Forthcoming

Evidence from a correspondence experiment

Abstract:   This study examines the intersection of sex and nonbinary gender identity in hiring discrimination using a resume audit study, where sex was signaled via first name and nonbinary identity via “they/them” pronoun disclosure. Results show male and female nonbinary applicants face similar discrimination levels, with patterns in sex-based discrimination resembling those of cisgender applicants with the same implied sex: female-named applicants are penalized in male-dominated occupations and vice versa. Applicants with intersecting minoritized identities—being both the non-dominant sex and disclosing nonbinary pronouns—face heightened discrimination. These findings highlight the importance of considering intersecting identities in understanding labor market disparities.

Current Version: available here
Twitter Thread: available here

To be published in AEA Papers & Proceedings (May 2025)

 A field experiment on pronoun disclosure and hiring discrimination

Abstract: Nonbinary people have a gender identity that falls outside the male-female binary. To investigate hiring discrimination against this group, thousands of randomly generated fictitious resumes were submitted to job postings in pairs where the treatment resume contained pronouns listed below the name and the control resume did not. Two treatments were considered: nonbinary "they/them” and binary "he/him” or "she/her” pronouns congruent with implied sex. Hence, discrimination is estimated against nonbinary and presumed cisgender applicants who disclose pronouns. Results show that disclosing "they/them" pronouns reduces positive employer response by 17%. There is also evidence that discrimination is larger (approximately double) in Republican than Democratic geographies. By comparison, results are inconclusive regarding discrimination against presumed cisgender applicants who disclose pronouns; if discrimination does exist, it is of lower magnitude than discrimination against nonbinary applicants who disclose pronouns. 

Current Version: available here

Twitter Thread: available here

Transgender "passing privilege" and labor market discrimination

Abstract:   Transgender people experience worse labor market outcomes compared to similar cisgender peers; recently, research has found causal evidence of discrimination against this group in various settings. I propose a two-stage correspondence field experiment involving fictitious workers with A.I.-generated headshots, where the extent to which individuals "pass" as cisgender is experimentally manipulated. In addition, while some transgender workers will indirectly disclose their identity, others will not. These workers will apply to real job postings in Germany, where applicants typically include headshots on their resumes. With this experiment, I aim to answer two primary research questions. First, do transgender women experience hiring discrimination in terms of employer attention (i.e., the extent to which employers acquire applicant information) and response (i.e., interview requests or similar)? Second, do transgender women who pass as cisgender experience "passing privilege" (i.e., less discrimination) compared to those who do not pass? 

How Khan Academy influences student math learning; with Philip Oreopoulos

Abstract: Recent experimental research has highlighted the promise of Computer-Assisted Learning (CAL): this technology has been shown to increase student learning and offers affordable scalability. However, this evidence is often focused on unrealistic dosage levels, implemented among teacher volunteers (who have ostensibly bought into CAL), or implemented among students at home as an add-on to classroom learning. We add to this literature using three years of novel administrative panel data from Khan Academy. Using variation in within-teacher use of Khan Academy over time, and controlling for teacher and student fixed effects, we estimate the impact of additional Khan Academy usage in a real-world setting. In particular, we are able to identify the impact of even small changes in platform usage and in a context where teachers both increase and decrease usage over time.