Associations between women's retention in STEM or STEM-related fields and their spouses' occupations and majors

The study addresses how spouses’ occupations and majors relate to married women’s retention in STEM or STEM-related (healthcare) occupations. Motivated by women’s persistent underrepresentation and higher exit rates from STEM, and by evidence that marriage and work–family dynamics shape women’s careers, the paper asks: What is the relationship between a spouse’s STEM or STEM-related occupation/degree and whether married women with STEM or STEM-related majors remain in those occupations? It further tests whether this association differs by field (STEM vs healthcare) and by the presence of children. The hypothesis, grounded in literature on spousal homogamy and informational support, is that having a spouse with a STEM or STEM-related occupation/major is positively associated with women’s retention in these fields. The work situates the question within U.S. labor market trends, gendered work–family constraints, and evidence that occupational similarity in couples can shape career development.

Prior research indicates that marriage and family dynamics, including spousal support, gender-role attitudes, and work–family conflict, influence women’s careers. Studies note a marriage premium differs by gender and that family constraints can hinder married women’s optimal career decisions (Dougherty, 2006; McKinnish, 2008). Spousal careers can negatively affect wives’ labor force participation yet provide resources aiding career attainment; occupational similarity can increase informational support (Bernardi, 1999; Wallace & Jovanovic, 2011). Spousal support relates to career success via work–family balance, and spouses’ aspirations shape career paths (Amin et al., 2017; Pluut et al., 2018). Women remain underrepresented in STEM due to stereotypes, norms, and climates, and have higher exit rates and more difficulty re-entering (Beede et al., 2011; Glass et al., 2013; Mavriplis et al., 2010; Casad et al., 2021). Family-related challenges (pregnancy, caregiving) are salient for women in STEM (Jean et al., 2015). Romantic goals may conflict with STEM persistence, though supportive partners can bolster women’s educational and career goals (Park et al., 2011; Barth et al., 2016). Educational/occupational homogamy is common in highly educated populations, including science and engineering doctorates (Case, 2013; Blossfeld, 2009; Arum et al., 2008). Few studies differentiate STEM from healthcare; this study does so, defining STEM to include life/physical/social sciences, engineering, math/IT, and architecture, and treating healthcare as STEM-related.

Data source: 2015–2019 American Community Survey (ACS) IPUMS USA cross-sectional microdata covering the U.S. population. Inclusion criteria and samples: (1) For analyses of marital status and retention, 159,290 women aged 25–55 with at least a bachelor’s degree in STEM or STEM-related fields, regardless of marital status. (2) Main analytic sample for spousal effects: 147,467 married women aged 25–55 with at least one STEM or STEM-related major, whose spouse is employed and college-educated (bachelor’s or higher). (3) Subsamples by field: STEM subsample (92,551 women) and healthcare subsample (56,829 women); some women appear in both due to multiple majors. Measures: STEM and STEM-related degree fields and occupations were classified using SOC Policy Committee recommendations and the CTE Act recognizing architecture as STEM. Dependent variables: (a) Retention: binary indicator of working in a STEM or STEM-related occupation; in subsample analyses, indicator of working in the same classification as one’s degree (STEM vs healthcare). (b) Employment status: binary indicator of being employed. Key independent variables: spouse has a STEM or STEM-related occupation (binary); spouse has a STEM or STEM-related degree field (binary). Additional key variables in some models: marital status categories; presence of children and interaction terms with spouse’s STEM/STEM-related indicators. Controls: census year; woman’s age, race, U.S.-born, state, metro residence; woman’s and spouse’s education above bachelor’s; specific most-advanced degree field(s) for woman and spouse; spouse’s sex; number of children and number under age 5 (omitted when modeling child interactions); spouse’s wage and salary income (per $10,000) and usual weekly hours in selected models. Statistical analysis: Multiple groups of logistic regression models estimated odds of outcomes. Model groups: (1) Association between marital status (including whether spouse has STEM occupation/degree) and women’s retention. (2) Effects of spouse’s STEM occupation/degree on retention, with/without controls for spouse’s income and hours. (3) Effects on women’s employment (to distinguish nonemployment from non-STEM employment). (4) STEM and healthcare subsamples: effects of spouse’s same-class vs cross-class occupation/degree on retention in the field of one’s major. (5) Models including presence of children and interaction terms with spouse’s STEM occupation/degree. Software: Stata used for descriptive statistics, figures, and logistic regressions.

• Descriptives: Among married women with STEM/STEM-related majors, 56.63% were in STEM/STEM-related occupations when spouses were in STEM/STEM-related occupations vs 47.01% when spouses were not. In the sample, 40.8% of spouses worked in STEM/STEM-related. Women in couples with a STEM/STEM-related spouse worked fewer hours and had higher incomes on average; spouse income and total family income were higher. Degree field distribution showed healthcare (Medical and Health Sciences and Services) was most common (37.26%); among healthcare majors, 14.94% worked in high-status clinician roles, while 46.30% were in nursing and similar roles. - Marital status and retention (Table 6): Relative to never-married women, being married to a spouse with a STEM/STEM-related occupation increased the odds of staying in a STEM/STEM-related occupation by 1.240. Being married to a spouse without a STEM/STEM-related occupation decreased the odds (OR 0.787). Having a spouse with only a STEM/STEM-related degree (no occupation alignment) was not significantly related to retention. - Spousal occupation/degree and retention (Table 7): Spouse in STEM/STEM-related occupation positively associated with women’s retention (OR 1.591 without, 1.603 with spouse income/hours controls). Spouse with STEM/STEM-related degree also positive but smaller (OR 1.180 without, 1.195 with income/hours). Spouse wage and salary income and usual hours were each negatively associated with women’s retention (OR per $10,000 ≈ 0.994–0.996; OR per hour ≈ 0.991–0.992). - Employment status (Table 8): A spouse’s STEM/STEM-related occupation was negatively associated with women’s overall employment (OR 0.890; with income/hours controls OR 0.925–0.950), while spouse’s STEM/STEM-related degree alone was not significant. Spouse income and hours also negatively related to women’s employment (income OR 0.988; hours OR 0.973). - Field-specific subsamples (Table 9): Same-class spousal occupation strongly increased retention: STEM subsample spouse STEM occupation OR 1.921; healthcare subsample spouse healthcare occupation OR 1.990. Cross-class spouse healthcare occupation reduced STEM retention (OR 0.523). Spouse degree field had no significant effect in the STEM subsample but was strongly positive in healthcare (spouse healthcare degree OR 2.525). - Children as moderator (Table 10): Presence of children was negatively associated with women’s retention (OR ~0.885–0.929). Interaction terms indicated moderation when the spouse had a STEM/STEM-related occupation (interaction OR ~0.920–0.922), but the core positive association between spouse’s STEM occupation/degree and women’s retention remained stable or slightly strengthened after accounting for spouse income/hours and child interactions.

Findings support the hypothesis that spousal alignment in STEM or healthcare is positively associated with married women’s retention in those occupations. The positive association persists after controlling for spouse work hours, income, and child-related factors, suggesting that informational support, shared professional interests, and assortative matching (spousal homogamy) may facilitate women’s continued employment in their trained fields. While spouse income and longer work hours are negatively related to women’s employment and retention—consistent with potential work–family trade-offs—these factors do not negate the positive effect of spousal occupational similarity. Field-specific analyses show the strongest associations when spouses share the same classification (particularly pronounced in healthcare) and even negative effects when spouses are in a cross-class field (e.g., spouse in healthcare associated with lower STEM retention), highlighting the importance of field congruence. Overall, the results align with prior work on homogamy and previous evidence linking spouses’ STEM employment to women’s STEM retention.

For college-educated married women with STEM or STEM-related majors, having a spouse employed in the same broad field (STEM or healthcare) is associated with higher odds of remaining in that field, and this positive association outweighs potential negatives linked to spouse work intensity and earnings. The effect is robust across models, present in both STEM and healthcare subsamples, and remains when children are present, with marginal strengthening in some specifications. The study contributes by distinguishing STEM from healthcare, quantifying spousal effects on retention, and highlighting that field-congruent spousal employment is more predictive than spousal degree alone. Future research could unpack mechanisms such as STEM-specific informational/emotional support, examine other occupational domains, and incorporate richer measures of women’s own work hours and earnings.

• Women’s own salary and working hours were not included, introducing potential omitted-variable bias of unknown direction. - Limited child-related demographics were modeled; excluding richer child characteristics may underestimate positive spousal effects. - The study cannot identify causal mechanisms or directly measure spousal support; reasons for the positive association remain speculative. - The focus is restricted to STEM and STEM-related (healthcare) fields and is not necessarily generalizable to other occupations. - The analysis relies on cross-sectional ACS data, limiting causal inference.