Reducing the Biological and Psychological Toxicity of Poverty-related Stress: Initial Efficacy of the BaSICS Intervention for Early Adolescents

Abstract

This proof-of-concept study tests the initial efficacy of the Building a Strong Identity and Coping Skills (BaSICS) intervention, a selective prevention of internalizing problems program for early adolescents exposed to high levels of poverty-related stress. Eighty-four early adolescents (M_age = 11.36 years) residing in very low-income neighborhoods were randomized to receive the 16-session intervention (n = 44) or to an assessment-only control condition (n = 40). BaSICS teaches coping skills, social identity development, and collective social action to empower youth with the ability to connect with members of their communities and cope with poverty-related stress in positive and collaborative ways. Pretest–posttest analyses showed that intervention adolescents acquired problem-solving and cognitive-restructuring skills and reduced their reliance on avoidant coping. In addition, HPA reactivity was significantly reduced in the intervention youth, but not controls. Finally, intervention youth’s internalizing and somatic symptoms as reported by both youth and their parents, showed significant reductions over time, whereas control youth had no such changes. Results provide strong support for this approach to strength-building and symptom reduction in a population of early adolescents exposed to poverty-related stress.

Introduction

Despite pervasive socioeconomic (SES) and race-based mental and physical health disparities (Wolfe, Evans, & Seeman, 2012), there is much resilience within the population of youth exposed to poverty and its associated noxious stressors (McBride Murry, Berkel, Gaylord-Harden, Copeland-Linder, & Nation, 2011). As emphasized by Coll et al. (1996), social stratification shapes children’s lived experiences in both positive and negative ways. As such, developmental research must take into account adaptive cultures and competencies—acknowledging and considering both promoting and inhibiting contexts. Cole (2009) further asserts that individual-level differences must be understood in light of broader systems-level factors, such as structural inequality, and should attend to sources of strength as well as strain. This proof-of-concept article describes and provides initial efficacy for a new intervention designed specifically to meet youth in poverty where they are skill- and development-wise, and to build promotive and protective factors to help them grow and flourish.

Adaptation to Poverty-related Stress

The Adaptation to Poverty-related Stress model (APRS; Wadsworth, Raviv, Santiago, & Etter, 2011) expanded strain models of economic hardship (Conger, Elder, Lorenz, Simons, & Whitbeck, 1994) to include the experiences and voices of children as well as parents and to focus on sources of protection in the face of risk. This developmentally and ecologically informed model proposes that poverty shapes children’s skills, abilities, and stress physiology to manage and adapt to stress in ways that serve them well in the short-term but can place them at long-term risk for mental and physical health problems (Santiago & Wadsworth, 2011). According to APRS, coping skills and resources can confer protection in the face of poverty-related stress (PRS) and mitigate its effects on physical and mental health problems. The APRS also asserts that there are no universally adaptive or maladaptive coping strategies, and in order to fully understand risk and resilience processes, a child’s adaptation must be viewed in light of their social context.

A growing body of empirical research suggests that PRS often interferes with the development of a wide and flexible coping repertoire that can be applied to a variety of stressors, and instead promotes coping specifically tailored to dangerous, unpredictable environments (Wadsworth, 2015). Whereas engagement coping, which includes primary (e.g., problem-solving, emotion regulation) and secondary control (e.g., acceptance, cognitive reappraisal) strategies, is typically associated with better psychological functioning across various ecological contexts (Wadsworth, Raviv, Compas, & Connor-Smith, 2005), disengagement coping (e.g., avoidance) is generally associated with poorer functioning in many, but not all contexts (Wadsworth & Compas, 2002). PRS appears to constrain the development of engagement coping strategies, leading to greater reliance on disengagement, which while well suited for coping with danger and violence, is inadequate and often inappropriate for situations not involving physical threat or imminent harm. Hypervigilant and avoidant coping, for example, are functional and serve a critical purpose in the context of toxic levels of stress, but tax a child’s mind and body, predicting the development of anxiety, depression, and health problems over time (Santiago & Wadsworth, 2011). It appears, therefore, that under conditions of chronic PRS, typically maladaptive coping is fostered at the expense of strategies that can be applied to a wide range of stressors. Interventions striving to meet such youth where they are cannot simply take a deficit approach, assuming that youth facing PRS need to learn the “right” coping skills and unlearn the “wrong” ones. Rather, programs should foster flexible response to a diverse array of stressors.

The conditions that interfere with the development of a flexible engagement-heavy coping repertoire also lead to dysregulation in the hypothalamic–pituitary–adrenal axis (HPA), which orchestrates physiologic responses during threatening encounters (Del Giudice, Ellis, & Shirtcliff, 2011). The autonomic nervous system and HPA work in tandem to regulate our biological functioning while at rest and during threatening encounters. When over-taxed, the HPA’s feedback mechanisms can become modified. This process initially leads to an exaggerated response or what has been termed hyperactivation (inability to downregulate a stress response) of the HPA and over time eventuates in a hypoactive response (inability to activate the stress response; Booij, Bouma, de Jonge, Ormel, & Oldehinkel, 2013). Both HPA patterns (hypo- and hyperactive) are associated with internalizing psychopathology. Furthermore, chronic poverty (Fernald & Gunnar, 2009) and social class and racial discrimination (Zeiders, Doane, & Roosa, 2012) are known to lead to these atypical HPA responses. Atypical responses, while maladaptive from a psychopathology perspective actually function to protect the developing child from immediate damage in the face of danger (physical or psychological; Ellis, Del Giudice, & Shirtcliff, 2013), are therefore adaptive in such contexts. For example, being hypervigilant to signs of impending conflict may allow a child to leave a situation (e.g., their home) in time to avoid a potentially dangerous episode of violence, but also causes wear and tear on psychological and physiologic systems.

Culturally Appropriate Coping Resources

The negative messaging children and adults receive about being members of marginalized groups (people of color, individuals in poverty) can erode internal resources such as self-esteem and pride in one’s family—resources sorely needed for coping with stressors stemming from poverty and discrimination, for example (Spencer, Fegley, Harpalani, & Seaton, 2004). The development of broad, highly diversified coping repertoires which allow poverty-affected and minority children to cope effectively with PRS likely requires a positive sense of self-grounded in an adaptive culture that promotes collective and individual agency. According to Coll et al. (1996), adaptive cultures arise out of the collective historical experiences of marginalized groups, which foster culturally defined coping.

The exploration of, identification with, and connection to one’s cultural heritage are key aspects of cultural identity development likely to promote resilience across social categories and cultural contexts salient to poor youth. Especially important are the stories, practices, ideals, and values that have guided past struggles and led to triumphs over adversity as they validate one’s experience of marginalization, celebrate resilience, and encourage agency. These serve as the ideal vehicle for promoting coping in context among youth who may differ in many ways, but share common experiences growing up in poor families and communities. The APRS model therefore leverages these collective strengths, proposing that cultural identity can serve as a protective asset for youth coping with PRS (Wadsworth, Ahlkvist, McDonald, & Tilghman-Osborne, 2018). For a strength-based (rather than deficit-focused) coping intervention for poor youth, it is vital to leverage context-specific protective factors to build upon their preexisting individual and collective assets and experiences (Zimmerman et al., 2013).

Transformative Intervention

Advocates of intersectionality (Rosenthal, 2016), social justice education (Hahn Tapper, 2013), and community psychology (Prilleltensky & Fox, 2007) remind us that intervention can serve the purposes of either “domestication or transformation,” meaning that it can either serve to keep marginalized people powerless and passive or it can be used for empowerment and action. Taking empowerment seriously requires that interventions sow the seeds of transformation—seeds that can grow to support social action and community change. As proposed by Shin et al. (2017) interventions qualify as “transformative” when they address the intersection of multiple social identities in the context of structural inequalities and call for social justice. Connecting with others and learning how to participate in collective social action are empowering ways of coping with systemic stressors that can enable individual agency and foster belongingness (Zimmerman et al., 2013). While youth living in poverty may not be able to change the economic circumstances of their family or avoid witnessing crime in their neighborhood, they can join with others to take action in prosocial ways to work toward improving their community (Burns, Merritt, Chyu, & Gil, 2019). Following Gutiérrez’ (1994) call to frame individual ways of coping with stress in the context of collective empowerment, we have sought to develop an intervention that fosters “collaborative coping” whereby learning fundamental coping skills such as problem-solving, emotion regulation, and acceptance is a means to contributing to collective action that changes the social context and is transformative for individuals.

According to the APRS and as noted in Wadsworth et al. (2018), an intervention designed to address the needs of marginalized adolescents such as those living in poverty would include the following: (a) Inputs: the curriculum would embrace a dialectic of acceptance and change—too often interventions invalidate the legitimacy of children’s anger and outrage over injustice. It would use principles of social justice education to infuse children’s knowledge and experiences into the fabric of the intervention during every session (Sylvestre, Notten, Kerman, Polillo, & Czechowki, 2018). Intervention leaders would be sufficiently clinically trained so as to be comfortable naming injustice, able to validate the hurt, anger, and aggressive impulses that stem from marginalization, and skillful in sowing the seeds of empowerment to help children develop socially acceptable ways to address stressors in their lives. (b) Outputs: the program would build both personal and collaborative coping skills, help children identify and connect with positive aspects of their cultural backgrounds, and give children direct experience with community action to promote belongingness, agency, hope, and a sense of purpose. (c) Outcomes: the intervention would attend to both psychological and biological mechanisms and outcomes. Emerging knowledge about the biological embedding of stress and of the biological substrates of coping suggest that potent adaptive coping skills should be evident at multiple levels of analysis. The ability to use relaxation techniques or cognitively reframe negative thoughts should therefore be accompanied by efficient downregulation of stress physiology and improvement on indices of anxiety and depression (Wadsworth et al., 2013).

Building a Strong Identity and Coping Skills (BaSICS) is the first intervention to explicitly apply the APRS and is designed to improve and expand preadolescents’ coping and self-regulation skills and to (re-) calibrate the physiologic stress response system (especially the HPA). We theorized that these individual-level goals would be best achieved through a social justice intervention in which coping and self- regulation skills are conceptualized as resources to facilitate positive identity development and proactive engagement with others to bring about social change (Quinn, Pacella, Dickson-Gomez, & Nydegger, 2017). In line with social justice education and empowerment approaches, content is derived from children’s own lived experiences and issues that matter to them, such as problems in the community. This approach helps raise awareness of social injustice and lays the groundwork for how they as “collaborative copers” can address important issues together (Roy, Raver, Masucci, & DeJoseph, 2019). The BaSICS curriculum provides a structure to help youth learn and apply coping skills and develop a socially embedded sense of self, culminating with a youth-designed community action project that builds community assets and/or reduces community stressors. Via the social action work, which is made possible by the previous two modules’ coping and identity skill building, youth voices are heard, acknowledged, valued, and used for action. They become engaged in agentic activities to solve real problems in their community (Burns et al., 2019). Agency, belonging, sense of purpose are hypothesized to empower youth in prosocial ways. In this way, BaSICS is designed to transform youth into agents of change and authors of their own stories rather than victims.

The Current Study

The current study describes the first efficacy trial of BaSICS. Intent-to-treat analysis with all 84 early adolescents randomized to participate in the 16 session BaSICS intervention (n = 44) or an assessment-only control group (n = 40) and analyses with “completers” (i.e., early adolescents who completed both pretest and posttest) were conducted. First, this study examines the acquisition of core coping and self-regulation skills. Second, pre–post changes in early adolescents’ cortisol reactivity and recovery responses to the Trier Social Stress Test (TSST) and on several indices of internalizing symptoms in intervention versus control participants were examined. We hypothesized that as follows: (a) early adolescents in the intervention would acquire and be able to demonstrate increased primary and secondary control coping skills (e.g., problem-solving, emotion recognition, distraction, and positive thinking) and show decreased reliance on disengagement coping (e.g., avoidance) relative to controls; (b) intervention youth’s cortisol patterns would change from pretest to posttest to resemble a more typical physiologic stress response, whereas control adolescents would not show changes over time; and (c) improvement in overall internalizing problems as well as specific indices of depression and somatic symptoms would occur for intervention youth, but not for controls.

Because cortisol levels fluctuate with age, pubertal development, and by gender, it is recommended that these variables be accounted for in analyses capturing variability in cortisol levels (Shirtcliff et al., 2012). In addition, internalizing symptoms and disorders are known to increase in prevalence across adolescence, to be higher in those with off-time pubertal maturation, and to see a shift toward greater prevalence for girls than boys by age 13 (Hankin; Parent, Forehand, Pomerantz, Peisch, & Seehuus, 2017). Older age, more advanced maturation, and female sex are also associated with higher levels of cortisol in adolescence (Natsuaki, et al., 2009; Shirtcliff et al., 2012). Given the importance of age, sex, and pubertal maturation to our outcomes, these variables were therefore included in all statistical models.

Method

Study Procedures

All study procedures were approved by the Pennsylvania State University’s institutional review board. All adult participants provided informed consent, and all youth assented to participate. Parent–youth dyads were recruited from two neighboring urban communities in central Pennsylvania that have very high levels of poverty (40–44% of children under age 18 live below the poverty line; U.S. Census Bureau, 2017), as well as crime, trauma, and violence exposure (violent crime rate in both communities is more than 250% U.S. average). Both communities have similar racial/ethnic diversity with the largest Census-identified groups being African American (43%/52%), White (45%/39%), and Hispanic/Latino (17%/20%).

Participants were recruited via flyers posted at youth-serving community agencies and neighborhood schools and in person using informational brochures and goody bags distributed at community and school events attended by recruiting staff who are local residents. Recruiters targeted an array of locations and events where parents of early adolescents were likely to be present, including back to school and parent–teacher events at neighborhood schools, church-sponsored events, community festivals and block parties, and recreational youth programs, and sent brochures home from school with youth. While time-consuming, these extensive efforts to canvas the two communities were deemed necessary to maximize the generalizability of the study’s findings. The recruiting protocol described the purpose of the study as learning how the BaSICS Program can build upon a child’s natural ability to handle stress and how these skills promote physical and mental health. Parents were informed that participation is voluntary and confidential and that information about them and their child would not be shared with the youth’s school. Recruiters were trained to explain the research procedure, monetary incentive, time commitment, and answer questions about the BaSICS intervention and participation in its evaluation. Interested families were contacted by program staff to screen children for eligibility.

Eleven and twelve year-olds without clinically elevated anxiety or depression, or evidence of an autism disorder or intellectual disability from families with an income ≤200% FPL, and one parent or guardian, who agreed to complete the intervention should they be randomized to the intervention condition, were eligible for the study. As shown in Fig. 1, out of the 183 youth assessed for eligibility, 24 did not meet the inclusion criteria and 68 of those eligible declined to participate. Overall, 37% of recruited youth were recruited from middle schools and 63% were recruited from community locations and events. Of the refusals, 12.5% (8) were recruited at schools and the remaining 87.5% (60) were recruited in the community.

Fig. 1.

Consort table.

Ninety-one eligible child–parent dyads were scheduled for a 3-hour pretest assessment appointment during which they were consented for the study and agreed to be randomized. At the in-person pretest visit, children were assessed at their school or the project office located in one of the communities between the hours of 4:00 PM and 7:00 PM. Participants completed interviews, questionnaires, and the modified Trier Social Stress Task (TSST-M; Yim, Quas, Cahill, & Hayakawa, 2010). Parent–child dyads received $40 upon completion of pretest. While children were completing measures and procedures, parents completed self-report and parent measures on tablets.

Following the pretest assessment, youth were randomly assigned to intervention or an assessment control condition using Research Randomizer (randomizer.org) to create 24 sets of numbers (0 or 1), which were printed out, placed in sealed envelopes, and handed to youth in the order generated by the program following completion of the pretest. Forty-four of the 51 intervention-assigned youth participated in the 16 session BaSICS intervention over 8 weeks while the 40 youth assigned to the control group did not (see Fig. 1). Five groups with a mean size of 8.4 youth and average attendance of 12.6 sessions were led by two lead facilitators and one assistant facilitator and met for 2 hours twice per week. Groups were conducted at a community health center and two middle schools all of which are located in neighborhoods with some of the highest levels of violent and drug-related crime, blight, and residential vacancies in the area, according to official police records. All participants completed a posttest two weeks after the date of the final intervention session. This ten-week delay between pre- and posttest has been shown to eliminate practice effects of the TSST (Petrowski, Wintermann, & Siepmann, 2012). Posttest assessment followed the same procedure as pretest, with parent–child dyads receiving $60 as payment.

Participants

Ninety-one parent–child dyads were recruited into the study and completed all pretest measures (62.6% of children and 91.6% of caregivers were female), but seven dropped out of the study prior to starting the intervention, leaving an analysis sample of N = 84. No differences between the dropped participants and those remaining in the study were evident on indices of child internalizing, parental depression, or perceived economic hardship. All of the 7 who dropped out after pre-assessment were single mothers with an average income of $8,300/year in comparison with $21,000/year for the remaining dyads. These parents all reported that their schedules no longer accommodated the intervention, despite agreeing to be available during the screening process. All data reported in tables and text were restricted to the analysis sample. Children were on average 11.32 years old (SD = 0.59) and were 45.1% Hispanic/Latino, 35.5% Non-Hispanic Black, 12.9% Non-Hispanic White, and 6.5% Other/Multiracial. Parent’s age ranged from 26 to 63 years of age (M = 39.05, SD = 8.20), and they were 37.9% Hispanic/Latino, 43.9% Non-Hispanic Black, 12.2% Non-Hispanic White, and 6.1% Other/Multiracial. The average annual household income was $19,742 (SD = 11,897), which is below the 2017 federal poverty threshold for a family of four set at $28,780 (U.S. Census, 2017). Most caregivers were employed (53.9% either full-time or part-time), 22.4% were unemployed, 14.5% were on disability and 11% were homemakers, full-time students, or retired. Thirty percent of parents did not complete high school, 30.3% had a high school diploma or GED, and 32.9% had schooling beyond high school. The majority of participating families received public assistance (82.9%) and reported food insecurity (57.1%). Lastly, 31% of parents were single, 18% were cohabitating, 24% were married, and 21% were widowed, separated or divorced.

Sixteen dyads did not return for their posttest assessment. Most attriters (75%) were from the control group. Hence, intervention retention was 91%, while control retention was 70%. Completers and attriters differed on CBCL, F(1, 62) = 3.93, p < .055 and YSR total problems, F(1, 62) = 4.0, p < .05, reflecting poorer pre-intervention functioning for completers—this was not different for intervention versus control participants, all ps > .10.

Measures

Demographics

Parents completed questionnaires on a tablet using Qualtrics software. Parents reported basic demographic information about themselves (sex, race/ethnicity, immigrant status, marital status, annual family income, education level, and receipt of public assistance) and about their child (age, sex, and race/ethnicity).

Salivary Cortisol

Six saliva samples were taken via passive drool (Davis, Bruce, & Gunnar, 2002); immediately preceding TSST-M (T1; 40 minutes post-arrival at laboratory), immediately following TSST (T2; 15 minutes after T1), and at four 10-minute intervals thereafter (T3–T6). Given the 10- to 30-minute delay in cortisol’s appearance in saliva, T3 represented the average peak cortisol level in the sample (Kirschbaum, Pirke, & Hellehammer, 1993). However, as recommended by Miller et al. (2018), to account for heterogeneity in individual peaks, we calculated reactivity and recovery metrics using each child’s own peak (T3, T4, or T5). To control for external factors that influence cortisol, participants were directed not to eat a large meal, eat dairy products, brush their teeth, or drink sugary beverages an hour before their appointment. Saliva samples were stored in a biomedical grade freezer prior to the analysis at the CORE Biomarker Lab at Penn State University. Cortisol levels were determined using duplicates, with the detection levels in the range of 0.003–3.0 kg/dL (range, 0.08–82.77 nmol/L) using a commercial expanded-range high-sensitivity enzyme immunosorbent assay kit (Lot #’s 1410510 and 1502503; Salimetrics, LLC, State College, PA, USA).

Data Reduction and Preprocessing

Outliers more than 3 SD from the mean were winsorized to 3 SD (Smyth, Hucklebridge, Thorn, Evans, & Clow, 2013). In 6 saliva samples, insufficient saliva was collected for analysis (occurring randomly across participants). These 6 values were replaced with the mean cortisol value for the corresponding time point. Natural log (ln) transformation was used to normalize all salivary cortisol data (Smyth et al., 2013).

Cortisol Reactivity and Regulation

Area Under the Curve with respect to the increase (AUC_i) is the most common metric for capturing the amount of cortisol released between baseline and peak, and then from peak back to baseline (Pruessner, Kirschbaum, Meinlschmid, & Hellhammer, 2003). As such this method was used to compute both reactivity and recovery scores. Area Under the Curve with respect to the ground (AUC_g) refers to the total amount of cortisol secreted during a proscribed time period (i.e., starting at an absolute zero). AUC_i captures only changes in cortisol that occur starting at the baseline level (not at zero) and is therefore calculated by subtracting the baseline cortisol level from AUC_g. This is accomplished using the formula below. AUC_ireact was calculated by subtracting T1 baseline cortisol from the AUC_greact (e.g., T1 to peak) trapezoid area, and AUC_irecov was calculated by subtracting T6 from AUC_grecov (e.g., peak to T6).

$$\text{AU}{\text{C}}_i=\left(\sum _{i=1}^{n-1}\frac{(m_{(i+1)}+m_i)*t_i}{2}\right)-\left(m_i*\sum _{i=1}^{n-1}{t}_i\right).$$

Coping Skills Acquisition

The Coping Skills Scale (CSS) was used to assess knowledge of primary and secondary control coping (Raviv & Wadsworth, 2010). The first section assesses acquisition of knowledge about primary control coping (problem-solving). Participants’ ability to describe the problem-solving steps taught during the intervention was assessed. The second section assesses secondary control coping (distraction activities and positive thinking). Participants’ ability to describe the steps to engage in positive thinking and distraction were assessed. Coping skill acquisition coding was accomplished by trained coders who were blind to experimental condition. Thirty percent of protocols were coded by two coders to test inter-rater reliability. Sum scores were calculated and inter-rater reliabilities for primary control (pre-κ = .70; post- κ = .90) and secondary control coping (pre-κ = .86; post-κ = .94) were adequate.

Responses to Stress Questionnaire

The 57-item Responses to Stress Questionnaire (RSQ; Connor-Smith, Compas, Wadsworth, Thomsen, & Saltzman, 2000) assesses primary control (e.g., problem-solving, emotion regulation) and secondary control (e.g., acceptance, distraction) coping in addition to disengagement coping (e.g., avoidance, wishful thinking) as well as involuntary responses to stress. Only primary control, secondary control, and disengagement factors were used in the current study. All items were rated on a Likert scale from 1 (not at all) to 4 (a lot). Ratio scores, which capture the amount of each factor taking into account the full stress response, were calculated for each factor by dividing the sum of each factor by the total score. Adequate internal consistency was found at pre and post, respectively, for primary control (α = .74; α = .84), secondary control (α = .76; α = .83), and disengagement coping (α = .76; α = .83).

Child Behavior Checklist and Youth Self-Report

Parents completed the Child Behavior Checklist (CBCL; Achenbach & Rescorla, 2001), and youth completed the Youth Self-Report (YSR; Achenbach & Rescorla, 2001). The CBCL contains 113 items and the YSR has 112 items that describe behavioral and emotional problems, which are rated on a three-point Likert scale, with 0 (never true), 1 (sometimes true), and 2 (very often true). Both measures include 8 narrow-band syndromes and 3 broadband factors. This study focused on internalizing, and therefore, the following scales were used: Somatic Complaints, Affective Problems, Internalizing Problems, and Total Problems. Cronbach alphas on the CBCL range from .78 (Somatic Complaints) to .97 (Total Problems) and on the YSR from .80 (Somatic Complaints) to .95 (Total Problems).

Covariates

Covariates include sex (0 = male, 1 = female), puberty, and financial strain.

Puberty

Pubertal status was assessed using the parent-report Pubertal Development Scale (PDS; Petersen, Crockett, Richards, & Boxer, 1988), indexed by menarche, height, skin changes, body hair growth, and breast growth for girls, and by voice changes, height, skin changes, body hair growth, and facial hair growth for boys. Girls’ average PDS score was 2.42 (SD = 0.57; range 1.0–3.6), and boys’ was 1.76 (SD = 0.44; range 1.0–2.6).

Financial Strain

The 7-item Financial Strain Index from the parent-report economic strain RSQ (Connor-Smith et al., 2000) was used to index the degree of financial strain in the home in the previous 6 months. Parents indicated which of seven major financial problems such as losing a job occurred (0 = no, 1 = yes). Sample average was 1.65 (range 0–6), indicating that the average family in the study had experienced one or two of these life events in the past 6 months.

Intervention: Building a Strong Identity and Coping Skills (BaSICS) Program

The three modules addressing fundamental coping skills, cultural identity development, and community action, which comprise the BaSICS intervention, are described briefly below. The program is designed to be delivered either weekly or biweekly in sixteen total 2-hour sessions and is appropriate for either in-school or after-school delivery (Herrenkohl, Hong, & Verbrugge, 2019). The full BaSICS curriculum is described in more detail in Wadsworth et al. (2018).

Module 1: Fundamental Coping Skills

During the first module of the intervention, participants learn and practice foundational coping skills individually and collaboratively. Raviv and Wadsworth’s (2010) Families Coping with Economic Strain program (FaCES), which teaches primary control coping (e.g., problem-solving), secondary control coping (e.g., cognitive restructuring), and relaxation techniques (e.g., guided imagery, diaphragmatic breathing), was adapted for use in BaSICS. In the current investigation, stressors related to social inequality (e.g., discrimination, community violence) were addressed in greater detail.

Module 2: Community and Cultural Identity

The second module of the intervention uses participants’ growing coping awareness and competence to identify community stressors and assets and to explore their personal and cultural identities. Participants discuss possible changes in their community that would help reduce social stressors and increase social assets. Multiple sessions are devoted to exploration of participants’ family and cultural root systems that nourish and support them and give them their unique strengths and talents.

Module 3: Collaborative Coping

In the final BaSICS module, participants apply their coping skills and draw on what they have learned about their community and themselves to design and carry out a community action project of their own choosing. Participants earn the funding for their project with homework completion, on-task behavior, and positive contributions during group meetings. The projects are developed and implemented by the early adolescents in partnership with a community organization such as their school, with scaf-folding and support from facilitators, teachers, school administrators, parents, and other community members. Past community projects have included creating an ongoing dialogue between the local police department and middle school students via suggestion boxes and community forums, and planting a community garden at the school and linking it with the school science club to facilitate sustainability. They learn the purpose and principles of social action as a constructive, group-based response to community stressors, which we termed “societal stress.”

Fidelity

Several strategies were used to maintain provider skills and ensure fidelity to the intervention material. First, the program has a detailed, step-by-step, session-by-session intervention manual as well as a behavior management manual. Second, a licensed clinical psychologist versed in the intervention conducted weekly supervision with facilitators to review past week’s content and preview upcoming content. Adherence to protocol was monitored via coding of videotapes of sessions and in-session raters. As rated by at least two raters per session, 89.9% of required elements were covered. Inter-rater agreement was κ = 95%.

Analysis Plan

To examine whether there were differences in coping skills acquisition, cortisol response to the TSST, and symptoms of psychopathology between intervention and control participants, GLM repeated measures analyses of variance (ANOVAs) were conducted with pre–post scores. Time was a within-subjects factor and group was between subjects. Analyses were conducted twice in order to ensure robust effects. First, efficacy analyses were conducted to determine if the intervention was effective with study completers—those youth who completed both pre- and post-intervention assessments and attended at least one intervention session if assigned. Second, we conducted intent-to-treat (ITT) analyses to ensure that our treatment effects were not overestimated due to differential dropout. Missing values were replaced with pretest values. Third, we computed reliable change improvement (RCI) and deterioration (RCD) scores to index the proportion of early adolescents in each group who demonstrated clinically reliable change on psychopathology and coping behaviors (Jacobson & Truax, 1991).

Results

Preliminary Analysis

Baseline Descriptive Statistics

There were a few differences between the control and intervention groups on demographic variables at pretest. Parents in the control group were more likely to be employed [t(49) = 1.78, p < .08], and youth in the intervention group were more likely to be girls [t(66) = −2.11, p < .05] with greater pubertal development [t(64) = −2.08, p < .05]. All analyses therefore control for financial strain, youth pubertal status, and youth biological sex. ITT analyses yielded similar results unless specified otherwise. Differences in power resulted in slight variations in F and p values as noted below. Means, standard deviations, and correlations among study variables at pretest are reported in Table 1.

Table 1.

Pretest means, standard deviations, and correlations among study variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
1. Sex (1 = female)	‐
2. Age	.12	‐
3. Puberty	49**	.35**	‐
4. CSS Primary	−.05	.02	−.07	‐
5. CSS Secondary	.09	−.12	−.07	.20	‐
6. RSQ Primary	.30**	.02	.01	.07	.03	‐
7. RSQ Secondary	.05	.14	.09	−.04	.08	.17	‐
8. RSQ Disengage	−.25*	.05	−.04	.04	−.14	−.43**	−.09**	‐
9. sC Reactivity	.15	.01	.03	.18	−.02	.08	−.03	−.06	‐
10. sC Recovery	.24*	.08	.13	.02	−.13	.08	.12	.01	.19	‐
11. CBC Affect	−.02	−.13	.01	.05	.19	.01	.24*	−.02	.02	−.11	‐
12. CBC Somatic	.01	−.18	.05	−.07	.03	.11	.13	−.10	.05	.02	.74**	‐
13. CBC Int	−.03	−.14	.02	−.04	.11	.03	.23*	−.10	.03	−.01	.85**	84**	‐
14. CBC Tot	−.17	−.23*	−.09	−.02	.09	−.03	.13	−.04	−.06	−.12	.76**	70**	87**	‐
15. YSR Affect	−.15	−.21	−.20	−.11	.10	−.35**	−.36**	.10	−.07	−.17	−.05	.03	−.02	.01	‐
16. YSR Somatic	−.12	−.27*	−.13	−.01	.20	−.22*	−.44**	.11	.01	−.24*	−.02	−.01	−.04	−.01	.76**	‐
17. YSR Int	−.09	−.19	−.15	−.08	.21	−.26*	−.51**	.13	−.07	−.28*	−.10	−.07	−.04	−.01	78**	.83**	‐
18. YSR Tot	−.09	−.16	−.16	−.08	.09	−.34**	−.44**	.15	−.06	−.27*	−.05	−.01	−.01	.03	79**	.69**	.88**	‐
Mean		11.82	2.17	0.24	0.11	0.183	0.245	0.154	−0.03	0.39	57.27	58.61	55.03	55.75	56.04	59.84	54.51	53.84
SD		0.59	0.59	0.077	0.38	.039	0.051	0.030	0.55	0.67	9.09	9.78	12.20	12.16	6.71	9.47	10.81	10.46

Affect, affective problems; CBC, Child Behavior Checklist; CSS, Coping Skills Scale; Disengage, Disengagement Coping; Int, internalizing problems; Primary, Primary Control Coping; RSQ, Responses to Stress Questionnaire; Secondary, Secondary Control Coping; Somatic, somatic complaints; Tot, total problems; YSR, Youth Self-Report.

^*p < .05.

^**p < .01.

Changes in Behavioral and Physiologic Responses to Stress

Coping Skill Acquisition and Use-Hypothesis 1

Results support our first hypothesis that early adolescents who participated in the intervention would acquire and be able to demonstrate increased primary and secondary control coping skills and decreased reliance on disengagement coping relative to youth in the control condition. To assess differences in early adolescents’ knowledge of the targeted engagement coping skills, pretest scores on the CSS were compared to posttest scores (see Table 2).

Table 2.

Pretest and posttest mean scores by group and BaSICS intervention effects analyzed with 2 (group) × 2 (time), GLM repeated measures ANOVAs (controlling for sex, puberty, and financial strain)

	BaSICS (n = 40)				Control (n = 28)
	Pretest		Posttest		Pretest		Posttest		Time × group
Variables	M	SD	M	SD	M	SD	M	SD	F(1, 63)	${\eta }_p^2$
CSS Primary	0.25	0.83	1.90	1.92	0.14	0.52	0.11	0.42	11.49**	.18
CSS Secondary	0.07	0.26	1.22	1.31	0.14	0.47	0.11	0.31	13.58**	.21
RSQ Primary	0.18	0.04	0.19	0.04	0.19	0.04	0.19	0.03	0.03	.001
RSQ Secondary	0.25	0.05	0.25	0.04	0.24	0.05	0.25	0.05	0.13	.003
Disengage. Coping	0.15	0.03	0.15	0.02	0.15	0.03	0.16	0.02	3.64+	.07
Cortisol React (Ln)	0.09	0.32	−0.28	0.64	−0.05	0.55	0.09	0.47	5.74*	.10
Cortisol Recov (Ln)	0.47	0.73	0.27	0.58	0.30	0.55	0.29	0.63	0.16	.003
CBC AFFECT	58.33	10.48	55.54	9.08	56.66	8.01	56.32	6.09	4.10*	.08
CBC SOMATIC	59.40	10.96	58.07	10.26	57.66	8.44	55.96	6.93	0.91	.02
CBC INT	56.26	13.25	51.63	14.07	54.00	11.38	53.48	10.66	4.00*	.08
CBC TOT	56.52	12.81	51.54	14.89	55.16	11.93	56.48	8.29	5.72**	.10
YSR AFFECT	56.26	7.27	55.51	6.32	55.84	6.09	56.11	7.19	1.00	.02
YSR SOMATIC	59.53	9.54	56.68	7.46	60.03	9.40	59.78	7.74	5.84*	.11
YSR INT	53.26	11.10	51.27	11.30	55.55	10.72	55.00	13.21	0.51	.01
YSR TOT	53.16	10.81	50.85	11.35	54.50	10.00	54.04	12.78	0.95	.02

AFFECT, affective problems; CBC, Child Behavior Checklist; CSS, Coping Skills Scale; Disengage, disengagement coping; INT, internalizing problems; Primary, primary control coping; RSQ, Responses to Stress Questionnaire; Secondary, secondary control coping; SOMATIC, somatic complaints; TOT, total problems; YSR, Youth Self-Report.

⁺p < .10;

^*p < .05;

^**p < .01;

^***p < .001.

Significant time × condition interactions were found for acquisition of primary control coping knowledge [F(1,51) = 11.49, p = .001, ${\eta }_p^2=.18$] and secondary control coping knowledge [F(1, 51) = 13.58, p = .001, ${\eta }_p^2=.21$] as shown in Table 2. At pretest, the majority of youth in the intervention (86%) and control (90%) groups could not list any of the problem-solving steps taught in the intervention. However, at posttest youth in the intervention on average listed two of the five problem-solving steps (41.5% were able to list three or more of the problem-solving steps) whereas those in the control group were still unable to list any problem-solving steps (93%). Similarly, in terms of secondary control coping, at pretest the majority of youth in the intervention (92%) and control (90.0%) groups were unable to list any specific ways to engage in positive thinking and distraction. However, at posttest youth in intervention on average listed more than one of the targeted ways to think positively (34% were able to list two or more), while the majority of those in the control group were still unable to list anyways to engage in positive thinking (89%).

Pretest scores on the RSQ were compared to posttest scores to examine whether there were differences in the proportional use of primary control, secondary control, and disengagement coping skills between the intervention and control groups. As shown in Table 2, there were no significant differences in RSQ reports of primary and secondary control coping between the groups. The time × condition effect on RSQ disengagement coping approached statistical significance [F(1, 55) = 3.64, p = .06, ${\eta }_p^2=.07$], and in ITT analysis this effect reached significance [F(1, 72) = 5.23, p = .03, ${\eta }_p^2=.07$]. ITT results revealed that intervention youth had substantial reductions in the proportion of their stress responses that comprised disengagement coping while control youth showed increases.

Cortisol Reactivity and Regulation-Hypothesis 2

To test the hypothesis that intervention youth’s cortisol patterns would change from pretest to posttest to resemble a more typical physiologic stress response, whereas control youth would not, we examined changes in early adolescents’ physiologic stress response by comparing pretest and posttest cortisol reactivity and recovery AUCi scores. As indicated in Table 2, there were no significant differences in cortisol recovery between the groups, but the time × condition effect on cortisol reactivity was significant [F(1, 54) = 5.77, p = .02, ${\eta }_p^2=.10$]. Youth in the intervention group had reductions in their reactivity from pretest to posttest while youth in the control group did not. Overall, there were less extreme activation patterns for the intervention group in comparison to the control group at posttest. This finding suggests that youth who participated in the intervention had a different physiologic response pattern at posttest relative to pretest, whereas control youth did not, supporting the hypothesis.

Mental Health Symptoms-Hypothesis 3

Parents and youth reported on early adolescents’ depressive problems, somatic problems, internalizing symptoms, and total problems. Changes in mental health symptoms were assessed by comparing pretest and posttest T-scores on the CBCL and YSR. The final hypothesis that improvement in symptoms of psychopathology would occur for intervention youth, but not for controls was also supported. As shown in Table 2, there were significant time × condition effects on parent-report of depressive problems, internalizing symptoms, and total problems. ITT analyses showed similar patterns for depressive problems and total problems, though the ITT time × condition interaction for internalizing symptoms was only marginally significant. Parents of youth in the intervention group reported decreases in depressive symptoms, internalizing symptoms, and total problems over time, while parents of youth in the control group reported no change in mental health problems from pretest to posttest. In terms of youth’s report of mental health symptoms, there was a significant time × condition effect on somatic complaints, with youth in the intervention reporting reductions in somatic problems as compared the control group.

Reliable Change Indices-Hypothesis 3 (Table 3)

Table 3.

Reliable change index improvement and deterioration effects

	Improvement		Deterioration
Measure	Intervention	Control	Intervention	Control
CSS primary control coping	46	4	2	4
CSS secondary control coping	59	11	2	14
RSQ primary control coping	12	11	7	7
RSQ secondary control coping	10	4	10	4
RSQ disengagement coping	17	7	5	11
CBCL affective problems	14	4	0	13
CBCL somatic complaints	5	0	3	0
CBCL internalizing	37	8	3	13
CBCL total problems	38	22	11	17
YSR affective problems	5	8	0	6
YSR somatic complaints	10	8	3	20
YSR internalizing	15	16	7	16
YSR total problems	36	20	18	16
Average	23.4	9.5	5.5	11.6

CSS, Coping Skills Scale; RSQ, Responses to Stress Questionnaire; CBCL, Child Behavior Checklist; YSR, Youth Self-Report. Numbers in the table reflect percentage of youth within the condition.

Reliable change improvement (RCI) and deterioration (RCD) scores were computed (Jacobson & Truax, 1991) for mental health and coping, and the percentage of youth surpassing a t-ratio of 1.96 on each was calculated. These scores support the main findings from the GLMs and suggest additional reliable positive outcomes not yet achieving statistical significance (YSR Total Problems and Internalizing). In most cases, the proportion of RCI was 2–5 times higher in intervention than in control. Further, the proportion of RCD was generally lower in intervention than control. In cases where the intervention RCD was not lower than the control RCD, (e.g., CBCL internalizing and Total problems, YSR Total problems), the accompanying RCIs were substantially higher in intervention than in control. These findings provide further support for Hypothesis 3.

Discussion

This study evaluated the preliminary efficacy of the BaSICS intervention, which was designed specifically to meet the needs of early adolescents living in poverty. Reflecting national trends (Bureau of Labor Statistics, 2018), more than half of the youth reside in families with employed parents whose wages are insufficient to raise their income above the poverty line. Results confirm the proof of concept that intervening with early adolescents using this developmentally informed, social justice-oriented psychosocial intervention can lead to improvements in early adolescents’ biological and psychological functioning. Specifically, compared to a control group, early adolescents in the intervention exhibited (a) significant improvements in their ability to demonstrate knowledge and application of the core coping skills taught in the intervention, (b) concomitant reduction in HPA axis reactivity, and (c) reductions in internalizing symptoms as reported by both youth and their parent. The pattern of associations provides support for the underlying logic model of change, namely that youth in the intervention condition will acquire coping skills that will be accompanied by changes in HPA axis function and psychopathology symptom improvement. These findings support recent calls to view the plasticity of the early adolescent brain as not only a source of risk, but also as a great opportunity to invoke lasting changes to youth’s self-regulation capacities (Steinberg, 2014).

Learning and Using Engagement (Control) Coping

Given that PRS exposure leads to the development of a restricted coping repertoire (Wadsworth, 2015), a central aim of the intervention is to improve youth’s engagement coping skills. The results show that it is possible to teach early adolescents the skills for coping with everyday stressors and PRS as well as what we call “societal stress,” or those situations and occurrences that occur in the greater community and over which an individual has little (or no) control. In this study, pre–post-intervention increases in primary and secondary control coping knowledge were consistent with past coping-based interventions with economically disadvantaged youth (Raviv & Wadsworth, 2010). Importantly, youth in the intervention condition also showed reductions in the relative use of disengagement coping. After the intervention, participants were better able to list the multiple problem-solving steps taught in BaSICS, generate more positive coping thoughts in response to a hypothetical stressor, and reduce their habitual reliance on disengagement coping. These changes were accompanied by positive changes in both mental health problems and cortisol reactivity. Acquisition of primary and secondary control coping seems to have been better captured by early adolescents’ ability to recall and apply skills to specific hypothetical scenarios than by the RSQ, which assesses youth’s general use of primary and secondary control coping skills. This could reflect that youth’s coping knowledge has not fully translated into widespread use of these skills in everyday life and are therefore not captured by the RSQ’s assessment of coping with the general domain of PRS.

Regulating Stress (HPA) Reactivity

In addition to promoting the acquisition of engagement coping skills including problem-solving and positive thinking, the intervention leads to less extreme salivary cortisol responses to the TSST (Kudielka, Hellhammer, & Kirschbaum, 2007). This study is one of the first to suggest that improving early adolescents’ coping skills has effects at the physiologic level and may present a route to repairing dysregulated HPA responses (Davidson & McEwen, 2012). While the intervention appears to have had the intended effect on cortisol reactivity, the predicted effects on recovery cortisol did not emerge. Whereas immediate physiologic benefits during reactivity can be realized from easy-to-implement strategies such as deep breathing, more difficult-toimplement strategies such as cognitive restructuring may take more practice before physiologic recovery benefits can be observed. Thus, recovery phase effects may emerge at subsequent assessments.

Improving Mental Health

Despite screening out adolescents with clinical levels of anxiety and depression, the youth in the study came in with elevated T-scores on most subscales of the CBCL and YSR, which is not surprising given their high levels of environmental risk exposures. Notably, improvements in psychological symptoms were evident for intervention youth in comparison to control youth. Parent-reported depressive symptoms, internalizing problems, and total problems, and youth-reported somatic symptoms all showed differential change from pretest to posttest, reflecting decreases for intervention, but not control youth. This may reflect the relevance of engagement coping in particular to internalizing problems (Hankin, Mermelstein, & Roesch, 2007).

It is noteworthy that changes in internalizing symptoms were evident across reporter, but the specific nature of the symptom changes varied by reporter consistent with context-driven rater effects (De Los Reyes, 2011). Parent-reported effects were specific to affective and behavior problems, whereas youth-reported effects were primarily isolated to somatic complaints. These reductions in somatic symptoms in particular, likely reflect the sustained attention given to interoceptive awareness and relaxation exercise in the intervention. These reductions are potentially quite important as we learn more every day about the psychobiological pathways from chronic stress in early adolescence to physical disease in adulthood (and earlier). The present findings suggest that some of the consequences of PRS may be reversible (Lebel & Beaulieu, 2011). Early adolescents graduate from BaSICS with new-found competence to combat stressors that may have previously been unmanageable without the skills acquired in the intervention.

Using Cohen’s (1988) rough benchmarks to describe their relative size [small (${\eta }_p^2=.01$), medium (${\eta }_p^2=.06$), and large (${\eta }_p^2=.14$)], the effects found in this study ranged from moderate to very large. Further, the effects were evident across reporters, measures, and measurement types (open-ended interviews, questionnaires) and were found in both biological and psychological systems. The effects were also reliable, robust to conservative intent-to-treat analysis, and consistent with our thoroughly researched logic model (Wadsworth et al., 2018). These all support the robust nature of the findings reported herein. The Reliable Change indices shed additional light on the importance of the outcomes of this evaluation, revealing a combination of prevention and treatment effects. The findings show that, without intervention, 11% of (control) early adolescents would show deterioration in functioning over the course of three months and 9.5% would show spontaneous improvements. In contrast, that deterioration rate is cut in half for youth receiving the intervention, and the improvement rate more than doubles for intervention youth. Given the significant risks faced by these youth every day, the findings are very compelling and suggest BaSICS has promise as a clinical intervention to prevent internalizing problems in youth facing the highest levels of environmental risk.

A note about parents. Why were not parents involved in this BaSICS intervention trial? First of all, parents were involved in the development of the program—in several focus groups and in consultation with our Community Advisory Board, parents and community members expressed an urgent need for an intervention like this, agreed on the principles taught in the program, and approved of the skills that their children needed help with. Many parents indicated that, while they would be interested in learning this type of material alongside their children, their work schedule and other responsibilities would likely preclude it. Hence, rather than develop an intervention only for children whose parents are able to attend too, we developed an intervention applicable for all youth affected by PRS. Second, our goal is that this program can be broadly implemented in schools as part of the curriculum, or as an after-school program, and can also serve as an adjunct to existing parenting interventions (Herrenkohl et al., 2019). Third, helping youth extend their support networks beyond the family is developmentally appropriate. The intervention’s identity exploration activities help each child identify the supportive people, practices, and institutions in their lives to help them connect with supportive adults, whether they are their parents or not. Schools, peers, and community agencies are all natural supports that are leveraged in the intervention, giving these youth a wider net of support with which to cope with PRS.

Study Limitations & Future Directions

As with any study, however, there are several limitations to the current study that should be addressed in future research. Changes in knowledge of problem-solving steps and ways to think more positively about stressors were evident when assessed with structured interviews requiring recall of learned content and application to hypothetical scenarios (CSS), but were not strongly evident in the RSQ self-report (though reductions in proportional use of disengagement coping were found). As noted above, this could reflect that regular use of these skills has not yet occurred and may be revealed over time. It could also reflect the inadequacy of a broad self-report measure to pick up on short-term changes resulting from an intervention. In addition, the findings from this study represent relatively short-term changes from pre- to post-intervention. Replication and long-term follow-up are planned. Coping skills are viewed as the critical foundation of this multi-component intervention, and this evaluation therefore focuses on coping and self-regulatory skills needed to engage productively with the increasingly complex social identity and social justice content that follows. Future studies will assess additional BaSICS skills and principles. Finally, multiple statistical comparisons were made to capture both parent and adolescent perspectives since it is critical to get youth reports in addition to parent reports on internalizing problems. However, with a relatively small sample size, this could increase Type I or II error. The reliable change analyses support that the effects found herein represent clinically meaningful change and that many more individuals showed positive changes in the intervention group and negative changes in the control group. In addition, as can be seen in Table 2, most effect sizes were moderate to large. Both of these facts increase confidence in the findings.

Cortisol levels are affected by many factors, including sex, age, puberty, time of day, steroid-based medications, and contamination from food or blood, for example. Steps were taken to minimize the effects of each of these. Time of day was carefully selected and strictly controlled—all assessments were conducted between 3:30 and 6:30 pm, the time of day with the flattest diurnal slope. Visits were timed to reduce saliva contaminants that would result from consumption of large meals, sugary treats, or tooth brushing. Sex and pubertal status were included as covariates in analyses to account for variability due to those factors. Pubertal status was measured via parental report on the PDS, which provides a non-invasive, but somewhat less precise assessment of pubertal development than physician examination (Dorn, Dahl, Woodward, & Biro, 2006). Finally, the sample size was too small to examine the effects of the intervention on these outcomes separately for boys and girls. This will be an important direction for future evaluations of this program, as sex differences have been found in rates of stress, psychopathology, and cortisol levels (Hankin et al., 2007; Kudielka & Kirschbaum, 2005).

Completers and attriters differed on total behavior problems, as early adolescents who participated in posttest assessments had higher behavioral problems at pretest than those who did not attend posttest. Hence, it seems that relatively healthy early adolescents were more likely to drop out of the study, perhaps reflecting a lower perceived level of need. Importantly, these patterns did not differ by experimental group. It is possible that youth with behavioral challenges or their parents saw potential benefits in participating, even if they were not receiving the intervention. Other interpretations are of course possible, including the possibility that youth with fewer emotional and behavioral problems are more likely to be involved in after-school activities such as sports, whose schedules interfered with participation in the after-school intervention and assessments. Either way, it is encouraging that the youth who appear to need the intervention most are more likely to remain in the study than those at lower risk.

Addressing the needs of highly vulnerable youth who face poverty-related toxic stress requires fresh ideas and perspectives—borrowing ideas and strategies from social justice education, as well as community, social, and clinical psychology made it possible to bring our best developmental science to a community in great need. We crossed disciplinary and sub-disciplinary boundaries in order to incorporate the physiologic stress response, attend to social identity, and address racism, discrimination, and prejudice (Sylvestre et al., 2018). Similarly, we worked side-by-side with youth and the adults in their lives and communities to develop collective approaches for addressing societal-level stress and promote the fundamental human needs of belongingness, agency, and hope (Quinn et al., 2017). The results of this research underscore the value and importance of bringing rigorous research methods to community work (Zimmerman et al., 2013). Finally, given the link between physiologic stress patterns and long-term mental and physical health, it is especially important to expand intervention evaluation methods to more fully capture changes in critical biological systems in addition to behavioral, cognitive, and emotional functioning. Decades of improved health and well-being could be the result (Steinberg, 2014).

Highlights.

• We examined the efficacy of the BaSICS preventive intervention for young adolescents in poverty.

• Adolescents randomized to receive BaSICS acquired new skills for coping with poverty-related stress.

• BaSICS participants showed significant reductions in internalizing problems over time.