). Participants reported how they think about themselves (i.e., I feel I am just as smart as others at my age and I do very well at my classwork) with response options ranging from 1 = strongly disagree to 4 = strongly agree. Reliability in the present sample was α = 0.61. Measures with few items often exhibit lower reliability, and we therefore determined that this reliability was acceptable for the illustrative purposes of the present study. Self-worth was measured by two items from the short measure of the Five Cs of Positive Youth Development (Geldhof, Bowers, Mueller, et al., 2014; original items from; Harter, 1988). Participants responded to I am happy with myself most of the time and I am very pleased with how I am, with response options ranging from 1 = strongly disagree to 4 = strongly agree, and the scale exhibited good reliability in this sample (α = 0.82). Positive identity was measured by two items from the short measure of the Five Cs of Positive Youth Development (Geldhof, Bowers, Mueller, et al., 2014; original items from; Leffert et al., 1998). Participants responded to All in all, I am glad I am me and When I am an adult, I'm sure I will have a good life, with response options ranging from 1 = strongly disagree to 4 = strongly agree. Reliability of this scale was α = 0.78 in the present sample. In addition to positive youth development, all models controlled for demographic covariates. Socioeconomic status was measured through student self-report of family economic standing ranging from 1 = very well off to 5 = very bad off, which were then aggregated into three response categories (1 = not so good, 2 = somewhat well-off, and 3 = well off). In order to accommodate both the limited response options selected and potential nonlinearity, we re-specified the SES variable as two dummy codes in which option three (well off) served as the reference group. Gender data were retrieved from country school departments with categories 1 = female and 2 = male. Last, participants were enrolled in one of three treatment groups (control, Dream School, Dream School + Mental Health Support Team). This variable was coded into two dummy codes in which Dream School + Mental Health Support Team served as the referent group. 2.3. Analyses We fit two two-level path analyses, with observations nested in classrooms and standard errors additionally corrected for nesting in schools. These analyses were performed using Mplus (i.e., TYPE = COMPLEX TWOLEVEL). Because we were primarily interested in Level 1 effects, and because the conditional ICCs after controlling for gender, SES, and treatment were less than 0.05, we saturated the Level 2 portion of the model by allowing all Level 2 variables to freely covary with each other. We did not explicitly model Level 3 (variance across schools) because the ICCs and conditional ICCs at that level were all lower than 0.02. At Level 1, the first model regressed the three outcomes on caring, gender, and the two SES dummy codes. We allowed the three positive youth development measures to freely covary with everything at Level 1 (Treatment did not vary at Level 1 and was therefore not included in this regression). The second regression model additionally regressed the outcome measures on the three positive youth development measures. The first model therefore tested the direct association between caring and the outcomes, whereas the second model tested the conditional association between caring and the outcomes, controlling for overall positive youth development. Both regression models were fully saturated at Level 1, meaning we estimated residual covariances among the three outcomes and allowed all predictors/covariates to freely covary. The models therefore had zero degrees of freedom and fit the data perfectly. Missing data (3.42% total missingness, range for individual predictors and outcomes [0.00%, 10.65%]) were accommodated using maximum likelihood and by specifying all predictors and outcomes as endogenous (y-side) variables. To verify robustness, we also replicated all results in SAS using PROC MIXED (three-level regression models that included random intercepts at both the classroom and school levels, results not reported here). 3. Results Table 1 provides means, standard deviations, and counts for all considered variables. Table 2 presents zero-order correlations between caring, the three measures of positive youth development, and the target outcomes. As Table 2 shows, caring was positively correlated with both the positive and negative developmental outcome variables prior to controlling for any covariates. Conversely, the three variables used as proxies for global positive youth development were negatively associated with anxiety and depressive symptoms while being positively associated with mental well-being. This finding supports the general assumption that adaptive and problematic outcomes should be at least somewhat negatively correlated. The positive youth development proxies were only weakly correlated with caring, supporting the idea that caring may be a multi-valenced construct (i.e., more is not always better). Table 3 contains raw and standardized parameter estimates from the two regression models. As the top part of the table shows, caring was positively associated with both anxiety and mental well-being but was not significantly associated with depressive symptoms after controlling for the demographic covariates. Adding the three proxies for positive youth development as predictors changed the associations between caring and the outcomes, with the association between caring and depressive symptoms becoming positive and statistically significant, the positive association between caring and anxiety increasing in magnitude, and the positive 5

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

Table 1 Descriptive statistics (N = 2384). Variable

Mean (SD)

Anxiety Depressive Symptoms Mental Well-Being Caring Academic Competence Self-Worth Positive Identity Gender Male Female Treatment Group Control Dream School Dream School + Mental Health Support SES Not so good Somewhat well off Well off

1.84 1.91 3.50 4.00 3.09 3.00 3.88

Count (% of Sample)

(0.84) (0.92) (0.85) (0.85) (0.64) (0.77) (0.85) 1315 (55.16) 1069 (44.84) 574 (24.06) 821 (34.41) 991 (41.53) 169 (7.22) 637 (27.22) 1534 (65.56)

Note: Missing data deleted at the item level. Table 2 Zero-Order Correlations (N = 2384; pairwise deletion).

1. 2. 3. 4. 5. 6. 7. 8.

Gender Anxiety Depress. MWB Caring Academic Self-Worth Pos. Identity

1.

2.

3.

4.

5.

6.

7.

8.

1.00 -.36∗∗∗ -.33∗∗∗ .16∗∗∗ -.35∗∗∗ .05∗ .27∗∗∗ .10∗∗∗

1.00 .79∗∗∗ -.41∗∗∗ .17∗∗∗ -.24∗∗∗ -.50∗∗∗ -.43∗∗∗

1.00 -.48∗∗∗ .12∗∗∗ -.31∗∗∗ -.53∗∗∗ -.51∗∗∗

1.00 .15∗∗∗ .36∗∗∗ .51∗∗∗ .54∗∗∗

1.00 .11∗∗∗ .02 .15∗∗∗

1.00 .40∗∗∗ .37∗∗∗

1.00 .61∗∗∗

1.00

*p < .05; **p < .01; ***p < .001. Note: Missing data deleted pairwise; Depress = Depressive Symptoms; MWB = Mental Well-Being; Academic = Academic Competence. Table 3 Level 1 regression coefficients from path analyses (N = 2384). Anxiety B (SE) Positive Youth Development Not Included Gender -.53 (.04)∗∗∗ SES1 .41 (.07)∗∗∗ SES2 .17 (.03)∗∗∗ Caring .08 (.03)∗∗∗ Positive Youth Development Included Gender -.33 (.04)∗∗∗ SES1 .14 (.06)∗ SES2 .03 (.02) Academic -.05 (.03) Self-Worth -.30 (.03)∗∗∗ Pos. Identity -.22 (.02)∗∗∗ Caring .15 (.02)∗∗∗

Depress

MWB

β

B (SE)

β

B (SE)

β

-.27 .13 .10 .08

-.53 (.04)∗∗∗ .51 (.08)∗∗∗ .22 (.03)∗∗∗ .03 (.03)

-.24 .15 .11 .02

.33 (.03)∗∗∗ -.29 (.08)∗∗∗ -.25 (.04)∗∗∗ .21 (.03)∗∗∗

.16 -.09 -.14 .20

-.16 .05 .02 -.04 -.28 -.24 .15

-.30 (.03)∗∗∗ .16 (.07)∗ .04 (.03) -.13 (.04)∗∗ -.31 (.02)∗∗∗ -.32 (.02)∗∗∗ .13 (.02)∗∗∗

-.13 .05 .02 -.09 -.26 -.31 .11

.11 (.03)∗∗ .04 (.07) -.07 (.03)∗∗ .18 (.03)∗∗∗ .27 (.03)∗∗∗ .29 (.03)∗∗∗ .11 (.02)∗∗∗

.06 .01 -.04 .13 .24 .31 .11

*p < .05; **p < .01; ***p < .001. Depress = Depressive Symptoms; MWB = Mental Well-Being; Academic = Academic Competence.

association between caring and mental well-being weakening but remaining statistically significant. As a post-hoc analysis, we fit a two-group model that modeled the same data in each group, but in which the model for the first group specified the above model that did not control for the positive youth development proxies and the model for the second group specified the above model that did control for the positive youth development proxies. This model allowed us to compare parameter estimates from the two models directly. Wald tests and single-degree-of-freedom likelihood ratio tests all indicated that each 6

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

regression coefficient from the unconditional model was significantly different from its counterpart in the model that did control for the positive youth development proxies, although it is not clear how strongly this two-group approach biased the group comparisons. When comparing results from the two-group model against results from the respective one-group models, the parameter estimates were identical across models. The standard errors were slightly smaller in the two-group model, however (e.g., .026 for the one group model versus .025 for the two-group model). 4. Discussion The Five Cs Model derives from a relational, systems-oriented worldview that defines individuals and contexts as inextricably fused components of the same developmental system (Geldhof, Bowers, Johnson, et al., 2014). Due to this embeddedness, models that artificially separate persons from contexts must—at a minimum—acknowledge mutually influential person ←→ context relations (i.e., developmental regulations; Lerner, 1998; 2015; Lerner et al., 2006) as the building blocks of human development and as the units of analysis in developmental research. Developmental regulations are agnostic in terms of whether the person and/or the context benefits as a result, and optimizing development requires a differentiation between developmental regulations that are adaptive (i.e., that benefit both person and context) and those that are not. When they manifest together, the Five Cs indicate a young person is on a positive developmental trajectory and is therefore likely to be experiencing adaptive developmental regulations. This general statement does not, however, mean the individual Cs must unambiguously and positively indicate adaptive development. In isolation, each C carries a latent potential for promoting maladaptive developmental outcomes. Young people with high levels of competence and confidence but low levels of the socio-emotional Cs (i.e., caring and character) run the risk of parasitic developmental regulations. As Larson (2011) noted, developmental outcomes like this “can be explained with one word: Machiavelli” (329). Similarly, youth who show high levels of empathy, sympathy, and caring but are otherwise poorly equipped to navigate their environments (e.g., by having inadequate competence and confidence) risk experiencing what we have called martyring developmental regulations. Taken together, these ideas closely reflect a concern expressed by Reed Larson in his 2010 presidential address to the Society for Research on Adolescence: The point is that positive development has many dimensions: ethical, interpersonal, civic, spiritual, etc. (Eccles & Gootman, 2002; Lerner, Phelps, Forman, & Bowers, 2009). We do not want teens to develop one and neglect others. We do not want agency skills without ethics; nor are ethics much good without agency (p.329)2. The present paper provides an empirical test of these ideas. Drawing from previous unexpected findings that caring may be positively associated with internalizing problems (Erentaitė & Raižienė, 2015), especially after controlling for overall positive youth development (Geldhof, Bowers, Mueller, et al., 2014; Holsen et al., 2017), we assessed whether this association could be purposefully replicated in a new dataset. Replicating this previous research, our models indicated a positive association between caring and anxiety and a conditionally positive association between caring and depressive symptoms (when controlling for indicators of global positive youth development). Thus, individuals who have a very high investment in the thoughts, feelings, and perceptions of others (i.e., are high on caring) are not necessarily ‘well-adapted.’ Although additional work should explore mechanisms that underlie the observed association, being over-invested in others may lead to anxiety about social situations or vice versa. Internalizing this anxiety might then lead to feelings of depression. However, these findings do not suggest caring is universally a negative characteristic. In addition to the aforementioned negative associations, caring was consistently and positively associated with our measure of mental well-being, even after controlling for our indicators of global positive youth development. Together, these findings confirm caring as a multi-valenced construct and support the conceptualization of caring as part of a larger, complexly connected developmental system. Future measurement work must therefore clarify which aspects of caring can be assessed independently among which persons and under which circumstances (see also Bornstein's, 2017, specificity principle). Only then can we fully understand how each facet predicts and is predicted by other indicators of thriving. 4.1. Implications for research and practice The present findings indicate that, at a minimum, caring can take on multiple contextually dependent meanings. It is likely that other indicators of thriving (e.g., the remaining four Cs) will do the same, and this paper gives guidance as to where researchers may wish to begin the move away from nomothetic, population-oriented analyses and toward more group- and person-centered analyses. For example, inductive group-centered analyses such as latent profile and latent class analyses should examine whether different types of individuals exhibit distinct profiles of the Five Cs and, consequently, different levels of key developmental outcomes. Personcentered analyses such as those utilizing the p-technique (Cattell, 1966; Lee & Little, 2012) could additionally discern how levels of each C and levels of important outcomes (e.g., depressive symptoms) covary within individuals across time. Moving to these groupand person-oriented techniques will offer greater precision for identifying youth at risk of experiencing negative developmental outcomes. Another implication of our findings is that existing measures of caring may be too coarse to detect the nuanced differences between individuals who exhibit caring as part of adaptive versus negative developmental regulations. Existing measures should be supplemented with items that more explicitly acknowledge potentially problematic aspects of caring, for instance by asking, “Do 2

Published as Larson, 2011. 7

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

others tend to not care about you as much as you care about them?” or, “Do you put too much emotional investment into others?” Having data from such items will directly inform whether participants themselves are aware that they “care too much” and will help researchers tease apart caring's different dimensions. Our findings also highlight the need for practitioners to treat unique indicators of thriving (e.g., the individual Cs) as components of a larger developmental system. For programs targeting youth with very specific needs (e.g., low emotional intelligence), it may be perfectly acceptable to focus explicitly on promoting socio-emotional skills. However, programs that serve a broader audience must emphasize adaptive development as a holistic process and therefore cannot focus exclusively on promoting a limited range of personcontext associations. For instance, programs that only teach the importance of caring will likely lead some young people to overinvest in others in ways that are detrimental to themselves (e.g. by eliciting anxiety about others' needs). 4.2. Limitations and future directions The above findings provide unique insight into the nature of caring, but several factors limit their generalizability. First, we relied on the same measure of caring used in previous research that found positive associations with anxiety and depressive symptoms. Although this measure is well established, it is likely that we have only captured a narrow slice of the caring construct (e.g., general caring for all people), and it is unclear whether our findings would replicate using alternative measures. This is especially true for measures that capture potentially problematic aspects of caring. Second, and as implied above, we relied on a strictly nomothetic statistical model in order to intuit how complex processes may differ between diverse individuals. The absence of idiographic nuance means the findings may not apply to individuals with extreme circumstances (e.g., those in a clinical setting) and may not apply within individuals as they develop. Thus, although the present research provides strong justification for future exploration of our ideas, additional longitudinal and person-oriented studies that include diverse individuals and more nuanced measures will ensure the greatest real-world impact. Complementing the present study with such research will therefore provide robust information about the multidimensional nature of the individual Cs while also helping answer the primary question facing developmental science today (e.g., Lerner et al., 2012): which individuals with which characteristics and embedded in which contexts benefit in which ways from which programs, interventions, and experiences? Acknowledgements This work was supported in part by a grant from the Norwegian Ministry of Education to Dr. Torill Larsen (PI) and a grant from the Stupski Foundation to Dr. G. John Geldhof (co-PI). References Bornstein, M. H. (2017). The specificity principle in acculturation science. Perspectives on Psychological Science, 12, 3–45. Bowers, E. P., Li, Y., Kiely, M. K., Brittian, A., Lerner, J. V., & Lerner, R. M. (2010). The five Cs model of positive youth development: A longitudinal analysis of confirmatory factor structure and measurement invariance. Journal of Youth and Adolescence, 39, 720–735. Box, G. E. (1976). Science and statistics. Journal of the American Statistical Association, 71(356), 791–799. Catalano, R. F., Berglund, M. L., Ryan, J. A. M., Lonczak, H. S., & Hawkins, J. D. (2002). Positive youth development in the United States: Research findings on evaluations of positive youth development programs. Prevention & Treatment, 5(1), Retreived from http://psycnet.apa.org/fulltext/2002-14078-001.pdf. Clarke, A., Friede, T., Putz, R., Ashdown, J., Martin, S., Blake, A., ... Stewart-Brown, S. (2011). Warwick-Edinburgh Mental Well-Being Scale (WEMWBS): Validated for teenage school students in England and Scotland. A mixed methods assessment. BMC Public Health, 11(487). Cattell, R. B. (1966). Psychological theory and scientific method. In R. B. Cattell (Ed.). Handbook of multivariate experimental psychology (pp. 1–18). Chicago, IL: Rand McNally. Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113–126. Eccles, J., & Gootman, J. (Eds.). (2002). Community programs to promote youth development. Washington, DC: National Academy Press. Eisenberg, N., Fabes, R. A., Murphy, B., Karbon, M., Smith, M., & Maszk, P. (1996). The relations of children's dispositional empathy-related responding to their emotionality, regulation, and social functioning. Developmental Psychology, 32(2), 195. Erentaitė, R., & Raižienė, S. (2015). Lithuanian version of measure of positive youth development based on the Five Cs model. European Journal of Developmental Psychology, 12, 701–717. Geldhof, G. J., Bowers, E. P., Johnson, S. K., Hershberg, R., Hilliard, L., Lerner, J. V., et al. (2014). Relational developmental systems theories of positive youth development: Methodological issues and implications. In P. C. M. Molenaar, R. M. Lerner, & K. M. Newell (Eds.). Handbook of developmental systems theory and methodology (pp. 66–94). New York, NY: Guilford Press. Geldhof, G. J., Bowers, E. P., Mueller, M. K., Napolitano, C. M., Callina, K. S., & Lerner, R. M. (2014). Longitudinal analysis of a very short measure of positive youth development. Journal of Youth and Adolescence, 43(6), 933–949. Gestsdottir, S., Geldhof, G. J., Lerner, J. V., & Lerner, R. M. (2017). What drives positive youth development? Assessing intentional self-regulation as a central adolescent asset. International Journal of Developmental Science, 11(3–4), 69–79. Harter, S. (1988). Manual for the self-perception profile for adolescents. Denver, CO: University of Denver. Heck, K. E., & Subramaniam, A. (2009). Youth development frameworks. Davis, CA: 4-H Center for Youth Development, University of California [Monograph]. Holsen, I., Geldhof, G. J., Larsen, T., & Aardal, E. (2017). The five Cs of positive youth development in Norway: Assessment and associations with positive and negative outcomes. International Journal of Behavioral Development, 41, 559–569. Jeličić, H., Bobek, D., Phelps, E.,D., Lerner, J. V., & Lerner, R. M. (2007). Using positive youth development to predict contribution and risk behaviors in early adolescence: Findings from the first two waves of the 4-H Study of Positive Youth Development. International Journal of Behavioral Development, 31(3), 263–273. King, P. E., Dowling, E. M., Mueller, R. A., White, K., Schultz, W., Osborn, P., et al. (2005). Thriving in Adolescence: The voices of youth-serving practitioners, parents, and early and late adolescents. The Journal of Early Adolescence, 25(1), 94–112. Larsen, T., Urke, H. B., Holsen, I., Anvik, C. H., Olsen, T., Waldahl, R. H., et al. (2018). COMPLETE – a school-based intervention project to increase completion of upper secondary school in Norway: Study protocol for a cluster randomized controlled trial. BMC Public Health, 18(1), 340. https://doi.org/10.1186/s12889-0185241-z. Larson, R. W. (2011). Positive development in a disorderly world. Journal of Research on Adolescence, 21(2), 317–334. Lee, I. A., & Little, T. D. (2012). P-technique factor analysis. In B. Laursen, T. D. Little, & N. A. Card (Eds.). Handbook of developmental research methods (pp. 350–362).

8

Journal of Adolescence 71 (2019) 1–9

G.J. Geldhof et al.

New York: Guilford. Lerner, J. V., Phelps, E., Forman, Y., & Bowers, E. (2009). Positive youth development. In (3rd ed.). R. M. Lerner, & L. Steinberg (Vol. Eds.), Handbook of adolescent psychology: Vol. 1, (pp. 524–558). Hoboken, NJ: Wiley. Leffert, N., Benson, P. L., Scales, P. C., Sharma, A. R., Drake, D. R., & Blyth, D. A. (1998). Developmental assets: Measurement and prediction of risk behaviors among adolescents. Applied Developmental Science, 2(4), 209–230. Lerner, R. M. (1998). Theories of human development: Contemporary perspectives. In W. Damon, R. M. Lerner, & Vol (Eds.). Handbook of child psychology: Vol. 1. Theoretical models of human development (pp. 1–24). New York: Wiley. Lerner, R. M. (2015). Promoting positive human development and social justice: Integrating theory, research and application in contemporary developmental science. International Journal of Psychology, 50(3), 165–173. Lerner, J. V., Bowers, E., Minor, K., Boyd, M., Mueller, M., Schmid, K., et al. (2012). Positive youth development: Processes, philosophies, and programs. Handbook of Psychology, 6, 365–392. 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(3), 172–180. Lerner, R. M., Freund, A. M., De Stefanis, I., & Habermas, T. (2001). Understanding developmental regulation in adolescence: The use of the selection, optimization, and compensation model. Human Development, 44(1), 29–50. Lerner, R. M., Lerner, J. V., Almerigi, J., Theokas, C., Phelps, E., Gestsdóttir, S., et al. (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. The Journal of Early Adolescence, 25(1), 17–71. Lerner, R. M., Lerner, J. V., Almerigi, J., Theokas, C., Phelps, E., Naudeau, S., et al. (2006). Towards a new vision and vocabulary about adolescence: Theoretical, empirical, and applied bases of a “Positive Youth Development” perspective. In L. Balter, & C. S. Tamis-LeMonda (Eds.). Child psychology: A handbook of contemporary issues (pp. 445–469). New York: Psychology Press/Taylor & Francis. Lerner, R. M., Lerner, J. V., Bowers, E., & Geldhof, G. J. (2015). Positive youth development and relational developmental systems. In W. F. Overton, & P. C. Molenaar (Eds.). Theory and Method, Volume 1 of the Handbook of Child Psychology and Developmental Science (pp. 607–651). (7th ed.). Hoboken, NJ: Wiley Editor-in-chief: R. M. Lerner. 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. Overton, W. F. (2010). Life-span development: Concepts and issues. In R. M. Lerner, W. F. Overton, & Vol (Eds.). Handbook of life-span development: Vol. 1. Cognition, biology,and methods across the lifespan (pp. 1–29). Hoboken, NJ: Wiley. Overton, W. F. (2015). Processes, relations, and relational‐developmental‐systems. In W. F. Overton, P. C. M. Molenaar, Volume, & R. M. Lerner (Eds.). Handbook of child psychology and developmental science (pp. 9–62). (7th ed.). Hoboken, NJ: Wiley. Phelps, E., Zimmerman, S., Warren, A. E. A., Jeličič, H., von Eye, A., & Lerner, R. M. (2009). The structure and developmental course of Positive Youth Development (PYD) in early adolescence: Implications for theory and practice. Journal of Applied Developmental Psychology, 30(5), 571–584. Piko, B. F., & Pinczés, T. (2014). Impulsivity, depression and aggression among adolescents. Personality and Individual Differences, 69, 33–37. Smith, R. L. (2015). Adolescents' emotional engagement in friends' problems and joys: Associations of empathetic distress and empathetic joy with friendship quality, depression, and anxiety. Journal of Adolescence, 45, 103–111. Tambs, K., & Moum, T. (1993). How well can a few questionnaire items indicate anxiety and depression? Acta Psychiatrica Scandinavica, 87(5), 364–367. Smith, R. L., & Rose, A. J. (2011). The “cost of caring” in youths' friendships: Considering associations among social perspective taking, co-rumination, and empathetic distress. Developmental Psychology, 47(6), 1792. Tone, E. B., & Tully, E. C. (2014). Empathy as a “risky strength”: A multilevel examination of empathy and risk for internalizing disorders. Development and Psychopathology, 26(4pt2), 1547–1565. Truskauskaitė-Kunevičienė, I., Žukauskainė, R., & Kaniušonytė, G. (2014). Positive youth development links to satisfaction with life, resilience and internalizing and Externalizing Problems. Social Work: Research Journal, 13(1), 98–109. Van Lissa, C. J., Hawk, S. T., Koot, H. M., Branje, S., & Meeus, W. H. (2017). The cost of empathy: Parent–adolescent conflict predicts emotion dysregulation for highly empathic youth. Developmental Psychology, 53(9), 1722.

9