Getting Inclusion
Do Equal Employment Opportunity Statements Encourage Racial Minorities? Evidence from a Large Natural Field Experiment
Andreas Leibbrandt & John List
European Economic Review, May 2025
Abstract:
Labor force composition and the allocation of talent remain of vital import to organizations. For their part, governments and companies around the globe have implemented equal employment opportunity (EEO) regulations to influence labor market flows. Even though such regulations are pervasive, surprisingly little is known about their impacts. We use a natural field experiment conducted across 10 U.S. cities to investigate if EEO statements affect the first step in the employment process, application rates. Making use of data from over 2,000 job seekers, we find that the presence of an EEO statement in job advertisements does not encourage racial minorities’ willingness to apply for jobs. Our results highlight that if one goal of EEO regulations is to enhance the pool of minority applicants, then it is not working as we also observe discouragement effects in some cities.
Approaching or avoiding? Gender asymmetry in reactions to prior job search outcomes by gig workers in female- versus male-typed job domains
Tiantian Yang, Jiayi Bao & Ming Leung
Social Forces, forthcoming
Abstract:
Despite recent increases in females entering male-typed job domains, women are more likely to exit these jobs than men, leading to a “leaky-pipeline” phenomenon and contributing to continued occupational gender segregation. Extant work has demonstrated that women are less likely to reapply to employers who previously rejected them for jobs in male-typed job domains. However, these studies leave unexamined whether women will reapply to other employers in those job domains and, if so, whether this pattern differs in female-typed job domains, hampering our confidence in the contribution of these patterns to gender segregation. This paper investigates whether employer rejection dampens women’s job-seeking persistence more than men’s for all employers and across male versus female job domains. Regression analyses of more than 700,000 applications for over 200,000 job postings by roughly 70,000 freelancers in an online contract labor market demonstrate that women are more likely than men to reduce job-seeking activity from all employers following rejections in the male-typed IT and programming job domain. Women are also more likely than men to seek jobs in other domains outside IT and programming following job-seeking rejection. By contrast, female freelancers in female-typed writing and translation jobs do not exhibit similar gendered behavior patterns. Implications for research on gender segregation, careers, and hiring are discussed.
Sex-Based Wage Gaps in Nursing
Pablo Estrada, Sara Markowitz & Alexia Witthaus
NBER Working Paper, February 2025
Abstract:
Males comprise a small fraction of the nurse labor force, yet across the distribution of wages, male nurses earn more than females. In this paper, we use nurse survey data to decompose the sex-based wage gap and to explore why male nurses earn a premium in a female-dominated profession. We consider the role of traditional factors such as human capital and family structure, along with explanations that are more specific to nursing. Results indicate that overtime pay is a significant factor, particularly among hospital workers, but otherwise, after accounting for an extensive set of job-related characteristics, the wage gap persists.
Scale dichotomization reduces customer racial discrimination and income inequality
Tristan Botelho et al.
Nature, forthcoming
Abstract:
Online platforms are rife with racial discrimination, but current interventions focus on employers rather than customers. We propose a customer-facing solution: changing to a two-point rating scale (dichotomization). Compared with the ubiquitous five-star scale, we argue that dichotomization reduces modern racial discrimination by focusing evaluators on the distinction between ‘good’ and ‘bad’ performance, thereby reducing how personal beliefs shape customer assessments. Study 1 is a quasi-natural experiment on a home-services labour platform (n = 69,971) in which the company exogenously changed from a five-star scale to a dichotomous scale (thumbs up or thumbs down). Dichotomization eliminated customers’ racial discrimination whereby non-white workers received lower ratings and earned 91 cents for each US dollar paid to white workers for the same work. A pre-registered experiment (study 2, n = 652) found that the equalizing effect of dichotomization is most prevalent among evaluators holding modern racist beliefs. Further experiments (study 3, n = 1,435; study 4, n = 528) provide evidence of the proposed mechanism, and eight supplementary studies support measurement and design choices. Our research offers a promising intervention for reducing customers’ subtle racial discrimination in a large section of the economy and contributes to the interdisciplinary literature on evaluation processes and racial inequality.
Race Discrimination in Internet Advertising: Evidence From a Field Experiment
Neil Sehgal & Dan Svirsky
University of Pennsylvania Working Paper, December 2024
Abstract:
We present the results of an experiment documenting racial bias on Meta's Advertising Platform in Brazil and the United States. We find that darker skin complexions are penalized, leading to real economic consequences. For every $1,000 an advertiser spends on ads with models with light-skin complexions, that advertiser would have to spend $1,159 to achieve the same level of engagement using photos of darker skin complexion models. Meta's budget optimization tool reinforces these viewer biases. When pictures of models with light and dark complexions are allocated a shared budget, Meta funnels roughly 64% of the budget towards photos featuring lighter skin complexions.
Gender Differences in Choices Involving Own Ability: Evidence from Competitive Powerlifting
Scott Abrahams
LSU Working Paper, December 2024
Abstract:
An analysis of the results from thousands of powerlifting meets establishes several behavioral patterns reflecting gender differences in choices that are functions of how one assesses one's own ability and updates that assessment in response to feedback: 1) men more often fail to lift a weight of their own choosing given a strong incentive to pick one within their ability, 2) men more often try for a heavier weight immediately after failing to lift a lighter one, and 3) men more often fail later attempts conditional on earlier misses. While these behaviors decrease with experience in aggregate, the gender gaps in them do not. These gaps persist even when lifters are competing primarily against themselves as the only entrants in their division. On the extensive margin, women are no less likely to compete again in the future after the ultimate failure of being disqualified.
Vanguard: Black Veterans and Civil Rights after World War I
Desmond Ang & Sahil Chinoy
NBER Working Paper, February 2025
Abstract:
Nearly 400,000 Black men were drafted into the National Army during World War I, where they toiled in segregated units and received little formal training. Leveraging novel variation from the WWI draft lottery and millions of digitized military and NAACP records, we document the pioneering role these men would play in the early civil rights movement. Relative to observably similar individuals from the same draft board, Black men randomly inducted into the Army were significantly more likely to join the nascent NAACP and to become prominent community leaders in the New Negro era. We find little evidence that these effects are explained by migration or improved socioeconomic status. Rather, corroborating historical accounts about the catalyzing influence of institutional racism in the military, we show that increased civic activism was driven by soldiers who experienced the most discriminatory treatment while serving their country.
Patent Outcomes and the Gender Composition of Teams
Talia Bar & Heshan Zhang
University of Connecticut Working Paper, February 2025
Abstract:
Using a United States Patent Office dataset with (name-based) inventors’ gender attribution, we examine gender differences in U.S. patent outcomes -- forward citations, triadic grants (related patents in Europe and Japan), and patent renewals. On average, patent teams with at least one woman inventor have weaker outcomes. We find that differences in assignees (the organization that initially owns the patent) explain a significant part of the gender gap. After accounting for technology, application years, patent examiners and assignees, we show that for solo-inventor patents there are no significant gender differences in any of the outcomes. However, for patent teams whose first inventor is a man, mixed gender teams have on average weaker outcomes than men-only teams. There is a small gap in citations and triadic grants even when we control for the identity of the first inventor. We discuss potential mechanisms and implications of these findings.