Academic Commons

Theses Doctoral

Changes in gendered social position and the depression gap over time in the United States

Platt, Jonathan M.

Introduction: There is a large literature across disciplines aimed at understanding the causes of the depression gap, defined as an excess of depression among women compared with men. Based on the totality of evidence to date, social stress appears to be an important explanation for the depression gap. Social stress theory highlights women’s disadvantaged social position relative to men, positioning gender differences in socio-economic opportunities as social stressors, while also acknowledging how gender socialization teaches women to respond to stressors in depressogenic ways from an early age. This dissertation applied social stress theory to better understand the social causes of the depression gap with three related aims. Aim 1 summarized the evidence for variation or stability in the depression gap in recent decades, through a systematic review and meta-regression of depression gap studies over time and by age. Aim 2 examined the evidence for a changing depression gap across birth cohorts, and tested the extent to which any changes over time were mediated by changing gender differences in education, employment, and housework rates, three indicators of broader trends in gendered social position through the 21st Century. Aim 3 examined whether women in the workforce with competing domestic labor roles were at increased risk of depression, and whether pro-family workplace benefits buffered the effects of competing roles.

Methods: In aim 1, depression gap estimates were extracted through a systematic review of published literature (from 1982-present). Analytic datasets were comprised of 76 diagnostic-based estimates and 68 symptom-based estimates. For each dataset, meta-regression models estimated time and age variation in the depression gap, as well as the interaction between time and age group, to estimate the variation in the gap over time by age. Data from the National Longitudinal Surveys were utilized for aims 2 and 3. Depression was measured with the Center for Epidemiologic Studies Depression Scale (CESD), and the depression gap was defined as differences in mean CESD scores for women vs. men. The aim 2 sample included 13,666 respondents interviewed from 1992-2014. Hierarchical mixed models estimated the magnitude of the gender depression gap over time, and its relationship with 10-year birth cohort (range: 1957-1994) and whether any variation was mediated by gender differences in: those with a college degree or more, those who were employed full-time, and the average number of hours spent doing housework per week. The sample in aim 3 was limited to employed women ages 17-57 (n=3993). Generalized estimating equations estimated the relationship between competing roles and depression, and the interaction between competing roles and pro-family employee benefits on depression. Interaction results were compared to models estimating the effect of non-family-related benefits on the relationship between competing roles and depression.

Results: In aim 1, there was no evidence of change in the depression gap over time. Compared with the reference group (i.e., respondents ages 60+), the age effect was appreciable among the youngest age group (age 10-19) (RR=1.44; 95% CI=1.19, 1.74), but did not differ for any other age groups. The age by time interaction was elevated for youngest age group (RR=1.27; 95% CI=1.0, 1.61), suggesting that, compared to the oldest age group, the diagnostic depression gap had increased among the youngest ages from 1982 to 2017. There was no evidence of time changes among any other age group. Results were similar for symptom-based studies.

In aim 2, there was a linear decrease in the depression gap by 0.18 points across birth cohort (95% CI= -0.26, -0.10). The results of the mediation analysis estimated that an increasing ratio of college degree attainment mediated 39% of the gender depression gap across cohorts (95% CI= 0.18, 0.78). There was no evidence of mediation due to changing employment or housework ratios.

In aim 3, there was evidence that women in competing roles reported a 0.56-point higher CESD score (95% CI= 0.15, 0.97), compared with women not in competing roles. The interaction between pro-family benefits and competing roles was associated with CESD scores (B=-0.44, p=0.023). More specifically, among women without access to pro-family benefits, those in competing roles reported a 6.1 point higher CESD score (95% CI=1.14, 11.1), compared with those not in competing roles, however, among women with access to these benefits, there was no association between competing roles and CESD scores (difference=0.44; 95% CI=-0.2, 1.0). Results were similar for non-family-related benefits. Women in competing roles without non-family-related benefits reported a 3.59 point higher CESD score than those not in competing roles (95% CI=1.24, 5.95) while among women with access to these benefits, there was no association between competing roles and CESD symptoms.

Conclusion: This dissertation provided evidence to partially support the hypothesis that the depression gap is changing over time and is meaningfully related to the social environment, through which gender roles, responsibilities, and opportunities available to women and men are defined and reinforced. The results of these studies suggest that the depression gap may be expanding and contracting over time for different age groups. Understanding the social causes of the depression gap is important to reduce the present and future burden of the depression gap, and to understand the fundamental processes through which depression disparities may be perpetuate or attenuated in adolescence and beyond.


  • thumnail for Platt_columbia_0054D_15622.pdf Platt_columbia_0054D_15622.pdf application/pdf 811 KB Download File

More About This Work

Academic Units
Thesis Advisors
Keyes, Katherine M.
Ph.D., Columbia University
Published Here
November 11, 2019