Do Single‐Sex STEM Programs Have Merit? If So, for Whom, on What Measures?

ABSTRACT Single‐sex STEM programs—defined here as voluntary, gender‐segregated extracurricular or supplemental activities (e.g., summer camps, workshops, robotics clubs, internships, or citizen science initiatives)—have experienced heightened popularity and scrutiny amid efforts to increase diversity in STEM fields. Empirical evidence regarding their efficacy has been inconsistent. Using a nationally representative dataset of 15,725 US undergraduates, we employed rigorous Propensity Score Weighting (PSW) to estimate the effects of attending single‐sex STEM programs during high school on STEM affinity outcomes (i.e., STEM interest, identity, and career intentions). Our findings demonstrated that attending single‐sex STEM programs was associated with a closing of gender gaps. After PSW adjustment, the interaction between being female and attending a single‐sex program was strongly positive for STEM interest ( β = 0.465, p < 0.001), STEM identity ( β = 0.496, p < 0.001), and, most notably, STEM career intentions ( β = 1.778, p < 0.001; odds ratio = 5.92). Sensitivity analysis showed these effects were highly robust to unmeasured confounding ( E ‐values up to 6.03). Conversely, boys attending all‐boys programs exhibited significantly lower STEM affinity in all three outcomes, compared with boys in coed programs or girls in single‐sex programs. Overall, single‐sex STEM programs provided substantial gains for girls while having mixed‐to‐negative implications for boys. We suggest expanding access to all‐girls STEM programs and critically investigating the pedagogical features driving these differential effects.