Positive affect, a measure of psychological wellbeing (Boehm & Kubzanksy, 2012; Kashdan, Biswas-Diener, & King, 2008), is characterized by happiness, feeling satisfied and free from tension, and a hopeful outlook on life (Waterman, 2008; Watson, Clark, & Stasik, 2011). There is limited research on the distribution of wellbeing in the adolescent population, but studies do suggest that positive emotions are associated with numerous social-emotional outcomes. Positive wellbeing is associated with benefits to both short and long term physical health (Howell, Kern, & Lyubomirsky, 2007). Positive emotions are protective against symptoms of depression (Seligman, 2006). Beyond health outcomes, positive emotions have been found to have a broadening effect on cognitive processes (Fredrickson & Branigan, 2005) and are associated with higher reports of quality social interactions (Harker & Keltner, 2001).

Understanding positive emotions in adolescents is particularly important due to its influence on health across the life course (Masten & Tellegen, 2012). Yet much of mental health research focuses on mental illness (Brodhagen & Wise, 2008; Nolen-Hoeksema, Girgus, & Seligman, 1992) creating a tacit assumption that the absence of mental illness represents the presence of mental wellbeing. This is an oversimplification of psychological wellbeing and undermines the importance of understanding psychological wellbeing as a distinct form of mental health. Particularly for those in social work, understanding both ends of the psychological spectrum is crucial to understanding mental health and illness.

Studying the correlates of wellbeing can be used to promote mental health and create interventions that are not just geared towards decreasing symptoms of disorder (Diehl, Hay, & Berg, 2011; Watson et al. 2011) but are instead focused on increasing wellbeing. Lerner’s relational developmental model (Lerner, 2006) suggests a child’s inner life is connected to their environment; thus, examining contextual factors are essential to research on adolescent mental health. Two factors widely accepted to affect the social emotional health of adolescents is school and family support (Grusec, 2011; Maimon & Kuhl, 2008), and they are used here to explore their connection to positive affect.

The importance of school for adolescents cannot be overstated and, with family being first, is the second strongest socializing agent in their lives (Simons-Morton, Crumo, Haynie, & Saylor, 1999). In the school setting, adolescents are exposed to various life skills such as organization, teamwork, and critical thinking. In addition, stress buffering resources such as mastery, mattering, and coping mechanisms are often learned, practiced, and perfected within the school setting. Much of the research on school support examines the association to adolescent risk behaviors (McNeely & Falci, 2004) and academic outcomes (Klem & Connell, 2004). Evidence has also demonstrated that low perceptions of school support are associated with poor mental health. Newman, Newman, Griffen, O’Connor, and Spas (2007) found that a decline in a sense of school belonging was associated with an increase in depressive symptoms, and that being bullied increased distress in adolescents. Yet, school support may also be associated with psychological wellbeing. Studies have found that integration into school is connected to better mental health and lower disorder (Byck, Bolland, Dick, Ashbeck, & Mustanski, 2013; Murray & Greenberg, 2000).

Two dimensions of school support are teacher–student relationships and student engagement. Teacher–student relationships are important to student performance as well as their evaluation of the supportive nature of their school context (Furrer & Skinner, 2003; Skinner, Furrer, Marchand, & Kindermann, 2008). Having caring, involved teachers who provide consistent expectation and are fair in their disciplinary actions is crucial to the way a student appraises their relationships with their teachers and affects their level of school engagement as well as likelihood to engage in risky behaviors (Klem & Connell, 2004; Skinner & Belmont, 1993; Voelkl, 1995). These relationships may also be associated with positive affect. Adolescents who have strong, positive relationships with their teachers may have higher levels of optimism and positive emotions than those who do not. Student engagement refers to student participation in school, school integration, and academic achievement. High achievement and participation has been found to have positive effects on school adjustment and overall academic outcomes. This is often because students who are high achievers and have high grades are often rewarded by teachers and given access to additional opportunities that benefit them.

In addition to the school, the family is one of the most vital socialization agents in a youth’s life and is thought to be essential to the development of positive psychological wellbeing into adulthood (Grusec, 2011). Higher levels of parental support are consistently found to be associated with lower depression (Colarossi & Eccles, 2003; Newman et al., 2007) and higher self-esteem (Hoffman, Ushpiz, & Levy-Shiff, 1988). Cheng et al. (2014) found that among adolescents, perceptions of having a caring adult in the home was positively associated with hope and negatively associated with depressive and post-traumatic stress symptoms.

Communication and closeness are two dimension of family support that have been demonstrated to affect wellbeing. Open communication refers to members of the family feeling heard and able to openly share their own thoughts and opinion. Being able to express one’s opinions, having input in decision making, and a space to talk about personal issues may be important to feelings of happiness and optimism. Closeness refers to the amount of physical and emotional closeness created within the family. A supportive family setting in which youth feel close, connected and supported by family and parents is central to development of self-esteem and other skills necessary for positive mental health outcomes (Bean, Bush, McKenry, & Wilson, 2003; Erikson, 1968). Family closeness, created by positive family interactions helps youth feel a part of a unit, safe, and stable. Such feelings are related to positive decision making and reduced negative behavior (Ackard, Neumark-Sztainer, Story, & Perry, 2006).

In terms of school and family support, research has indicated race and gender influence perception and utilization of support. An extensive body of research has demonstrated teacher perceptions of students are biased against racial ethnic minorities, low income students, and males (Brady, Winston, & Gockley, 2014). This suggests the relationship between school support and positive affect may differ between racial ethnic minority and White adolescents. Further, research has indicated that racial/ethnic minorities and females tend to report higher levels of family support than their counterparts. Research has found family support is integral to young Black Americans academic achievement and self-esteem (Brooks-Gunn, & Markman, 2005; Causey, Livingston, & High, 2015; Taylor & Roberts, 1995). Among Latinos, family support is associated with advantages in mental health (Mulvaney-Day, Alegria, & Sribney, 2007). Walen and Lachman (2000) found family support reduced the effects of stress for girls more so than boys. Consequently, lower perceptions of support have been found to have greater negative effects on girls than boys (Operario, Tschann, Flores, & Bridges, 2006; Størksen, Røysamb, Holmen, & Tambs, 2006). These documented differences suggests that the association between school and family support and positive affect may also differ by race and gender.

The present study examined the relationship between school and family support and positive affect in adolescents in order to better understand the correlates of psychological well-being in the general population. In addition, the population was divided by race (White, Black, and Latino) and gender (males and female) in order to examine whether there were patterns of difference or similarity across sub-populations. The research is guided by the following research questions:

  1. RQ1

    What facets of school and family support are associated with positive affect?

  2. RQ2

    Are those relationships consistent across race and gender?

Method

Sample and Procedure

The present study utilized data from the National Comorbidity Survey: Adolescent Supplement (NCS-A). The NCS-A is a nationally representative, face to face survey of 10,148 adolescents in the United States. Data was collected between February 2001 and January 2004 (Kessler et al., 2009). The NCS-A was carried out at the request of the National Institute of Mental Health (NIMH) as a late addition to the National Comorbidity Survey Replication (NCS-R) to meet a request from Congress to provide national data on the prevalence and correlates of mental health indicators among US youth (Kessler et al., 2009). The NCS-A was designed to estimate the lifetime-to-date and current prevalence, age-of-onset distributions, course, symptoms, and comorbidity of DSM-IV disorders among adolescents in the United States; to identify risk and protective factors for the onset and persistence of these disorders; to describe patterns and correlates of service use for these disorders; and to lay the groundwork for subsequent follow-up studies that can be used to identify early expressions of adult mental disorders (Kessler et al., 2009). The overall response rate was 82.9%.

To ensure the target sample of 10,000 adolescents was reached, the NCS-A relied on a dual frame design that added a school based sample to the household sample. In the household sample, one random adolescent was selected by a computer program when more than one adolescent resided in the household. In the school sample, the adolescent was identified by the school roster. A representative sample of all accredited eligible schools was selected with probabilities proportional to the size of the student body in the classes relevant to the target sample in each of the counties or county clusters that made up the primary sampling units (PSUs) of the nationally representative National Comorbidity Survey-Replication sample. Within each school, a random sample of 40–50 eligible students was selected for sampling using a systematic selection procedure. Recruitment and consent procedures were overseen and approved by both the Human Subjects Committees of Harvard Medical School and the University of Michigan.

After data collection, cases were weighted for variation in household probability of selection in the household sample and residual discrepancies between sample and population sociodemographic and geographic distributions. The household sample weights were already developed for the National Comorbidity Survey-Replication. They were added to the adolescent data and adjusted for differential probability of selection of adolescents in the household. This data was then compared with nationally representative census data on basic socio-demographic characteristics for purposes of post-stratification. Weighting for the school sample was based on weights that controlled for three sets of variables: quality education data which includes data on the characteristics of all schools in the US, public use microdata sample (PUMS), and block group (BG) level data. More detailed information regarding weighting procedures can be found in Kessler et al. (2009).

The sample consisted of 10,148 adolescents (5648 Whites, 1955 Blacks and 1922 Latinos). Due to the small number of respondents who identified as “other” (n = 623) and the potential ethnic variation within that subgroup, this category was omitted from the analysis. The sample included 4965 males and 5183 females. The average age was 15.2 years (Table 1). Recruitment and consent procedures were approved by the Human Subjects Committees of both Harvard Medical School and the University of Michigan.

Table 1 Demographic characteristics of sample
Full size table

Measures

Positive Affect

Positive affect was a four item scale measured with the following questions: (1) In the past 30 days, how often did you feel confident? (2) In the past 30 days, how often did you feel optimistic? (3) In the past 30 days, how often did you feel happy? (4) In the past 30 days, how often did you feel full of life? Response choices ranged from 1 (“all of the time”) to 5 (“none of the time”).

School Attachment

Four questions were used to assess school attachment: (1) Most of my teachers treat me fairly? (2) Getting good grades is important to me? (3) I like my teachers? (4) I try hard at school? Response choices included “Very” “Somewhat” “Not very” “Not at all”. Each item was treated as a separate indicator.

Family Communication

Three questions were used to assess family communication (1) How often family members easily expressed opinions? (2) How often did each family member have input in major family decisions? (3) How often family members talk about feelings? Response choices included “All of the time” “Most of the time” “Some of the time” “Never”. Each item was treated as a separate indicator.

Family Closeness

Three questions were used to assess family closeness: (1) How often family members felt close to each other? (2) How often family members did things together? (3) How often family shared interests and hobbies? Response choices included “All of the time” “Most of the time” “Some of the time” “Never”. Each item was treated as a separate indicator.

Demographics

Race/ethnicity was a categorical variable that measured self-reported racial identification: Hispanic, non-Hispanic Black, non-Hispanic White (reference category), or other. Gender was a binary variable coded male “0” and female “1”. Age was a continuous variable that ranged from 13 to 18.

Results

In the baseline model (Table 2, model 1), I examined the correlation between school and family support and positive affect for the total sample (RMSEA .037 CFI .953). All of the indicators were predictors of positive affect suggesting student–teacher relationships, student engagement, and family communication and closeness were associated with positive affect. Age and race were not predictors of positive affect, but gender was an indicator of positive affect. Males reported higher positive affect than females.

Table 2 SEM of school and family support predicting positive affect by gender
Full size table

To examine the differential effects of gender on the association between the focal school and family support variables and positive affect (Table 2 model 2), a structural equation model was conducted on the male and female subsamples (RMSEA .037 CFI .961). Among the male subsample, having a teacher treat them fairly (β = .06, p < .001), liking their teachers (β = .06, p < .001), and their own efforts (β = .08, p < .001) was associated with positive affect. Getting good grades was not a predictor of positive affect in males. Being in a family where members had input on family decision (β = .03, p < .001) and they could talk about their feelings (β = .03, p < .001) was associated with positive affect, but being in a family where member easily expressed their opinions was not associated with positive affect in males. Feeling close to family members (β = .06, p < .001), being in a family where members do things together (β = .05, p < .001), and being in a family where members shared interests and hobbies (β = .03, p < .001) was associated with positive affect in males. Family members having input in family decisions (β = .03, p < .001) and being in a family with frequent communication (β = .03, p < .001) were associated with positive affect in males, but feeling as if they could easily express their opinions was not associated with positive affect.

All of the school support variables were associated with positive affect in the female subsample. Being in a family where members easily expressed their opinions (β = .03, p < .01) and where family members talked about their feelings (β = .04, p < .001) was associated with positive affect while member input was not. Doing things together as a family was not associated with positive affect in the female subsample but feeling close to family (β = .09, p < .001) and having shared interests and hobbies (β = .02, p < .05) was associated with positive affect. Age had opposite effects on positive affect in males and females. Aging was associated with increases in positive affect in males and deceases in positive affect in females. Race had no effect on positive affect on females, but among males, racial/ethnic minorities reported lower positive affect than Whites.

To examine the differential effects of race on the association between the focal school and family support variables and positive affect (Table 3 model 3), a structural equation model was conducted on the White, Black, and Latino subsamples (RMSEA .037 CFI .961). White adolescents closely resembled the total sample. Getting good grades was the only measure of support that was not a significant predictor of positive affect in Whites. All other measures of school and family support were associated with positive affect in the White subsample. Among Black adolescents, teacher treatment (β = .09, p < .001) and their own engagement (β = .14, p < .001) was associated with positive affect whereas liking teachers was not. Being in a family that frequently discussed feelings (β = .04, p < .05) was associated with positive affect but being in families where members could express their opinions or where members had input on family decisions was not associated with positive affect. Feeling close (β = .08, p < .001) to members of family was associated with positive emotion; doing things together as a family or being in a family with shared interests had no association with positive affect.

Table 3 SEM of school and family support predicting positive affect by race
Full size table

Among Latinos, having teachers treat them fairly, was not associated with positive affect. Their own appraisal of whether they liked their teachers (β = .09, p < .001) and their school engagement were predictors of positive affect. Being in a family where family members had input on family decisions (β = .04, p < .05) and families that frequently talk about their feelings (β = .04, p < .05) was associated with positive affect. Having shared interests and doing things together was not associated with positive affect among Latinos but feeling close to family members (β = .09, p < .001) was associated with feelings of happiness among all racial groups, males reported higher positive affect than females.

Discussion

The primary purpose of the present study was to examine the correlates of positive affect in a nationally representative sample of adolescents in the United States. Particular attention was paid to school and family support because they are the primary sources of support and socialization agents in adolescents’ lives. When examining the total sample, findings showed that dimensions of school and family support were associated with positive affect in adolescents. Findings from this study suggests teacher-student relationships characterized by fair treatment by the teacher and students liking their teachers were robust predictors of positive affect. As gatekeepers and authority figures, teachers are forceful figures in adolescents’ lives (Murray & Greenburg, 2000; Wentzel, 1997). This analysis demonstrates adolescents’ relationships with them are associated with not only their achievement but also their mental health. But it is not only relationships with authority figures that makes school support so integral to positive affect, adolescents’ own engagement was also a strong predictor of positive affect. This may be because adolescent who try hard and buy into the importance of grades may be reaping benefits such as, positive attention from teachers and additional educational opportunities, which in turn is related to positive affect.

The family was also a key source of support for adolescents and the present study examined the dimensions of family support that were associated with positive affect. Analysis demonstrated that being in a family that felt physically close as well as a family with open communication was associated with positive affect. But the importance of emotional closeness cannot be understated. Families that valued talking about emotions and felt close to each other were robust predictors of positive emotions in the total sample and across race and gender subsamples. This suggests that feelings of optimism and happiness in adolescents is dependent not just on the physical time or spent with family members. It is the quality of that time, and the centrality of emotional wellbeing that is key.

When the population was disaggregated by gender, the factors of school support that were predictors of positive affect were overwhelmingly similar between males and females with one exception. For males, getting good grades was not a predictor of positive affect; however, it was a predictor of positive affect for females. Previous research has found that girls are more likely to be high achievers and work for good grades than boys (Duckworth & Seligman, 2006). This research adds to that research by suggesting that the extra academic engagement produced by girls has effects on their emotional wellbeing in addition to their scholastic achievement.

The facets of family support that were associated with positive affect were specific to males and females as well. For both males and females, families that talked about feelings was associated with happiness. However, equity in family decision making was associated with positive affect in males but not in females, and ease in expressing opinions was associated with positive affect in females and not males. These differences suggest males and females are differentially impacted by communication within the home. Interestingly, aging was associated with increases in positive affect in males but decreases in positive affect in females. This suggests the aging process along with its physiological and social changes have stronger negative effects on female positive wellbeing than their counterparts. Race was not a predictor of positive affect among females, but among males, identifying as racial/ethnic minority was associated with lower positive affect than White males.

When the population was disaggregated by race, the facets of school and family support that were correlated with positive affect varied among the different racial groups. Whites most closely resembled the total sample. Except for getting good grades, all of the dimensions of school and family support were correlates of positive affect among White adolescents. While Black adolescents’ perceptions of whether or not they liked their teachers was not correlated with positive affect, being treated fairly by their teachers was a strong predictor of positive affect. A substantial body of literature has demonstrated teachers hold biases against Black students, often affecting how they treat them such as recommendations for tracking or disciplinary actions (Monroe, 2005). This study demonstrates that how teachers treat Black students has a marked effect on their feelings of happiness and optimism. Interestingly, among, Latinos, teacher treatment was not correlated with positive affect. This is despite that fact that research has found that teachers are most often biased against Latinos as well (Skiba, Horner, Chung, Raush, & May, 2011). Instead, Latino adolescents own appraisal of their teacher was a stronger predictor of positive affect.

For Blacks and Latinos, being in families where there was open communication about feelings was a stronger predictors of positive affect than being in families that were physically close. Needing to be do things together and shared interests were not predictors of positive affect in either Blacks or Latinos. Talking about feelings was marginally significant for both groups. Among all groups, females had lower positive affect than males.

These differences suggest the relationship between support mechanisms and psychological wellbeing are not the same across race and gender. Although the school and family are generally considered to be sources of support, when specific dimensions of support are examined there are notable differences across the population. Differences between males and females and among different racial/ethnic groups demonstrates that evaluation of psychological wellbeing and its correlates must also consider that the correlates of mental health are greatly affected by demographic characteristics, cultural memberships, and identities.

These findings are not interpreted without caution. The NCS-A is a cross-sectional data set which limited the ability to examine the relationship between school and family support and positive affect across time. As adolescents age, the relationship between support systems and psychological wellbeing may change as they rely on newly developing friendship groups and begin to leave school and become independent of their families. Future studies should utilize longitudinal data in order to understand the nature of these relationships across time. Second, without consideration of other explanatory variables the findings are interpreted with caution. Positive affect is a composite measure that captures feelings of happiness, optimism, and feeling free from tension. This study should be replicated with various measures of self-esteem, optimism, happiness, mattering and other facets of psychological wellbeing in order to understand these relationships more in depth. Finally, while teacher–student relationships and student engagement were used as measures of school support and family closeness and communication were used as measures of family support, the school and family support networks are complex multidimensional networks that are not fully represented here. Other dimensions of school and family support should be evaluated in its relation to psychological wellbeing in order to gain more information about the correlates of mental health. Finally, the omission of such variables may partially explain the small magnitudes of the coefficients as well as the low amount of explained variance. Notwithstanding the limitation, this study has implications for assessment and intervention.

Implications

The findings from this study have implications for assessment of psychological wellbeing in adolescents. As researchers and practitioners pursue more information about adolescent mental health, assessment of wellbeing is becoming as important as assessment of mental disorder and illness. Understanding the correlates of wellbeing can aid in understanding the distribution of mental wellbeing in the general population. These findings are also relevant in social work because they have significant implications for mental health interventions. Planning interventions requires a tailored message that targets the specific population being affected instead of generic messages with no intended audience. The nuances in family and school attachment by race and gender, suggests that each group would benefit from specific interventions. For girls, for example, interventions at the school level should include consideration of their academic engagement and grades whereas focus on grades may not be as essential for boys. In addition, school based interventions for Black students specifically should involve improving their relationships with their teachers. Practitioners working with this population should implement assessment and intervention strategies that are mindful of these differences.