My theoretical contributions are focused in two domains. One stream endeavors to reveal how perceptual factors regarding category membership affect how market participants make evaluative decisions, sometimes to dramatic effect. My second stream applies this cognitive perspective to theorize novel mechanisms of the labor market matching process. By conceptualizing social categories of job seekers, such as their race, gender, or country of origin, as cognitive constructs, I can apply decision theories to explain hiring outcomes.

Published/Forthcoming

11. Parasurama, P., Ming D. Leung, and Sharon Koppman. 2025. Applying While Black: The Collateral Effects of Racial Differences in Work Histories. Administrative Science Quarterly, Published May 20 online.  https://doi.org/10.1177/00018392251340351 

It is well known that hiring practices that treat job seekers differently by race contribute to racial disparities in employment. Yet hiring practices that treat job seekers equivalently by race may also contribute to racial disparities if there are preexisting racial differences. We focus on the prominent hiring practice of using the prior experience a job seeker lists on her résumé—that is, her work history—to make inferences about her suitability for a given job. Scholars and practitioners alike have long assumed work histories result from job seekers’ strategic choices about where to apply and what jobs to accept and are therefore race neutral. However, Black job seekers face a distinct set of structural constraints—namely, anticipating and experiencing racial discrimination—that restrict the job search strategies and resulting jobs available to them. As a result, they are less likely to construct the related and specialized work histories employers value compared to their White peers. These racial differences in work histories contribute to the racial disparities that Black job seekers experience. We test and find support for this argument using over 490,000 job applications for all 3,683 publicly posted jobs over seven years at two U.S. technology companies. This study uncovers a novel pathway through which race shapes employment, contributing to the literature on racial discrimination and categorization in labor markets.

10. Daviss, Claire and Ming D. Leung. 2025. “Gender, An Uphill Battle: Employer Learning, and the Persistence of Occupational Gender Segregation.” Organizational Science; Published March 28 online.  https://doi.org/10.1287/orsc.2023.17383   

Employers in online labor marketplaces prefer women over men due to stereotypes that women are more trustworthy. These stereotypes are especially salient in this context because of the uncertainty in online transactions. Yet, employers’ interactions with women and men workers might moderate the influence of trustworthiness stereotypes and, by extension, employers’ hiring preferences for individual women and men as well as gender categories in general. By exploiting rare access to large-scale, longitudinal hiring data from an online labor marketplace, we examine how employers’ interactions with workers shape their subsequent hiring preferences. Using linear probability models with job fixed effects, we find that employers prefer rehiring workers with whom they previously had positive interactions over unknown workers. Men workers benefit more than women workers from these positive interactions, suggesting that these interactions decrease the influence of negative stereotypes about men’s trustworthiness. We additionally find that employers prefer hiring workers of the same gender as workers with whom they previously had positive interactions, suggesting that employers’ interactions with individual women and men shape their preferences for hiring women and men in general. These findings point to a nuanced theoretical relationship between employers’ interactions with workers and their social category preferences in hiring under conditions of immense uncertainty.

9. Yang, Tiantian, Jiayi Bao, and Ming D. Leung. 2025 “Approaching or Avoiding? Gender Asymmetry in Reactions to Prior Job Search Experience.” Social Forces; Volume 103, Issue 4, Pages 1350–1373. https://doi.org/10.1093/sf/soaf011

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.

8. Reschke, Brian P.  and Ming D. Leung. 2021. "Variety is the Spice of Life: Heterogeneity in Evaluator Engagement and the Valuation of Atypicality," Research in the Sociology of Organizations, Volume 77, 163–186.

PDF

Since initial demonstrations that categories are consequential for evaluation, scholars of organizations and markets have attended to dynamics in audience evaluations of category spanning. We consider how heterogeneity in evaluator engagement in a market may alter their evaluation of atypical candidates. In markets where evaluators self-propagate theories of diversification, atypical candidates are advantaged because they present a distinct and efficient opportunity to diversify. We argue that evaluator market engagement will (positively) moderate valuations of atypicality, as such evaluators will be better positioned to recognize atypical candidates and their alignment with prevailing theories of value. We find support for our contentions with data from an online peer-to-peer lending market, Prosper.com. Consistent with our hypothesis, we find that lender evaluation of these atypical borrowers is increasing in their market engagement: whereas lenders new to the market devalue atypical candidates, those who have made many evaluations evaluate atypicality positively.

7. Hurley, Vanessa B., Hector P. Rodriguez, Stephen Kearing, Yue Wang, Ming D. Leung, and Stephen M. Shortell. 2020. "The Impact of Decision Aids on Adults Considering Hip or Knee Surgery." Health Affairs, 39 (1). DOI:10.1377/hlthaff.2019.00100 

PDF

Trials of decision aids developed for use in shared decision making find that patients engaged in that process tend to choose more conservative treatment for preference-sensitive conditions. Shared decision making is a collaborative process in which clinicians and patients discuss trade-offs and benefits of specific treatment options in light of patients’ values and preferences. Decision aids are paper, video, or web-based tools intended to help patients match personal preferences with available treatment options. We analyzed data for 2012–15 about patients within the ten High Value Healthcare Collaborative member systems who were exposed to condition-specific decision aids in the context of consultations for hip and knee osteoarthritis, with the intention that the aids be used to support shared decision making. Compared to matched patients not exposed to the decision aids, those exposed had two-and-a-half times the odds of undergoing hip replacement surgery and nearly twice the odds of undergoing knee replacement surgery within six months of the consultation. These findings suggest that health care systems adopting decision aids developed for use in shared decision making, and used in conjunction with hip and knee osteoarthritis consultations, should not expect reduced surgical utilization.

6. Barach, Moshe, Aseem Kaul, Ming D. Leung, and Sibo Lu. 2019. "Small numbers bargaining in the age of big data: Evidence from a two-sided labor matching platform." Strategy Science, 4 (4): 298-322.

PDF

In this study, we examine how firms engage with the big data capabilities of two-sided matching platforms. While such platforms can use artificial intelligence (AI) and big data techniques to help firms find better transaction partners, the effectiveness of these technologies depends on the concentration of transactions undertaken on the platform, giving rise to a potential small numbers bargaining problem, as users become increasingly dependent on the platform to find transaction partners. We argue that firms deal with this appropriation challenge through strategies of partial reliance: they make use of the platform’s AI driven recommendations to identify an initial set of generally acceptable partners, then rely on their own internal capabilities to select the best firm-specific match. We test this argument in the context of an online labor platform, using a regression discontinuity design to causally identify the effect of the platform’s recommendations on the hiring choices of firms at every stage of the recruiting process. Consistent with our theory, we find that firms rely heavily on the platform’s recommendations when screening job applicants to view, but ignore these recommendations when deciding whom to hire from among those interviewed, with the reliance on the platform’s recommendations being weaker for more specialized jobs and for more experienced employers. The study contributes to our understanding of how firms use big data technologies, highlighting the challenges these technologies pose for organizational governance and scope choice, and providing a bridge between research on AI and big data and work on two-sided markets.

5. Leung, Ming D. and Sharon Koppman. 2018. "Taking a Pass: How Proportional Prejudice and the Decision Not to Hire Reproduce Sex Segregation in an Online Labor Market." American Journal of Sociology, 124 (3): 762-813. 

PDF

We propose and test a theory of how decisions not to hire reproduce sex segregation through what we term proportional prejudice. We hypothesize that employers are less likely to hire anyone when the applicant pool contains a large proportion of gender atypical applicants – that is, applicants from a different gender than the typical job holder – because they view this as a signal of a poor quality applicant pool. Analyses, of over seven million job applications for over 700,000 jobs by over 200,000 freelancers on an online platform for contract labor support our contention. A survey experiment isolates the mechanism: Applicant pools with a larger proportion of gender atypical applicants were perceived as less likely to contain people who “seemed skilled enough for the job.” We conclude by demonstrating how our theory explains the mixed findings as to whether gender atypical job seekers are disadvantaged in the hiring process. 

4. Leung, Ming D. 2018. “Learning to hire? Hiring as a Dynamic Experiential Learning Process in an Online Market for Contract Labor.” Management Science, 64 (12): 5461-5959. 

PDF

We know a job applicant’s social category affects an employer’s likelihood of hiring them, but we do not know whether, or how, employers update their beliefs regarding these social categories. I examine how prior negative and positive hiring experiences of employees from particular countries affect an employer’s subsequent likelihood of hiring applicants from those countries. I hypothesize that employer reactions will reflect loss aversion – that employers will react more strongly to negative hiring experiences than positive ones. Furthermore, I expect that the similarity of the prior job will moderate this effect. Analyses of 3.9 million applications, from freelancers worldwide, for over 290,000 jobs on an online labor market demonstrate that employers are 15% less likely (versus 8% more likely) to hire freelancers from a country following a prior negative (versus positive) experience. Prior negative experiences with similar jobs (versus dissimilar jobs) lead employers to be 82% less likely (versus 8% less likely) to hire from that country. Conversely, positive experiences with similar jobs (versus dissimilar jobs) lead employers to be 25% more likely (versus 3% more likely) to subsequently hire from that country. The consequences for switching countries, following negative experiences, are analyzed and wage differences to compensate for employer reactions, are calculated. Contributions to the hiring discrimination, impression formation, and gig-economy literatures are discussed.

3. Leung, Ming D. 2014. “Dilettante or Renaissance Person? How the Order of Job Experiences Affects Hiring in an External Labor Market.” American Sociological Review, 79 (1): 136-158.

PDF 

Social actors who move across categories are typically disadvantaged relative to their more focused peers. Yet candidates who compile experiences across disparate areas can either be appreciated as renaissance individuals or penalized as dilettantes. Extant literature has focused on the comparison between single versus multiple category members and on skill assessment, hindering its applicability. To discriminate between more versus less successful category spanners, I suggest that the order of accumulated experiences matters, because it serves as an indicator of commitment. I propose the concept of erraticism and predict that employers will prefer candidates who demonstrate some erraticism, by moving incrementally between similar jobs, over candidates who do not move and also over those with highly erratic job histories. Furthermore, I suggest this relationship holds for more complex jobs, less experienced freelancers, and is attenuated through working together. These issues are particularly salient given the rise of external labor markets where careers are increasingly marked by moves across traditional boundaries. I test and find support for these hypotheses with data from an online crowd-sourced labor market for freelancing services, Elance.com. I discuss how virtual mediated labor markets may alter hiring processes.  

2. Leung, Ming D. with Amanda J. Sharkey. 2014. “Out of Sight, Out of Mind? Evidence of Perceptual Factors in the Multiple-category Discount.” Organization Science, 25 (1) p.171-184.

PDF

Extant work shows that market actors who span multiple social categories tend to be devalued relative to their more specialized peers. Scholars typically explain this pattern of results with one of two arguments. Some contend that perceptual factors, namely the difficulties that buyers have in making sense of category spanners, contribute to the observed pattern of devaluation. Others argue that the penalty for category-spanning stems from the fact that those who do not focus their efforts narrowly tend to offer products that are of lower quality. Because these two mechanisms often co-occur, it has been difficult to provide definitive evidence of the perceptually-driven component of the multiple-category penalty. We employ a natural experiment on a peer-to-peer lending website to address this gap. Difference-in-difference analyses on matched samples show that category spanning is perceived negatively and can result in devaluation, even in the absence of underlying quality differences. This result supports the argument that perceptual issues contribute to the penalty for category spanning.  

1. Negro, Giacomo and Ming D. Leung. 2013. ""Actual" and Perceptual Effects of Category Spanning." Organization Science, 24(3): 684-696.

PDF

Literature to date has demonstrated that producers and products spanning multiple categories have inferior market performance. However, two related but distinct explanations exist as to the source of such a discount. One explanation suggests that “actual” skills are degraded when producers attempt to engage across diverse categories. Another explanation involves perceptual fit to category representations held by an audience as the cause. These two explanations tend to be confounded in archival studies because external observers, responsible for the evaluation of market performance, are often aware of both the identity of producers and the underlying characteristics of their products. This leaves researchers unable to empirically separate effects. We present an analysis conducted in a setting in which it was possible to distinguish the two mechanisms, critics’ ratings of the same wines through “blind” and “non–blind” tastings. The findings indicate that after controlling for the value of ratings assigned blindly, the wines made by wineries spanning styles continue to receive lower ratings in the non–blind situation. 

Under Review/Revision

A. Koppman, Sharon, Ming D. Leung, and Tingting Nian. 2026. “Breaking Type: How Generalist Careers Improve Gender Equality in Organizations”  Under 3rd Round Review

Building on research showing that organizational events that disrupt the status quo can reduce gender inequality, we develop a theory of how career generalism inside organizations might offset gender inequality in advancement by helping women “break type”—that is, subvert managers’ limited impressions of their potential. To test our theory, we draw on seven years of longitudinal data of over three million employee-month observations from a Fortune 500 high-technology firm headquartered in Silicon Valley. We find that the premium for a generalist career is stronger for women than for men, though this “advantage” only reduces their baseline disadvantage. Supplemental analyses show that this benefits women in roles traditionally associated with women, who are more likely to be typecast. This study updates the literature on career advancement, highlights how the consequences of category-spanning vary by gender, and provides practical implications to reduce gender advancement gaps.

B. Yang, Tiantian, Jiayi Bao, Tianna S. Barnes, and Ming D. Leung. 2026 "Looking the Part? How Candidates’ Race and Attire Shape Employer Hiring Decisions in a Low-Wage Labor Market" Under 1st Round Revision

Racial disparities in employment persist in the United States, particularly in low-wage labor markets. While employers often attribute these disparities to differences in job seekers’ self-presentation—particularly attire—we argue that this explanation obscures demand-side biases. Drawing on labor market discrimination and status characteristics theories, we argue that professional attire acts as a proxy for unobservable worker qualities and that racial stereotypes distort how these signals are interpreted. White candidates might benefit more from professional attire, whereas Black candidates receive weaker returns from employers. To test this argument, we conduct the first large-scale empirical analysis of candidates’ attire and hiring outcomes using a mobile gig-staffing platform (Jobmate), which includes 1,032,496 job applications for 60,636 job seekers, along with their profile photos. Our findings reveal that although professional attire improves hiring outcomes for both Black and White candidates, the effect is significantly weaker for Black job seekers, exacerbating racial hiring disparities. A controlled experiment further demonstrates that professional attire disproportionately benefits White candidates. These findings challenge the prevailing belief that self-presentation strategies can mitigate racial disparities. Our findings suggest that closing racial hiring gaps in lowwage labor markets requires cultural changes in hiring practices rather than individual-level adjustments by job seekers.

C. Yoon, Simon and Ming D. Leung. 2025 “Beyond Organizational Boundaries: Gender Segregation and Implications for Career Sequences in the Era of Remote Gig Work” Under 1st Round Revision

While processes of gender segregation in organizations have been well documented, much less work has examined whether and how these manifest in remote, gig working environments, an increasingly prevalent form of employment. We examine behaviors of women and men gig-workers on elance.com a platform that enables semi-skilled and skilled gig workers to be hired and work remotely. Analyses of the job application behaviors between 2008 and 2013 of over 70,000 gig workers reveal that women are more likely to begin working in job domains that pay less than jobs with a greater proportion of men gig workers. We hypothesize and find support that women are more likely than men to apply to job domains they have never previously worked in—i.e., novel job areas. Moreover, when applying to these novel jobs, women tend to target more distant domains, perhaps seeking better pay. The cumulative effect of these behaviors is that women gig workers are more likely to compile disadvantaged, erratic careers, composed of sequences of more dissimilar jobs, on the platform, harming their future hiring outcomes. Additional analyses reveal an earnings difference due to these erratic careers. These findings reveal how gender stratification fundamentally shapes how women gig worker can be trapped in a cycle of disadvantage and contributes to the work on gender segregation, skilled remote gig-work, and allocative discrimination.

D. Leung, Ming D., Wang, Jue (Kate), and Bergmann, Patrick. 2026. "Polarization and Participation: How Review Polarization Spurs Contributions to Online Forums" Under 1st Round Revision

We develop a theory of spontaneous solidarity that explains how visible disagreement mobilizes participation in online opinion markets. Departing from accounts that attribute reviewing to stable individual motives or inherent product traits, we argue that polarization among prior ratings, especially when made salient by platform design, socially constructs controversy that invites users to “pick a side.” Visible disagreement, therefore, mobilizes participation independent of an object’s intrinsic contentiousness. We test this theory by exploiting an interface change on Metacritic in August 2010 that displayed rating distributions through color-coded visualizations. Using difference-in-differences analysis, we compare 1,315 identical films on Metacritic and Rotten Tomatoes over 96 months, isolating the causal effect of polarization visibility from confounding movie characteristics. Heightened visibility increased user engagement by 20.67% more ratings, 54.54% more text reviews, and 2.24% more extreme contributions. We demonstrate that contentiousness emerges through social construction rather than residing in objects, revealing platforms as active constructors of opinion markets. Theoretically, we identify spontaneous solidarity as a novel form of collective mobilization arising from minimal social structures, advancing understanding of how ephemeral alignments produce consequential collective behavior in digital environments.


Advanced Working Papers

A. Leung, Ming D., Yoon, Simon, and Daviss, Claire. 2026. "Shaping Stereotypes: Categorical contrast learning and the malleability of gender beliefs in two gig-work platforms." Working Paper

Previous research shows that employers update their perceptions of individual workers based on direct observations of those workers, but offers conflicting evidence whether employers generalize these observations to entire social categories of workers. We reconcile these findings by proposing a theory of categorical contrast learning: that employers are more likely to update their beliefs of a social category when they interact with workers for jobs stereotypically incongruent with the worker’s social category, as these experiences are more salient. We examine gender-based hiring of freelancers in two gig-work settings, WorkNow (a pseudonym) and Elance.com. We expect that employers who hire women (men) freelancers are more likely to hire other women (men) freelancers following a positive hiring experience with women (men). Furthermore, we predict that employers who have hired a woman for a masculine-typed job will be more likely to hire women for future jobs of that type, compared to the increased likelihood of hiring men after having hired a man for the same masculine-typed job. We predict the opposite effect for hiring men in feminine-typed job domains. Results from linear probability regressions with employer fixed effects support our theory. We report several robustness checks. Our theory has implications for theories of gendered hiring bias and gender segregation, social categorization processes, and stereotypes.

B. Yoon, Simon and Ming D. Leung. 2025. "The Founder Gender Tax: Employee Retention Challenges in Women-Led Ventures." Working Paper

This paper examines how founder gender shapes employee retention in entrepreneurial ventures—a critical yet overlooked dimension of entrepreneurial scaling. Although entrepreneurship is often portrayed as a path for women to bypass workplace inequality, women-led ventures face distinct challenges in building and sustaining their teams. Using longitudinal data on over 54,000 U.S. startups and 5.7 million employees founded between 2000 and 2023, I find that ventures led by women exhibit systematically higher rates of employee turnover than those led by men. These disparities stem from gendered perceptions of founder legitimacy that weaken employees’ attachment to female-led organizations and constrain women’s ability to retain talent. Follow-on analyses show that elevated turnover substantially lowers the likelihood of key growth outcomes, including IPO exits. By shifting attention from financial capital to the human capital dimension of scaling, this study reveals how gendered legitimacy dynamics within entrepreneurial organizations reproduce inequality and create enduring disadvantages for female founders.

C. Jang, Marco Dongil, Alan Benson, Ming D. Leung. 2025 "The Cost of Specialist Signaling: Visibility and Opportunity of Online Professional Profiles." Working Paper

Professional profiles on online platforms like LinkedIn, GitHub, and Upwork make workers’ skills and experiences publicly visible, presenting career opportunities without requiring direct job applications. However, because workers cannot tailor their profiles to specific positions, they must strategically present themselves to the market. We ask: How does signaling as a specialist or a generalist affect the visibility and opportunity of online professional profiles? Although existing studies have explored the trade-off between specialists and generalists, theorizing and isolating the perceptual effects of signaling from the “actual” effects of skills has been challenging, as these two are naturally correlated. To address this, we propose the distinct risks and rewards of signaling and leverage an unannounced platform design change that exogenously blinded the degree of specialization in the initial screening stage, making specialist signaling indistinguishable from generalist signaling. When signaling types were identifiable, specialist signaling increased profile views by 33%, but the conversion rate from profile views to interview offers was 70% lower, resulting in a 59% decrease in interview offers compared to generalist signaling. Our findings provide evidence that specialist signaling in online profiles carries a true “signaling” cost—one that may initially increase visibility but ultimately lead to fewer opportunity