- Top of page
Although several longitudinal studies have demonstrated that having a disadvantaged family background is a risk factor for subsequent symptoms of depression, few studies have examined the mediating mechanisms that explain this long-term relationship. Thus, this study uses US national longitudinal data and integrates social stress theory with the life course perspective by focusing on two mediating mechanisms—the chronic stress of poverty and self-esteem during the transition to adulthood. Results reveal that self-esteem largely mediates the inverse relationship between parental education and levels of depressive symptoms in young adulthood. However, the inverse relationship between parental occupational prestige and depressive symptoms among young adults is not mediated by self-esteem, but rather long durations of poverty across 16 years. Overall, these findings suggest that different components of family socioeconomic status can leave a lasting imprint on mental health via the self-concept and the chronic stress of poverty throughout the journey to adulthood. Copyright © 2013 John Wiley & Sons, Ltd.
- Top of page
A number of longitudinal studies have demonstrated that having a disadvantaged family background is a risk factor for symptoms of depression in subsequent life stages (Duncan, Brooks-Gunn, & Klebanov, 1994; Elovainio et al., 2012; Gilman, Kawachi, Fitzmaurice, & Buka, 2002; Goosby, 2007; Harley & Mortimer, 2000; Harper et al., 2002; Kosidou et al., 2011; Lee, Wickrama, & Simons, 2013; McLaughlin et al., 2011; McLeod & Shanahan, 1993, 1996; Melchior et al., 2013; Mossakowski, 2008a; Power, Stansfeld, Matthews, Manor, & Hope, 2002; Ritsher, Warner, Johnson, & Dohrenwend, 2001; Stansfeld, Clark, Rodgers, Caldwell, & Power, 2010; Turner, Taylor, & Van Gundy, 2004). More longitudinal research, however, needs to examine the mediating mechanisms that explain this relationship early in the life course (Franziska, 2013). Furthermore, life course researchers argue that it is imperative to investigate longitudinally both risk factors and protective factors during the transition to adulthood because early adulthood is the stage at which the onset of depression can occur (Kessler et al., 2003). According to social stress theory, a major risk factor for depression is exposure to stress, whereas vulnerability to stress depends on protective factors, such as psychosocial resources (Pearlin, Menaghan, Lieberman, & Mullan, 1981). Guided by social stress theory and the life course perspective, the current study uses national longitudinal survey data to focus on two mediating mechanisms—self-esteem (a psychosocial resource) and the chronic stress of poverty (a risk factor) during the transition to adulthood—to help explain why disadvantaged family background can lead to symptoms of depression among young adults in the United States.
- Top of page
Table 1 shows the descriptive statistics, which include the means and ranges for the variables. As a sensitivity analysis, Table 1 compares the descriptive statistics for the sample (n = 8215) in column 1 with the smaller sample in column 2 (n = 6936). The sample in column 1 is used for the OLS regression analyses that examine whether prior self-esteem mediates the relationship between disadvantaged family background and depressive symptoms. The smaller sample in column 2 is used for the analyses of the mediating effect of poverty duration across multiple NLSY79 waves. The descriptive statistics in columns 1 and 2 are consistent, which suggests that the listwise deletion of missing data did not bias the results.
Table 1. Descriptive statistics: means and ranges
| ||n = 8215 *||n = 6936 *|| |
|Family background|| || || |
|Parental occupational prestige†||34.00||34.37||0–86|
|Demographics|| || || |
|Mental health|| || || |
|Prior Depressive symptoms||9.73||9.68||0–60|
|Poverty status|| || || |
According to Table 1, the average level of parental education is approximately 12 years (high school), and the average parental occupational prestige corresponds to occupations in sales. Column 1 indicates that the levels of depressive symptoms are within the range to be expected for a general population (Radloff, 1977, 1991), and the mean level of self-esteem is 32 with a minimum of 17 and a maximum of 40. Column 2 shows that the average duration of poverty from 1979 to 1994 is 3 years.
Model 1 in Table 2 evaluates the strength of the effects of parental education and parental occupational prestige on levels of depressive symptoms at ages 29 to 37 by controlling for gender, age, race/ethnicity and marital status. According to the standardized OLS regression coefficients in model 1, lower levels of parental education (beta = −0.054; p < 0.001) and parental occupational prestige (beta = − 0.059; p < 0.001) are statistically significant predictors of higher subsequent levels of depressive symptoms, adjusting for demographics. These findings confirm the first hypothesis of this study. Also, supplementary analyses reveal that parental education and parental occupation explain 2% of the variance in levels of depressive symptoms in young adulthood.
Table 2. Ordinary least squares regression models of family background, demographics,and prior mental health predicting symptoms of depression at ages 29 to 37
|Variables||Model 1||Model 2||Model 3||Model 4|
|Family background|| || || || |
|Demographics|| || || || |
|Prior mental health|| || || || |
|Prior depressive symptoms¶||—||—||0.387***||−0.083***|
Model 2 in Table 2 examines the second hypothesis, whether prior self-esteem at ages 15 to 23 is a mediating mechanism in the relationship between family background and depressive symptoms at ages 29 to 37. The effect of parental education on depressive symptoms is substantially reduced and no longer statistically significant when prior self-esteem is added to the equation, which suggests that it is a mediating mechanism (Baron & Kenny, 1986). More specifically, prior self-esteem explains more than half (53.7%) of the depressive effect of low parental education. As expected, lower levels of prior self-esteem predict (beta = −0.160; p < 0.001) increased levels of depressive symptoms. Supplementary analyses (available on request) indicate that parental education has a statistically significant and positive effect (beta = 0.180; p < 0.001) on self-esteem, which provides further evidence that self-esteem is a mediating mechanism in the relationship between earlier parental education and depressive symptoms in young adulthood (Baron & Kenny, 1986). Additionally, the Sobel test confirms that self-esteem is a significant partial mediator (z = −9.814; p < 0.001) (Baron & Kenny, 1986). Thus, the second hypothesis is proven for parental education. Interestingly, the relationship between low parental occupational prestige and depressive symptoms in model 2 is relatively unchanged (beta = −0.052; p < 0.001) compared with model 1 (beta = −0.059; p < 0.001). Prior self-esteem does not mediate that relationship.
Model 3 in Table 2 adjusts for prior depressive symptoms to evaluate the strength of the intergenerational social causation of depression by accounting for selection effects. A mental health selection effect is when earlier symptoms of depression mediate or explain why past parental SES is linked with subsequent symptoms of depression in young adulthood. Results in model 3 reveal that parental education and parental occupation still have statistically significant, negative relationships with depressive symptoms over and above controlling for prior depressive symptoms, which is evidence for social causation. Supplementary analyses that used the abbreviated 7-item CES-D in both waves indicated that disadvantaged family background predicted an increase in levels of depressive symptoms between the two time-points.
Model 4 in Table 2 adjusts for demographics, prior self-esteem and prior depressive symptoms. In model 4, the inverse relationship between parental education and depressive symptoms is further reduced and is not statistically significant (beta = −0.016) because of the inclusion of self-esteem. The inverse effect of parental occupational prestige on depressive symptoms, however, is slightly reduced and continues to be significant (beta = −0.044; p < 0.01), net of demographics and prior mental health.
Table 3 examines whether cumulative exposure to poverty status across 16 years is a mediating mechanism in the lasting relationship between parental SES and depressive symptoms in young adulthood. Model 1 in Table 3 is a sensitivity analysis because the sample size is smaller for the analyses of poverty duration because of listwise deletion of missing data. Results in model 1 of Table 3 do not substantively differ from the results in model 1 of Table 2: parental education (beta = −0.057; p < 0.001) and parental occupational prestige (beta = −0.059; p < 0.001) have significant negative effects on levels of depressive symptoms, adjusting for demographics.
Table 3. Ordinary least squares regression models of family background, demographics, prior mental health and poverty duration predicting symptoms of depression at ages 29 to 37
|Variables||Model 1||Model 2||Model 3|
|Family background|| || || |
|Demographics|| || || |
|Prior Poverty|| || || |
|Poverty duration§|| ||0.165***||0.085***|
|Prior mental health|| || || |
|Prior self-esteem¶|| || ||−0.072***|
|Prior depressive symptoms∥|| || ||0.356***|
Poverty duration is added to model 2 in Table 3. The depressive effect of low parental education decreases by 38.6% compared with model 1 and has a lower level statistical significance (p < 0.05). According to the Sobel test, poverty duration is a significant partial mediator (z = −6.882; p < 0.001). The depressive effect of low parental occupational prestige is reduced by almost 50% and is no longer statistically significant while controlling for poverty duration. Supplementary analyses indicate that low parental SES is associated with significantly longer durations of poverty. The Sobel test further verifies that poverty duration is a significant partial mediator (z = −7.509; p < 0.001) of the relationship between parental occupational prestige and symptoms of depression. Additionally, results in model 3 in Table 3 show that poverty duration is a significant risk factor (b = 0.085; p < 0.001) for depressive symptoms at ages 29 to 37, independent of demographics, prior depressive symptoms and prior self-esteem. This is further evidence for social causation. As a whole, the results in Table 3 support the third hypothesis of this study. Results indicate that poverty duration partially mediates the effect of parental occupational prestige at age 14 on depressive symptoms at ages 29 to 37, and partially mediates the effect of parental education but to a lesser extent than self-esteem as a mediator. Like the results in Table 2, the effect of parental education on depressive symptoms becomes substantially diminished and loses its statistical significance because prior self-esteem is included in model 3 of Table 3 and is a mediating mechanism.
- Top of page
Our knowledge remains limited about the mediating mechanisms that explain the lasting mental health consequences of having a disadvantaged family background in the United States. Thus, this study was inspired by social stress theory and the life course perspective to evaluate self-esteem and the chronic stress of poverty as mediating mechanisms. The central findings of this study are that these are important mediating mechanisms that help to explain this long-term relationship. Furthermore, the findings contribute to the literature by demonstrating two distinct social stress processes with national longitudinal data to bring us closer to understanding the intergenerational social causation of depression.
More specifically, the findings suggest that the two components of family background (parental education and parental occupational prestige) each have distinct mechanisms that link them to mental health in young adulthood. First, results reveal that self-esteem at ages 15 to 23 is a mediating mechanism that helps to explain why there is an inverse association between parental education and subsequent levels of depressive symptoms at ages 29 to 37. Yet, self-esteem does not mediate the relationship between parental occupation and depression. Prior studies have found that father's education has a positive effect on self-esteem in adolescence, but father's occupation does not significantly influence self-esteem (Rosenberg & Pearlin, 1978; Wiltfang & Scarbecz, 1990). According to social stress theory, self-esteem is a vital dimension of the self-concept through which the stress of disadvantaged SES can damage mental health (Pearlin et al., 1981). Second, the current study finds that the duration of family poverty status across 16 years largely mediates the depressive effect of disadvantaged parental occupational prestige and parental education. It is plausible that parents with low prestige jobs and limited educations could have household incomes that fall below the poverty line for long periods, and the family's exposure to the chronic stress of economic hardship and job insecurity could elevate their children's risk for symptoms of depression. From the standpoint of the life course perspective, chronic stress is a more powerful predictor of depression than acute stressors (Avison & Turner, 1988). Few studies have measured chronic stress with longitudinal data at multiple points in time. More longitudinal studies should examine specific components of parents' SES (e.g. education, occupational prestige, poverty and net worth) and identify other unique mediating mechanisms or pathways that link each to the onset, persistence and severity of the next generation's mental health problems (Adler & Newman, 2002; McLaughlin et al., 2011).
Although many American families have experienced persistent or episodic poverty in the course of their lives, only a few longitudinal studies have examined the influence of the duration of poverty on mental health (e.g., Duncan et al., 1994; Goosby, 2007; McLeod & Shanahan, 1993, 1996; Mossakowski, 2008a, 2008b). An advantage of the NLSY79 data used for the present study is that family poverty status was measured annually across 16 years. Findings indicate that the average duration of poverty status was 3 years. ‘The experience of poverty can thus be viewed within the wider context of the life course as a normative American event’ (Rank, 2001, p. 886). It is noteworthy that the duration of poverty measured in this study is not necessarily consecutive years, and poverty was assessed annually. However, the stress of poverty across a number of years can be perceived to be chronic even when poverty is episodic because these families are ‘living on the edge’ (Ross & Huber, 1985, p. 313). An annual household income that is slightly above the cusp of the poverty line can be a financial predicament that brings with it constant worries about slipping into poverty again, and thus, it can become a daily stressor. Future studies on mental health should measure consecutive years and/or continuous poverty throughout each year to better assess the stress of chronic poverty and use self-reports to determine whether it is actually perceived to be a chronic stressor. Another direction for future research is the use of biomarkers (e.g. cortisol) to test physiological levels of chronic stress due to poverty.
It is possible that the economic recessions in the 1980s and 1990s could have affected the durations of poverty experienced and the mental health of the NLSY79 cohort of younger American Baby Boomers examined in the present study. Research has shown that economic downturns are associated with mental health problems (Cooper, 2011; Tausig & Fenwick, 1999). Accordingly, the findings of this study are timely because they suggest that the recent economic recession in the United States and its detrimental impact on the SES of parents could have long-term and harmful mental health consequences for their children because symptoms of depression could eventually manifest during young adulthood. Furthermore, the Baby Boomers of the NLSY79 are now between the ages of 49 to 56, and research warns that if they experienced poverty and suffered from depression throughout the transition to adulthood and into midlife, they could have a significantly elevated risk of developing physical health problems (Power & Hertzman, 1997). Future research is also needed to examine the influence of poverty duration on the mental health of young adults today and other generations of young adults, such as Generation X and the Boomerang Generation, as well as cohorts in other countries. The findings could be different in other countries, economic climates, life stages and historical periods.
The NLSY79 used for this study is one of the premier datasets to examine SES during the transition to adulthood, but it was not originally intended to investigate the aetiology of mental illness. If depression was measured before the 1992 wave, it would have been a better test for selection effects. A limitation of this study is that simultaneous trajectories of depression, self-esteem and SES could not be analysed. Recent research has shown a socioeconomic gradient in long-term depression trajectories across 13 years, indicating that low SES can lead to persistent depression (Melchior et al., 2013), but the effect of childhood SES on symptoms of depression appears to weaken over time (Elovainio et al., 2012). By discovering the social and psychological antecedents of depression during the passage to adulthood, we can better target the prevention of mental illness in young adulthood and later life stages.
It should be acknowledged that there are other reasons not addressed by this study that could explain why coming from a family that had disadvantaged SES increases the risk of mental illness. One predominant theory is that the influence of parental SES on their children's well-being is explained by parenting behaviours (Conger et al., 1994; Harris & Marmer, 1996). For example, unsupportive parenting behaviours involve neglect, abuse, withholding affection and being unresponsive to a child's socio-emotional needs, which can be traumatic experiences. In contrast, parenting that provides a supportive and stimulating environment can nurture resilience in a child, which is useful for counteracting the stress of poverty (Bradley et al., 1994). Studies have found, however, that parenting attitudes and behaviours vary within economically deprived family environments (Bradley et al., 1994; Conger et al., 1994).
Another theory is that if parents do not cope adequately with their own stressful life events or financial strain, especially single parents, it can increase the risk that their children will develop depressive symptoms in adolescence and young adulthood, and this relationship is mediated by parents' depressive symptoms, parenting behaviours and SES (Barrett & Turner, 2005; Ge et al., 1994). A limitation of the current study is that the mental health of the parents of the NLSY79 young adults was not assessed. A study using data from the Children of the NLSY79 Survey, however, found that the influence of maternal poverty duration on an adolescent's depressive symptoms is mediated by the mother's depressive symptoms, regardless of her marital status and supportive parenting behaviours (Goosby, 2007). Thus, the stress of economic hardship can negatively affect child development by intensifying parents' distress.
To advance our understanding of the link between family SES and depression in young adulthood, we need to acknowledge perspectives that range from the microlevel to the macrolevel. A microlevel lens focuses on the social and psychological development of the child in the family environment and the interaction of stress with the child's genetics. Social scientists should not overlook that genetic expression can be triggered by stress in a family environment. Research indicates that childhood deprivation can exacerbate the inherited predisposition for depression, whereas protective genotypes can suppress the effect of childhood adversity, and social network structures (e.g. family social support) can minimize genetic risk (Caspi et al., 2003; Pescosolido et al., 2008). Next, the mesolevel lens expands the focus to include the economic conditions of the family's low-income neighbourhood context (e.g. social disorder and crime) and limited access to healthcare. Finally, the macrostructural perspective more broadly acknowledges SES at the level of the national economy and the cultural norms in a historical period. Future interdisciplinary studies that use structural equation modelling and hierarchical linear modelling may be ideal ways to explore these complex, multilevel indirect pathways from family SES to mental illness, which involve the intersections of unobserved latent variables, social antecedents, psychological predispositions, stressors and behaviours, as well as genetics.
There are public policy implications for the current study's findings that focus on family SES, the chronic stress of poverty, self-esteem and depression during the transition to adulthood. Public policy initiatives, social welfare programmes and medical interventions should not only target the reduction of poverty duration and improve access to mental health treatment but also aim to prevent poverty and mental illness early in the life course. From a macrostructural perspective, social policies that stimulate economic growth and reduce wealth inequality will benefit the economic and psychological well-being of the US population as a whole.
Public policies need to also target socioeconomic attainment at the individual-level, such as improving access to college educations and creating better bridges from school to work and to homeownership. First, to achieve a higher education and other credentials valuable in the labour market, poverty alleviation programmes need more funding to provide scholarships, grants, tutoring and improve schools in low-income neighbourhoods for youths. Second, for youths to find employment and remain employed, welfare-to-work programmes, employment agencies, retraining courses and internships with employers need to be expanded (Cooper, 2011). Third, family support programmes for low-income groups should include childcare allowances and better tax incentives for families to buy their first homes in safe neighbourhoods. More support is needed for neighbourhood stabilization initiatives designed to fund systemic reform to mitigate the impact of the foreclosure crisis, and more debt relief programmes are needed to help build a family's net worth to escape or prevent poverty (Cooper, 2011).
To prevent or treat symptoms of mental illness during the transition to adulthood, improving access to mental health services is an ongoing goal for healthcare reform. The Affordable Care Act now allows young adults to stay on their parents' health insurance plans until age 26. Unfortunately, young adults have the lowest access to employer-based health insurance because they often enter the labour market in low prestige jobs and part-time positions: approximately one-third of young adults are uninsured, which is a higher rate than any other age group in the United States (U.S. Department of Labor, 2013). To prevent mental health problems among youths and young adults, affordable community mental health services need more funding, as well as interventions in elementary schools, high schools and colleges so that there is better access to free psychological counselling during school days to address coping with stressors from different environments—school, family and their community—to enhance resiliency, such as via self-esteem (Reback, 2010). Yet, some scholars question American culture's preoccupation with an individual's self-esteem as a psychosocial resource and caution that it may not necessarily be the key to success and health (Baumeister et al., 2003; Crocker & Park, 2004).
In conclusion, more longitudinal research is needed to explore other resources for coping with stress (e.g., perceived social support and a sense of personal control over life) to guide public health policies, medical interventions and social welfare programmes. Reducing socioeconomic health disparities requires policy initiatives that acknowledge the different dimensions of family SES and their distinct pathways early in the life course to prevalent mental disorders, such as depression. Finally, international research should investigate whether long durations of disadvantaged family SES lead to depression in countries that have less of an emphasis on self-esteem and very different social welfare policies than the United States.