Abstract
Adolescent cognitive and behavioral regulation is influenced by multidimensional and multidirectional processes within and across biological and contextual systems that shift throughout development. Key among these influences are distal processes such as early life socioeconomic position (SEP), and proximal processes such as pubertal development, but questions remain concerning how links between SEP, pubertal development, and cognitive and behavioral regulation accumulate and unfold over adolescence. In the current study, and in line with Dr. John Schulenberg's foundational work, direct associations between SEP, puberty, and adolescent cognitive and behavioral regulation were examined; then pubertal timing and tempo were considered as moderators and mediators of links between SEP and adolescent cognitive and behavioral regulation. Data were drawn from the NICHD Study of Early Child Care and Youth Development, a longitudinal study of 970 youth (52% male; 80% White, 13% Black, and 7% another race/ ethnicity). Cognitive and behavioral regulation was measured using direct assessments of working memory, planning, risky decision-making, and impulse control at age 15. SEP included maternal education and family income-to-needs and was averaged from birth to 54months old; estimates of pubertal timing and tempo were derived using logistic growth curve models from age 9 to age 15. SEP was directly associated with cognitive and behavioral regulation. Pubertal development tended to moderate those links, but rarely mediated them. Specifically, socioeconomic disadvantage along with earlier timing or faster tempo tended to be associated with worse cognitive and behavioral regulation. Overall, findings suggest that pubertal timing and tempo may exacerbate existing environmental constraints.
The ability to reason logically and abstractly and to regulate one's cognitions and emotions is the hallmark of adolescence, but it emerges slowly (Dahl et al., 2018; Steinberg, 2005) and demonstrates significant individual differences (Carlson et al., 2008; Schulenberg & Maslowsky, 2015). This may because cognitive and behavioral regulation is influenced by multidimensional and multidirectional processes within and across biological and contextual systems that shift throughout development (Ahmed, Chaku, et al., 2023; Graber & Brooks-Gunn, 1996; Schulenberg et al., 2018). Some of these influences are more distal but persist over time, such as socioeconomic position (SEP), or the socially derived economic factors that influence the positions that individuals hold in society. Other influences are more proximal and emerge later during biological transitions such as puberty.
SEP (e.g., parent education, family income) is a robust predictor of adolescent adjustment; its influence spans multiple levels of development and accumulates over time (Ahmed et al., 2021; Hoyt et al., 2019; Lawson et al., 2018). Similarly, puberty, and especially its timing (onset compared to same-aged, same-sexed peers) or tempo (how fast one progresses through puberty), has long been linked to cognitive outcomes both directly and via biological and social mechanisms (Chaku & Hoyt, 2019; Stumper et al., 2020). However, much less is known about the extent to which these processes unfold and influence each other throughout development. Additionally, it remains unclear how these distal (SEP) and proximal (puberty) influences interact over time to affect adolescent development. In the current study, and in line with Dr. John Schulenberg's legacy, we explored these questions using three different theoretical perspectives on the role of puberty in the lifespan.
The importance of adolescent cognitive and behavioral regulation
Adolescent cognitive and behavioral regulation comprises many cognitive skills involved in the deliberate control of goal-oriented behavior such as simple ‘executive functions’ (i.e., inhibition, working memory, and cognitive flexibility; Miyake et al., 2000) as well as complex cognitive processes such as planning, problem-solving, and the top-down regulation of impulsive or risk-taking behaviors (Doebel, 2020; Nigg, 2017; Perone et al., 2021). Together, these cognitive and behavioral skills are broadly thought to be involved in the control and regulation of attentional and memory systems, facilitate the storage and processing of information, and work together to flexibly adapt to environmental demands (Karr et al., 2018; Rothbart & Posner, 1985).
Cognitive and behavioral regulation continues to develop over childhood and adolescence, and there is particular interest in its development during adolescence for several reasons (Best & Miller, 2010). First, adolescence marks an important period of accelerated cognitive development (Blakemore & Choudhury, 2006; Zelazo & Carlson, 2012). For example, recent studies (including those using nationally representative samples) find rapid growth in working memory and inhibitory control across adolescence relative to other developmental periods (Meisel et al., 2019; Myers et al., 2022); other studies find variability in the rate of cognitive development, highlighting that youth may develop different cognitive and regulatory abilities at different rates or different points in their development (Boelema et al., 2014; Chaku et al., 2021; Lensing & Elsner, 2018). This suggests that adolescence may be a formative period of the lifespan to investigate cognitive and behavioral regulation and its progression.
Second, adolescent cognitive and behavioral regulation has been shown to be a robust predictor of adult outcomes. Indeed, although most work conducted to date has only considered early cognitive and behavioral regulation, often finding that early cognitive skills such as working memory and delay of gratification demonstrate robust associations with educational outcomes (among others; see Cortés Pascual et al., 2019 for review), recent work highlights the importance of considering adolescent cognitive and behavioral regulation as well (Ahmed, Kelly, et al., 2023). These proximal cognitive processes may be uniquely predictive of later outcomes. Indeed, early adolescent cognitive and behavioral regulation predicts educational attainment, occupation and career success during early and middle adulthood (Ahmed et al., 2021; McClelland et al., 2013; Spengler et al., 2018) as well as future health and wellbeing even after controlling for early life characteristics (Huang et al., 2022; Poon, 2017). Thus, adolescent cognitive and behavioral regulation may be a unique predictor of future outcomes that could be considered with and separately from earlier life skills.
Socioeconomic position and executive function
Given the evidence that cognitive and behavioral regulation is a unique and prominent predictor of numerous life outcomes in adolescence and beyond, many researchers have sought to identify aspects of the social context that serve to support or hinder its development (see Hackman et al., 2010 for review). One candidate factor that has received a great deal of attention in the literature is SEP. Indeed, a large body of research has documented associations between SEP and family-level factors, such as parental beliefs and behaviors, as well as neighborhood and community characteristics, like school quality and neighborhood disadvantage (Conger & Donnellan, 2007; Davis-Kean et al., 2019; Evans, 2004). Each of these factors has been implicated, to varying degrees, in children's cognitive and regulatory development (DavisKean et al., 2021; Hyde et al., 2020). As such, SEP can be thought of as a valuable starting point for identifying aspects of the social context that may be influential in shaping cognitive development (Waters et al., 2021).
A growing body of research has documented associations between indicators of SEP and children's cognitive and behavioral regulation in early and middle childhood (see Lawson et al., 2018 for review). For example, Cuartas et al. (2022) observed SEP-related differences of 0.22–0.32 SD for self-regulation skills in early childhood, and Albert et al. (2020) observed a difference of 0.37 SD between SEP and self-regulation skills in middle childhood. However, much remains to be learned about whether and how SEP might relate to cognitive and behavioral regulation during adolescence. Indeed, adolescence is a sensitive period for cognitive development (Ahmed, Kelly, et al., 2023; Best & Miller, 2010; Luna et al., 2010) as well as for increases in self-regulation (Steinberg et al., 2008), suggesting that these skills might be particularly susceptible to contextual influences throughout this period (Schulenberg & Maslowsky, 2015).
The influence of SEP on adolescent cognitive and behavioral regulation could take several forms. First, it is possible that cognitive and behavioral regulation disparities observed in early childhood neither grow nor diminish over time, suggesting that early life processes may have lasting consequences for cognitive development. Second,it is plausible that SEP-related disparities in cognitive and behavioral regulation may widen from early childhood to adolescence as the persistent effects of socioeconomic disadvantage accumulate over time. These perspectives would emphasize the continuity of childhood characteristics over more proximal processes (see Schulenberg & Zarrett, 2006). Alternatively, though, SEP-related disparities could potentially narrow or be altered by adolescent experiences as children from lower-SEP backgrounds “catch up” to their more advantaged peers, move to more advantaged contexts, or experience common developmental transitions. This would emphasize the discontinuity of developmental experiences from childhood to adolescence (Schulenberg et al., 2003). Albeit limited, there is some empirical support for stability (e.g., Hackman et al., 2014), widening (e.g., Spielberg et al., 2015), and narrowing (e.g., Boelema et al., 2014) of SEP-related disparities in adolescent cognitive and behavioral regulation, highlighting the need for more research in this area.
The role of puberty
Puberty marks the biological entrance into adolescence. Its timing, or onset compared to same-sex and same-aged peers, and to some extent, its tempo or progression through pubertal stages, has been linked with both positive and negative outcomes. Specifically, youth with earlier timing and faster tempo tend to report more depression and anxiety as well as more externalizing problems across adolescence (Dorn & Beltz, 2023; Hoyt et al., 2019; Ullsperger & Nikolas, 2017). Less work has investigated pubertal development and cognition, but there is some evidence that earlier pubertal timing is associated with better cognitive outcomes, such as higher three-dimensional mental rotation performance in boys only (Beltz & Berenbaum, 2013) and faster growth in attention skills across genders (Chaku & Hoyt, 2019). Similarly, in academic settings, evidence suggests that early-developing youth demonstrate larger gains in mathematics and reading compared to on-time and late-maturing youth (Koerselman & Pekkarinen, 2018). Links are inconsistent though, with some evidence that earlier (or off-time) development is associated with poor initial EF (Chaku & Hoyt, 2019; Stumper et al., 2020) and worse concurrent and future academic performance (Cavanagh et al., 2007; Goering & Mrug, 2022). Currently, there is little evidence regarding the role of pubertal tempo, potentially because its measurement must be longitudinal and there is little consensus on how to best measure or conceptualize it (Mendle, 2014).
How links between puberty and cognitive and behavioral development unfold over time is an open question. Earlier pubertal timing and faster tempo might have direct effects on cognitive development. The hormonal influence hypothesis suggests that hormonal changes (specifically increases in testosterone, estradiol, and DHEA) at the onset of puberty are associated with changes in brain development and thus, cognitive performance (Beltz & Berenbaum, 2013; Laube & Fuhrmann, 2020). Specifically, pubertal hormones contribute to the reorganization of the prefrontal cortex, which is implicated in the ongoing development of cognitive and behavioral regulation (Drzewiecki & Juraska, 2020; Goddings et al., 2019). The hormonal changes of puberty may also operate indirectly via social experiences. Early developing youth are at risk for a host of negative social experiences (e.g., peer victimization, affiliation with deviant peers) that are associated with increased risk-taking behavior as well as more mental health problems such as depression and anxiety (Smith et al., 2013; Stattin et al., 2011). These negative social experiences can disrupt the development of cognitive and behavioral regulation (Pfeifer & Allen, 2021); further, their subsequent outcomes, such as substance use and depression are associated with worse cognitive and behavioral regulation as well (Gustavson et al., 2019; Kim-Spoon et al., 2017).
Early pubertal timing and faster tempo might also amplify environmental effects, and especially the effects of, for example, SEP on cognitive and behavioral regulation. Indeed, the contextual amplification hypothesis suggests that earlier timing or faster progression through puberty may exacerbate the effect of current negative environmental conditions on outcomes. This has mostly been tested within contexts that youth are directly or frequently in contact with, such as family, peers, or school systems (Ge & Natsuaki, 2009; Skoog & Kapetanovic, 2022). Thus, lower maternal education or lower family income and early pubertal timing or faster pubertal tempo may be associated with worse cognitive and behavioral regulation than just lower SEP, early pubertal timing, or pubertal tempo alone. Some work on internalizing and externalizing behaviors in adolescence bears this out (see Vijayakumar et al., 2018 for review), finding primarily that higher neighborhood income and early development were associated with more internalizing behaviors in girls (Deardorff et al., 2021) and more externalizing behaviors in boys (Niu et al., 2023), but this work has not yet been extended to cognitive or other behavioral outcomes.
Alternatively, early pubertal timing may also mediate links between SEP and cognitive and behavioral regulation. There is some evidence that harsh early environments, specifically maternal harshness, father absence, and adversity, predict earlier pubertal onset (Pham et al., 2022) and potentially faster pubertal tempo (Belsky et al., 2007). Evolutionary theories suggest that these early life insults are a symptom of scarce environmental resources, setting a biological imperative for youth to reach maturation earlier and thus reproduce earlier (Belsky et al., 2007; Hamlat et al., 2021). Alternatively, these early life insults may dysregulate the biological systems underpinning stress responses such as the hypothalamic–pituitary–adrenal or HPA axis, which itself plays a role in earlier pubertal timing (King et al., 2017; Pham et al., 2022). Several studies have borne out these hypotheses, finding that in addition to family-level adversities, such as maternal psychopathology, SES also predicts earlier pubertal timing (Obeidallah et al., 2004; Oelkers et al., 2021; Sun et al., 2017). One study even found that youth from low socioeconomic backgrounds (derived from maternal education, household income, and free/reduced lunch program eligibility) predicted earlier pubertal timing which in turn predicted worse performance on attention and other cognitive tasks (Stumper et al., 2020).
CURRENT STUDY
In the present study, we explored links between SEP, pubertal development, and cognitive and behavioral regulation using the Study of Early Childcare and Youth Development (SECCYD). In line with Dr. Schulenberg's scholarship on proximal effects (Maggs & Schulenberg, 2005), developmental turning points (Schulenberg et al., 2004), and discontinuity (Schulenberg et al., 2003), we tested three competing theoretical perspectives that assessed whether proximal effects (puberty) negated, reversed, or enhanced more distal effects (socioeconomic disadvantage) on adolescent development (Schulenberg et al., 2004). This theoretical lens was based on Dr. Schulenberg's research on developmental turning points and developmental disturbances, whereby proximal effects (puberty) can counteract developmentally distal effects (socioeconomic disadvantage) and prompt discontinuity in development (Schulenberg et al., 2018).
First, we examined direct links between SEP, puberty, and cognitive and behavioral regulation (Laube & Fuhrmann, 2020); this perspective emphasizes that puberty could act as a developmental turning point that may fundamentally alter behavior. Second, we examined whether pubertal development amplified or exacerbated the negative effects of early life socioeconomic disadvantage on adolescent outcomes (Ge et al., 2002); this perspective emphasizes the discontinuity of adolescent experiences or that puberty may provoke short (or long)-term instability in behavior. Third, we examined whether puberty operated as a biological pathway linking SEP to adolescent outcomes (Dorn et al., 2019); this perspective emphasizes the continuity of experiences; that is, pubertal onset and its progression contains echoes of one's context and past and could mediate links between SEP and later outcomes.
We hypothesized that higher SEP would be related to better cognitive and behavioral regulation (Duncan & Magnuson, 2012; Feinstein & Bynner, 2004; Waters et al., 2021). We additionally hypothesized that earlier pubertal timing would be related to worse cognitive and behavioral regulation, and that this effect would be larger for regulatory abilities than for cognitive abilities (Chaku & Hoyt, 2019; Laube & Fuhrmann, 2020). We expected that earlier pubertal timing would amplify the effects of lower SEP on cognitive and behavioral regulation, but that effects might differ across sexes (Chen & Raine, 2018; Shelton & Van Den Bree, 2010). We also expected that lower SEP would be associated with earlier timing and that this association would partially mediate links between SEP and cognitive and behavioral regulation (Stumper et al., 2020). Given the paucity of available data for pubertal tempo, we made no hypotheses related to the effects of pubertal tempo on cognitive and behavioral regulation nor on its moderating or mediating effects.
METHOD
Data were drawn from the National Institute of Child Health and Development Study of Early Child Care and Youth Development (NICHD SECCYD), a longitudinal study of childcare experiences, characteristics, and developmental outcomes from birth to young adulthood (https://www. icpsr.umich.edu/web/ICPSR/series/00233). Data were collected from 10 university hospital sites around the country. Additional details on study recruitment and data collection have been extensively reported on in previous studies (Bleil et al., 2017; Marceau et al., 2011, 2015; Pianta et al., 2002; Vandell et al., 2010) and can be obtained from the Early Child Care Research Network (http://www.nichd.nih.gov/ research/supported/seccyd/overview.cfm). This study was considered exempt by the institutional review board at Indiana University.
Participants
The initial sample included 1364 children assessed at birth, and then periodically until age 15. The initial sample was 51.7% male and 80% White, 13% Black, and 7% another race/ ethnicity. All youth who completed at least one cognitive and behavioral regulation outcome at age 15 were included in subsequent analyses (N=970). Thus, the final analytic sample was 49.5% male and 81.3% White, 11.9% Black, and 6.8% another race/ethnicity. Youth who were included in the analyses tended to come from families with higher mother education, t(1361)=4.63, p<.001 and from families with higher incomes, t(1353)=3.52, p<.001, but were not more likely to be White (p>.05) than those excluded from the analyses. Those included in the analyses were also not more likely to differ in their estimates of pubertal timing or tempo (all ps>.05) than those excluded from the analyses.
Measures
Cognitive and behavioral regulation
Working memory was assessed using the Operation Span task (OSPAN; Turner & Engle, 1989). In this task, which taps aspects of storage, updating, and retrieval, adolescents recall a series of letters between completing a set of calculations. Working memory was measured as the cumulative total memory score, with higher scores suggesting greater working memory (M = 32.39, SD = 17.09). Planning was measured using the Tower of London assessment (Berg & Byrd, 2002). In this task, adolescents were asked to move three balls on a computer screen to match a configuration shown on the computer screen in the fewest number of moves. The total number of trials solved was used to measure adolescents' planning skills (M = 10.59, SD = 2.75).
SEP
SEP included maternal education and family income-toneeds. Maternal education represented the total years of education completed by the participating child's mother and was measured when the child was 1 month old (M = 14.23 years, SD = 2.51). Income-to-needs was defined as the average from when the child was 1 month old to 9.5 years old (M = 3.55, SD = 2.78). An income-to-needs ratio of less than or equal to 1 indicates that a family is living in poverty (U.S. Bureau of the Census, 2023). These items were standardized and then averaged to create a composite measure of SEP.
Pubertal timing
Pubertal development was assessed via clinical exams, commonly known as Tanner Staging (Marshall & Tanner, 1969, 1970), a well-known and validated measure of pubertal status. Exams were completed by a trained doctor or nurse practitioner. They were conducted yearly starting at age 9.5 and continued until age 15l; they included measures of breast development (girls), genital development (boys), and pubic hair development (both). Youth were rated on a fivepoint scale where 1=prepubertal and 5=sexually mature. Exams continued until youth had reached Tanner Stage 5 (or sexually mature) on both indices. Thus, missing data after this point was recorded as Tanner Stage 5 following previous research (Chaku & Hoyt, 2019; Marceau et al., 2011).
Puberty data was first examined to ensure that only reliable and valid data were included. Following previous research (Beltz et al., 2014), only adolescents with at least two, increasing assessments were included in subsequent analyses; this removed 1% of all girls and 3% of all boys. Estimates of pubertal timing and tempo were derived using logistic growth curve models of pubic hair development in MPlus (following Beltz et al., 2014; Marceau et al., 2011). Thus, timing (i.e., the intercept) was conceptualized as age at the midpoint of puberty, or Tanner Stage 3, and higher scores indicated later timing or older age at Tanner Stage 3; tempo (i.e., the slope) was conceptualized as the rate of development at the midpoint of puberty, and higher scores indicated faster tempo, or progressing through more pubertal stages per year at the midpoint of puberty. Thus, a girl with an intercept of 12.45 and a slope of 1.10 would have reached Tanner Stage 3 at age 12.45 and would progress through approximately 1.10 (Tanner) stages at that age.
Analytic plan
Three sets of models were run in Mplus 8.0 (Muthén & Muthén, 1998-2011) using full information maximum likelihood estimation (FIML) to account for missing data, and the COMPLEX command to account for clustering by site. The first set of models assessed direct links between SEP, pubertal timing and tempo, and cognitive and behavioral regulation (working memory, planning; risk-taking propensity, impulse control). The second set of models assessed whether pubertal timing or tempo moderated links between SES and cognitive outcomes. Timing and tempo moderation effects were estimated in the same model, and significant interactions were probed at one standard deviation above and below the mean in follow-up analyses. The final set of models assessed whether pubertal timing or tempo mediated links between SES and cognitive outcomes. Timing and tempo mediation effects were estimated in the same model using 5000 bootstrapped samples. In all models, logistic growth curves of pubertal timing and tempo were estimated at the same time as the direct, moderating, and mediating effects.
Models were run separately by gender (girls, boys) and included White race/ethnicity as a dummy-coded covariate. This is common in puberty research given well-known differences in the onset and progression of puberty by gender and by race/ethnicity (Dorn & Beltz, 2023). Given the large number of models run, all p values were corrected using the Benjamini-Hochberg procedure (Noble, 2009). Analytic code is provided in the supplemental materials.
RESULTS
Descriptive statistics and a correlation table of all focal variables are presented in Table 1 by sex. Pubertal timing and tempo estimates were extracted from latent models for the correlational analyses only; raw values were used to increase interpretability. Income and mother's education are presented for interpretability (instead of SEP). There were some sex differences: First, although boys and girls did not differ in their working memory, planning, or impulse control, girls did demonstrate lower risk-taking propensity, t(953) = −4.71, p < .001. Second, as expected, girls had earlier timing (i.e., entered Tanner Stage 3, the midpoint of puberty) than boys, t(961) = −14.41, p < .001. Boys tended to experience faster tempo though, progressing through 1.00 pubertal stages per year compared to girls who progressed through .84 pubertal stages per year, t(961) = −10.44, p < .001. Gender differences in timing and tempo are similar to those reported in previous research using this dataset (Marceau et al., 2014). There were no significant differences in maternal education or incometo-needs by sex.
The correlational analyses suggested there were small positive associations between cognitive and behavioral regulations ( r s = .004–.30). As expected, working memory and planning were more likely to be positively correlated with each other than with risk-taking propensity or im - pulse control. This was true for both sexes. Maternal edu - cation and income-to-needs demonstrated small, positive associations ( r s = .02–.25) with most outcomes, but tended to be larger for working memory and planning ( r s = .07– .25) than for risk-taking propensity or impulse control (rs = .01–.07). This was also true for both sexes. In girls, later pubertal timing was significantly associated with better impulse control ( r = .10), higher maternal educa - tion ( r = .10), and higher income-to-needs ( r = .09), but not with pubertal tempo. In boys, later pubertal timing was significantly associated with White race/ethnicity ( r = .21) and higher income-to-needs ( r = .14), but again not with pubertal tempo.
The results for the focal analyses are presented in Table 2 for girls and Table 3 for boys. For each outcome, the main effects (of SEP and pubertal timing and tempo) are presented first, followed by the moderation model, and then the mediation model. Standardized betas, standard errors, and 95% confidence intervals for presented for each model.
Main effects models
For girls, higher SEP was associated with lower risk-taking propensity ( β =−.21, SE =.05, p <.01), better working memory ( β =.11, SE =.11, p <.05), and better planning ( β =.23, SE =.12, p <.001), but was not significantly associated with impulse control at age 15. Neither pubertal timing nor pubertal tempo were associated with cognitive or behavioral regulation. For boys, higher SEP was only associated with better planning (β=.41, SE=.10, p<.001) only, and similarly, pubertal timing and tempo were not associated with cognitive or behavioral TABLE 1 regulation. Overall, this suggests a direct effect of SEP on adolescent cognitive or behavioral regulation, especially for girls, but limited direct effects on pubertal timing or tempo. All models were run separately by maternal education and income-to-needs (see Tables S1 and S2); results are generally similar to those reported here.
Moderation models
For girls, several significant moderation effects emerged. First, pubertal timing moderated the effect of SEP on planning (β=−.11, SE=.04, p<.001). Probing the interaction at one standard deviation above and below the mean suggested that girls who came from families with lower SEP who also experienced early puberty had worse planning skills (b=−.41, SE=.11, p<.001) than those from families with higher SEP. There were several additional significant interaction effects for pubertal tempo. Specifically, pubertal tempo moderated the effect of SEP on impulse control (b=.10, SE=.04, p<.05), OSPAN (b=.11, SE=.04, p<.001), and working memory (b=.11, SE=.03, p<.001). In all of these cases, probing the interaction at one standard deviation above and below the mean suggested that girls from lower SEP households who also experienced faster pubertal tempo tended to demonstrate worse cognitive and behavioral regulation (e.g., working memory: b=−3.69, SE=1.12, p<.001) whereas girls from lower SEP households who experienced slower pubertal tempo tended to demonstrate better cognitive and behavioral regulation (e.g., working memory: b=5.50, SE=1.55, p<.001).
For boys, two significant moderation effects emerged. Specifically, pubertal timing moderated the effect of SEP on working memory (b=−.21, SE=.08, p<.001) and planning (b=−.26, SE=.01, p<.001). In all these cases, probing the interaction at one standard deviation above and below the mean suggested that boys from lower SEP households who also experienced earlier timing tended to demonstrate worse cognitive performance (e.g., working memory: b=−1.66, SE=.63, p=.01). Overall, this suggests that faster pubertal tempo in girls and earlier pubertal timing in boys may exacerbate links between SEP and adolescent cognitive and behavioral regulation. All models were run separately by maternal education and income-to-needs (see Tables S1 and S2); results are generally similar to those reported here.
Mediation models
In the mediation models, one significant effect emerged. In boys only, the effect of SEP on working memory was partially mediated by pubertal tempo (total indirect effect: b=−.05, SE=.03, p=.05). Specifically, higher SEP was associated with slower pubertal tempo, b=−.48, SE=.09, p<.001, which in turn was associated with better impulse control, b=−.04, SE=.02, p=.04. There were no other significant indirect effects of timing or tempo on cognitive and behavioral regulation for either sex. Again, all models were run separately by maternal education and income-to-needs (see Tables S1 and S2); results are generally similar to those reported here.
DISCUSSION
Adolescent cognitive and behavioral regulation predicts a host of future developmental outcomes including educational attainment, health, and wellbeing (Ahmed et al., 2021; Poon, 2017). SEP and pubertal development may play a role in the on-going development and maturation of these skills (Chaku & Hoyt, 2019; Hackman et al., 2010; Stumper et al., 2020), but there is little existing evidence as to how their influences accumulate, interact, and affect future cognitive and behavioral regulation. In the current study, distal processes such as early-life SEP and more proximal processes such as pubertal development were first considered as direct predictors of adolescent cognitive and behavioral regulation. Then, drawing on theoretical perspectives on adolescent development (Falk et al., 2013) and the role of puberty in the lifespan (Dorn & Beltz, 2023; Ge et al., 2002; Schulenberg & Maslowsky, 2015), pubertal development was additionally considered as a moderator and as a mediator of links between SEP and cognitive and behavioral regulation. Although effects were disparate, several consistent findings emerged that offer some directions for future research.
SEP but not puberty was directly associated with cognitive and behavioral regulation in adolescence
The direct effects of SEP differed by outcome and by gender. For girls, SEP had small, positive associations with most cognitive and behavioral regulation outcomes; for boys, this was only true for planning abilities. This was somewhat unexpected. SEP has long been hypothesized to have direct effects on cognitive and behavioral regulation; this effect has been evident in many previous cross-sectional and longitudinal studies (Duncan & Magnuson, 2012; Feinstein & Bynner, 2004; Hoyt et al., 2019; Waters et al., 2021). Thus, SEP effects might have been expected to be more robust (across both genders) than seen here. It may be that behavioral regulation measures such as the BART and Stoplight task may not be as reliable as the EF measures which have a large and robust literature of support (Karr et al., 2018; McCoy, 2019; Nyongesa et al., 2019). Alternatively, meta-analyses suggest that parenting practices, chaotic or disruptive home environments, and intelligence may mediate relations between early life SES and cognitive ability (Buckley et al., 2019; Letourneau et al., 2013). In some ways, adolescents are less exposed to these (potentially chaotic) home environments, spending more time in school or with peers (Ahmed, Kelly, et al., 2023); but they are also more exposed to their neighborhoods which carries their own unique (but unmeasured) risks (Molina-García & Queralt, 2017). Both explanations warrant further research.
It was also surprising that SEP was more robustly associated with outcomes for girls and not for boys, especially since SEP mechanisms might be thought to operate similarly on cognitive and behavioral regulation across genders (e.g., von Stumm et al., 2020). Many sex differences do first emerge in adolescence though, and there is some evidence that SEP effects on cognition might be more evident for girls (see Spielberg et al., 2015). It may also be that distinct aspects of SEP are associated with outcomes differently for girls than boys. For example, parents with lower educational status may engage in more conflict or provide fewer learning opportunities in the home (Weaver et al., 2015); adolescent girls who tend to be in the home more than boys may be disproportionately affected by these experiences (Lloyd et al., 2008; Varner & Mandara, 2014). Comparatively, lower income is typically associated with worse neighborhood quality (Hoyt et al., 2019; Niu et al., 2023) and adolescent boys tend to spend more time outside the home (Varner & Mandara, 2014; Worthen, 2012), potentially exposing them to more detrimental effects via neighborhood-level mechanisms (Niu et al., 2023).
Together, these findings suggest several key next steps. First, instead of averaging maternal education and incometo-needs together and across time, future research could test these associations with longitudinal measures of SEP that better account for the ways that SEP may dynamically fluctuate across time and have differential impacts on outcomes at different points of development (Ahmed, Chaku, et al., 2023; Falk et al., 2013; Hoyt et al., 2019). Second, future research could also test and incorporate distinct aspects of SEP (e.g., neighborhood, context, the built environment, subjective SEP) that may hold additional predictive value for understanding adolescent cognitive and behavioral regulation (Cirino et al., 2002; Conway et al., 2019). Finally, it would be important to connect adolescent cognitive and behavioral with later life outcomes, including SEP (e.g., Zhang et al., 2020). This would be critical to understand the role of adolescent behaviors on the lifespan (Falk et al., 2013).
Pubertal development moderates (but does not mediate) the relationship between SEP and adolescent cognitive and behavioral regulation
Pubertal development was not directly related with cognitive and behavioral regulation at age 15, but pubertal timing and tempo did moderate the effect of SEP on later outcomes in adolescence. More socioeconomic disadvantage along with earlier timing or faster tempo was associated with worse cognitive and behavioral regulation. Significant associations were evident for tasks that measured executive functions and those that measured behavioral regulation (albeit more consistently for executive function tasks). These findings add to a small, but growing literature suggesting that pubertal timing can increase risk for maladaptation (Beltz et al., 2014; Marceau et al., 2012; Tsai et al., 2014), extends findings to pubertal tempo (Dorn & Beltz, 2023; Mendle & Koch, 2019), and provides some initial evidence that links these effects to cognitive and behavioral regulation.
In all but one case, pubertal development did not mediate associations between SEP and cognitive and behavioral regulation. This is in contrast to one recent study that found that early pubertal timing mediated associations between low SEP (family income, maternal education, and eligibility for free or reduced-price lunch) and working memory (Stumper et al., 2020), but this study used longitudinal measurement of puberty and used different measures of working memory (e.g., digit span versus OSPAN), which of which may have contributed to differences in outcomes. Overall, the present findings are in line with the contextual amplification hypothesis which suggests that pubertal timing and tempo may exacerbate existing environmental constraints (Ge & Natsuaki, 2009; Vijayakumar et al., 2018), but did not provide much support for the hormonal influence or accentuation hypothesis (Laube & Fuhrmann, 2020).
There was, again, some evidence of gender differences. Specifically, for girls, the most significant interactions involved tempo (e.g., faster tempo and lower SEP were associated with lower working memory), whereas for boys, the most significant interactions involved timing (e.g., earlier timing and lower SEP were associated with more risk-taking propensity). This was somewhat surprising. First, girls' pubertal timing, which is well-studied (Mendle et al., 2007), did not generally moderate SEP but boys' pubertal timing did. Although early pubertal timing is not often considered negative for boys, recent evidence does suggest that early developing boys also experience elevated rates of anxiety and depression (Ullsperger & Nikolas, 2017). Extending this to cognitive and behavioral regulation, it may be that early pubertal development in boys is especially detrimental in disadvantaged contexts which may expose these young boys to social and neighborhood experiences they cannot yet handle. For girls, however, early pubertal development is universally associated with negative outcomes (Mendle et al., 2007; Ullsperger & Nikolas, 2017); thus, it may be that disadvantaged contexts do not exacerbate the effects of pubertal timing for them. There is limited research on tempo, and virtually no studies examining it in relation to cognitive and behavioral regulation. What exists suggests that faster tempo is detrimental for mental health and wellbeing (Mendle et al., 2010). We demonstrate that this may extend to cognitive and behavioral regulation as well, but only for girls. Clearly, more work is needed to understand how different aspects of puberty may be important for different genders.
These findings again point to several key areas for future research. For example, future research could consider SEP and puberty as part of a larger developmental cascade, looking at how early-life SEP and adolescent-SEP are related to pubertal development, the social context in which puberty unfolds, and then later cognitive and behavioral regulation. Another important future direction of this research involves understanding whether pubertal timing and tempo predict long-term adult outcomes. Although recent research has linked pubertal timing to a host of physical and health outcomes across adulthood (see Copeland et al., 2010; Hoyt et al., 2020 for review), few studies have examined how puberty relates to cognitive and behavioral outcomes during adulthood. Additionally, many of the existing studies in the literature focus solely on pubertal timing (e.g., early, typical, delayed), whereas fewer studies have investigated the role of pubertal tempo for short and long-term developmental outcomes. Future research would benefit from understanding how puberty moderates other developmental risk and protective factors to shape adolescent development and adult outcomes.
Limitations
There are several limitations of the present study that are important to consider. First, although the SECCYD is a national sample, it is not representative of the U.S. population, which might limit the generalizability of the results. For instance, although race/ethnicity was included as a covariate in these analyses given previous work reporting race/ethnic differences in pubertal development (Deardorff et al., 2021; Dorn & Beltz, 2023), race/ethnic differences were not interpreted given the limited number of racially/ethnically minoritized youth in the sample. No substantive theoretical perspectives on race/ethnic differences in adolescent cognitive and behavioral regulation exist that can be investigated in the current analyses (see Coll et al., 1996; Suzuki et al., 2021 for further elaboration). Future research would benefit from using larger, more representative samples and controlling for race/ ethnic differences as well as systemic inequalities (e.g., structural racism, discrimination, task development, and norms for different race/ethnic groups) that might affect these associations.
Second, the correlational nature of the study design prevented us from establishing causal links between the variables under study. Early life SEP might also capture a set of environmental risk factors related to both puberty and cognitive and behavioral regulation, which were not included in the present study. As such, an important future direction would be to examine the contextual mechanisms that might explain the observed associations between SES, puberty, and adolescent cognition. Third, maternal education and family income-to-needs were averaged, and a composite was created to assess SEP, but other SEP indicators such as family type, parent employment status, measures of economic insecurity, and census region could be used to provide a richer picture of contextual factors linked to cognitive and behavioral regulation. Fourth, because estimates of timing and tempo were derived separately by gender, they cannot be directly compared. This is common in puberty research but might limit interpretations. Further, there is limited evidence available for pubertal tempo and its measurement and additional pubertal indicators (e.g., status, synchrony) could also be considered. Finally, although the cognitive and behavioral regulation measures included in the present study are well-validated, they are not comprehensive. Future research would benefit from understanding how early life SEP and puberty are related to other cognitive and behavioral regulation outcomes not included in the present study, such as attention and cognitive flexibility.
CONCLUSIONS
In the current study, we examined whether pubertal timing and tempo moderated or mediated associations between early-life SEP and adolescent cognitive and behavioral regulation. We did so by interrogating different theories related to the role of pubertal timing/tempo. Overall, we found evidence that pubertal timing and tempo may operate as proximal effects that exacerbate existing vulnerabilities stemming from early life socioeconomic disadvantage. These findings support Dr. Schulenberg's theories on developmental turning points and disturbances, whereby proximal effects can enhance developmentally distal effects and prompt discontinuity in development (Schulenberg et al., 2018). However, the effects observed in the present study are disparate and may depend on gender or what aspect of SES and pubertal development are measured. A promising future direction of this research would be to extend these associations to adulthood, which is aligned with Dr. Schulenberg's perspective of “Taking the Long View” on adolescence, which connects childhood experiences, adolescent pathways, and long-term adulthood functioning (Schulenberg et al., 2018; Schulenberg & Maslowsky, 2015).