Spillovers and Scalability in Job Ad Experiments: Evidence from Gender-Neutral Language

Gendered-grammar languages are spoken by 39% of the global population, and recent years have seen increasing advocacy for adopting gender-neutral language to promote diversity. We present evidence from two experiments on the effects of gender-neutral language in job advertisements and its treatment spillovers. In a field experiment encompassing all job postings on a Spanish-language tech platform, ads randomly assigned gender-neutral language attracted more female applicants—but only when few other ads applicants viewed were also treated. A second experiment shows that gender-neutral language shapes female tech workers’ beliefs about job characteristics, particularly when the contrast with gendered language is salient. These findings are consistent with applicants interpreting gender-neutral language as a signal about job attributes, with effects that diminish as treatment becomes widespread. Short-run scalability is thus limited: small-scale interventions may produce meaningful impacts, but large-scale adoption may have negligible effects. However, longer-term effects may exist that our designs cannot capture.