CoChat
Sign Up
Cited by 8Names, Identity, and Discrimination ResearchLabour Market and MigrationLabor market dynamics and wage inequality

Getting a Foot in the Door: A Meta-Analysis of U.S. Audit Studies of Gender Bias in Hiring

Open access
Park, So Yun, Oh, Eunsil·2025·Sociological Science
DOI 10.15195/v12.a2
Open this paper in CoChat — free accountOpen PDF

Abstract

For the past three decades, scholars have conducted field experiments to examine gender-based hiring discrimination in the United States. However, these studies have produced mixed results. To further interpret these findings, we performed a meta-analysis of 37 audit studies conducted between 1990 and 2022. Using an aggregated sample of 243,202 fictitious job applications, the study finds no evidence of statistically significant gender discrimination at the study level. However, a series of more focused meta-analyses reveal important variations in the extent of discrimination by occupation type and applicant race. First, the gender composition of an occupation predicts gender bias in hiring. Second, the intersection of gender and race is critical—in female-dominated jobs, White female applicants receive more callbacks than their male counterparts, but Black female applicants experience no such benefit. The study contributes to the literature on labor market and gender (in)equality by synthesizing the findings of field experiments.

AI snapshot

Written from the paper’s full text.

TL;DR

This meta-analysis of 37 U.S. audit studies finds no statistically significant overall gender discrimination in hiring callbacks, but gender bias varies by occupation and race. Female-dominated occupations favor female applicants, yet this advantage appears for White women but not Black women.

Key findings

  • At the study level, the weighted average male/female discrimination ratio (logged risk ratio) was -0.05, indicating a slight preference for female applicants, but it was not statistically significant (95% CI: -0.10 to 0.01).
  • At the occupation level, the aggregate discrimination ratio was -0.07 (p < 0.05), again favoring female applicants; the female-to-male ratio of an occupation significantly predicted gender bias (p < 0.05), while skill level, median hourly wage, and occupational prestige did not.
  • For White applicants, the occupation-level discrimination ratio was -0.08 (p < 0.1), with White women receiving more callbacks than White men in female-dominated occupations; for Black applicants, the ratio was -0.03 and not statistically significant, and the interaction between race and gender composition was significant at p < 0.05.
  • The full data set included 66,145 White applicants and 34,328 Black applicants; the occupational-level data set comprised 85 data points from 23 studies.

Methods

  • The authors compiled 37 U.S. correspondence audit studies conducted between 1990 and 2022, covering 243,202 fictitious job applications, and constructed a separate occupation-level data set of 85 data points.
  • They used random-effect meta-regressions with the logged discrimination ratio as the dependent variable, calculated from callback rates for men relative to women, and included study-level demographic characteristics and occupation-level characteristics (skill level, female ratio, median hourly wage, prestige).
  • They ran pooled models, occupation-level models, and separate models for White and Black applicants; Hispanic and Asian applicants were excluded because of small sample sizes.

Limitations

  • The paper states that the data reflect only the initial screening stage, mostly entry-level positions, and do not follow up on callbacks, so the results may underrepresent discrimination in later hiring stages.
  • The paper notes that evidence for Black applicants is limited, with wide confidence intervals, and that Hispanic and Asian applicants were excluded due to small samples; it also calls for more research on AI-based screening.
  • The text makes evident additional constraints: most studies were conducted after 2010, limiting trend assessment; two studies lacked data-collection years and were imputed using publication year minus two; and occupation categories required crosswalking and matching, with some missing wage data filled by similar occupations.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.15195/v12.a2

Cites

51 of 51 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

5 of 24 cited sources checked · 2 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Field experiments that explore gender-based discrimination have produced mixed results.

    “Notably, field experiments that explore gender-based discrimination have produced mixed results.”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Deming et al. 2016 source

      The paper’s Main Results (page unavailable) report higher callback rates for women than men, while its callback tables show that gender differences are not significant in every job category. This supports variation in gender-related findings within this experiment, but the paper does not establish that field experiments generally have produced mixed results on gender-based discrimination. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites Pedulla 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Botelho and Chang, 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Interpretationcitation

    The finding that there is an interplay between occupational gender composition and race parallels results of ongoing studies that investigate both race and gender to explain hiring discrimination.

    “Second, our finding that there is an interplay between occupational gender composition and race parallels the results of ongoing studies that investigate both race and gender to explain hiring discrimination (Bertrand and Mullainathan 2004;Chavez, Weisshaar, and Cabello-Hutt 2022;Deming et al. 2016).”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Bertrand and Mullainathan 2004 source
    • cites Chavez, Weisshaar, and Cabello-Hutt 2022 source
    • cites Deming et al. 2016 source

      The Work section says the study randomized applicants’ race and gender, and the Results report callback differences by both characteristics. It also examines differences across occupations, but the passages do not establish an interplay between occupational gender composition and race. The citation supports the comparison to research on race and gender in hiring, not a parallel finding of that specific interplay. (LLM, conf=0.80)

      Experimental — accuracy not yet measured
  • Findingown data

    Only the ratio of females in the occupation predicts gender bias in callbacks.

    “In sum, based on field experiments conducted in the U.S. labor market, we find that only the ratio of females in the occupation predicts gender bias.”
    • Also stated in: § Discussion · plain
      “we do find a gender compositional effect, where the ratio of female employees predicts gender bias in the initial hiring stage.”
    • Also stated in: § The Intersection of Race and Gender · hedged
      “In sum, the findings indicate that occupational characteristics, specifically the gender composition of an occupation, influence gender discrimination in hiringthe more female-dominated an occupation, the more female applicants are preferred.”
    Quote in the paperSubstantial claim
  • Findingown data

    In female-dominated jobs, White female applicants receive more callbacks than their male counterparts.

    “in female-dominated jobs, White female applicants receive more callbacks than their male counterparts”
    Quote in the paperSubstantial claim
  • Findingown data

    Occupational prestige does not predict gender bias in hiring.

    “Prestige scores do not predict gender bias.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 19
  • May combine two claims. 23
  • The quote only partly states this claim. 2

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    The findings align with existing work on the persistence of gender segregation and racial inequality in the U.S. labor market.

    “Both of these findings align with existing work on the persistence of gender segregation (Zhu and Grusky 2022) and racial inequality (Quillian et al. 2017) in the U.S. labor market.”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Zhu and Grusky 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Quillian et al. 2017 source

  • May combine two claims.

    Study design and conductcitation

    Following prior literature, rates of 0 were changed to 0.5 to avoid dropping cases, reduce bias in the estimator, and prevent division by zero.

    “Following an approach used in prior literature (Schaerer et al. 2023), we changed rates of 0 to 0.5 to avoid dropping cases; this methodological choice also reduces bias in the estimator and prevents division by zero.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Schaerer et al. 2023 source

      Section 3.1.4 (Statistical analyses; page not supplied) says the authors added 0.5 to every cell of a 2×2 frequency table before calculating log odds ratios. It explicitly gives reducing estimator bias and avoiding division by zero as reasons. It does not say they changed zero rates to 0.5 or that the adjustment was made to avoid dropping cases. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    For Black applicants, the occupational-level discrimination ratio is -0.03 and is not statistically significant.

    “ratio calculated from the occupational-level data for Black applicants (-0.03) is not statistically significant.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The association between percent female in occupation and gender bias is statistically significant for White applicants but not for Black applicants.

    “As shown in Figure 6, a statistically significant outcome was observed for White applicants in contrast to the non-significant results for Black applicants.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There is a non-statistically significant trend such that female candidates are more likely to receive callbacks for lower-prestige jobs, and as job prestige increases, the preference for female applicants declines.

    “This result means that female candidates are more likely to receive callbacks for lower-prestige jobs, and as job prestige increases, the preference for female applicant declines.However, despite this overall trend, the results of the meta-regression are not statistically significant.”
    Quote in the paperNot a substantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    At the study level, the male/female discrimination ratio was -0.03 (95% CI -0.10 to 0.03) for White applicants and 0.01 (95% CI -0.10 to 0.12) for Black applicants, and neither was statistically significant.

    “The White applicant group sample yielded a male/female discrimination ratio of -0.03 (95 percent confidence interval: -0.10 and 0.03), whereas the Black applicant group sample was 0.01 (95 percent confidence interval: -0.10 and 0.12).”
    • Also stated in: § Additional information · plain
      “The results for both White and Black applicant groups were not statistically significant.”
    Quote in the paperNot a substantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    For White applicants, the occupational-level discrimination ratio is -0.08 and statistically significant at p < 0.1, indicating a preference for female job applicants.

    “the results from the occupational-level data for White applicants yield a statistically significant (p < 0.1) discrimination ratio of -0.08.”
    • Also stated in: § Additional information · plain
      “Thus, a weighted average of the occupation-specific-effect sizes among White job candidates across the different studies signals a preference for female job applicants.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    At the individual study level, some studies identified gender discrimination, some did not, and others observed gender bias only in conjunction with variables such as race/ethnicity, credentials, or parenthood status.

    “At the individual study level, some studies identified gender discrimination, some did not, and others observed gender bias only in conjunction with variables such as race/ethnicity, credentials, or parenthood status.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The result that only the ratio of females predicts gender bias holds in meta-regressions both when the ratio of females in occupations is controlled for and when it is not controlled for.

    “Such a result holds for both meta-regressions when the ratios of females in occupations are controlled for and are not controlled for.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The meta-analysis found two significant patterns of gender bias: an occupational characteristic significantly predicted gender bias in job callback rates and findings varied by race.

    “Importantly, however, the meta-analysis found two significant patterns of gender bias: an occupational characteristic significantly predicted gender bias in job callback rates and findings varied by race.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In more female-dominated occupations, White women received more callbacks than their male counterparts, whereas Black women did not experience a similar gender preference.

    “we find that in more female-dominant occupations, White women received more callbacks than their male counterparts, whereas Black women did not experience a similar gender preference.”
    • Also stated in: § The Intersection of Race and Gender · plain
      “White female applicants in more female-dominated occupations benefit from a preference in hiring, whereas Black female applicants in female-dominated occupations do not experience a similar benefit.”
    • Also stated in: § Discussion · plain
      “in female-dominated occupations, White women applicants received more callbacks than White men, but Black women did not experience a parallel benefit.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    The remaining 24 percent of studies did not manipulate gender but controlled for gender by sending resumes from only female or only male applicants to each job opening.

    “The remaining 24 percent of the studies did not manipulate gender but did control for gender by sending resumes from either only female applicants or only male applicants to each job opening.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The trend across occupations for Black applicants is less clear-cut: some more female-dominated occupations have positive discrimination ratios, whereas some more male-dominated occupations have negative discrimination ratios.

    “Furthermore, the trend across occupations is less clear-cut.Some more female-dominated occupations have positive discrimination ratios (male candidates received more callback offers than female candidates), whereas some more male-dominated occupations have negative discrimination ratios (women received more callbacks than men).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Occupations at the top of the forest plot (e.g., nursing and administrative positions) tend to be female-dominated, and their log discrimination ratios are mostly negative, indicating that resumes from women received more callbacks than resumes from men.

    “Occupations at the top of the forest plot (e.g., nursing and administrative positions) tend to be female-dominated.The log discrimination ratios for these occupations are mostly negative, indicating that resumes from women received more callbacks than resumes from men.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Occupations at the bottom of the plot (e.g., maintenance work and manufacturing) tend to be more male-dominated, and their log discrimination ratios are mostly positive, indicating that resumes from men received more callbacks than resumes from women.

    “Occupations at the bottom of the plot (e.g., maintenance work and manufacturing) tend to be more male-dominated.The log discrimination ratios for these occupations are mostly positive, indicating that resumes from men received more callbacks than resumes from women.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The education levels used in each study and occupation group of applications were comparable across White and Black applicants.

    “The education levels used in each study and occupation group of applications were comparable, with an average of 14.82 years of education for White applicants and 14.75 years of education for Black applicants.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The intersectional perspective emphasizes that status beliefs are not based solely on the applicant's gender but represent compounded beliefs about multiple groups.

    “A second conceptual framework concerning status beliefs is the intersectional perspective, which emphasizes that status beliefs are not based solely on the applicant's gender but rather represent compounded beliefs about multiple groups (Choo and Ferree 2010;Crenshaw 1989).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Choo and Ferree 2010 source
    • cites Crenshaw 1989

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Interpretationown data

    The findings support theories of tokenism and statistical discrimination.

    “Our findings further support these theories and are aligned with empirical studies that argue that persisting sex segregation in the workforce shapes and is shaped by genderbased stereotypes (Breda et al. 2020;Charles and Bradley 2002).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    The authors independently cross-coded the occupations and went through multiple rounds of calibration.

    “The authors independently cross-coded the occupations and went through multiple rounds of calibration.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Hiring discrimination against women in male-dominant jobs is the result of employers' rational choices, given that they seek to overcome uncertainty around applicant performance.

    “Statistical discrimination theory argues that hiring discrimination against women in male-dominant jobs is the result of employers' rational choices, given that they seek to overcome uncertainty around applicant performance (Arrow 1973).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Arrow 1973

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The two researchers cross-coded the data by reading each article to extract and code the variables necessary for the analysis.

    “After compiling the study list, the two researchers cross-coded the data by reading each article to extract and code the variables necessary for the analysis.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The results of existing field experiments on gender bias generalize predominantly to White female applicants, and Black female applicants follow somewhat different patterns.

    “These nuanced dynamics highlight that the results of existing field experiments on gender bias generalize predominantly to White female applicants and Black female applicants follow somewhat different patterns.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisereasoning

    The proportion of women in an organization or institution (i.e., occupation) is an important factor associated with employers' hiring bias.

    “A third key perspective is a composition-based status perspective, which posits that the proportion of women in an organization or institution (i.e., occupation) is an important factor associated with employers' hiring bias.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Gender discrimination occurs across the horizontal sex segregation of occupations in both female- and male-dominant occupations.

    “Such results suggest gender discrimination occurs across the horizontal sex segregation of occupations in both female-and male-dominant occupations.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    When women hold a proportional minority of positions within an organization, it induces gender inequality.

    “Organizational scholars also emphasize the composition of the workforce as a factor driving gender discrimination, arguing that when women hold a proportional minority of positions within an organization, it induces gender inequality (Kanter 1977).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kanter 1977 source

  • Classified as probably not a substantial claim.

    Premisereasoning

    The experimental research design allows researchers to explore decisions made by real employers, thus offering a high internal validity level.

    “The experimental research design allows researchers to explore decisions made by real employers, thus offering a high internal validity level.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The results suggest a mixed pattern of gender-biased hiring decisions, which may stem from the diversity in institutional contexts and occupational characteristics across studies.

    “In sum, the results suggest a mixed pattern of gender-biased hiring decisions, which we posit may stem from the diversity in institutional contexts and occupational characteristics across studies.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    98.8 percent of Fortune 500 companies use applicant tracking systems.

    “Research conducted by Jobscan (Myers 2023), a recruitment solutions company, showed that 98.8 percent of Fortune 500 companies use applicant tracking systems-software that makes hiring decisions by using algorithms to automatically filter submitted resumes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Myers 2023

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisereasoning

    Demand-side explanations of hiring discrimination are particularly useful in explaining the results of prior field experiments.

    “Thus, demand-side explanations of hiring discrimination, which focus on employers' decisions to hire, promote, and evaluate employees, are particularly useful in explaining the results of prior field experiments”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Audit studies use fictitious job applications with randomly assigned characteristics to audit employment outcomes and assess discrimination in hiring based on manipulated characteristics.

    “In audit studies, researchers use fictitious job applications with randomly assigned characteristics (e.g., gender, race/ethnicity, and criminal record) to audit employment outcomes and thereby assess discrimination in hiring based on manipulated characteristics (Gaddis 2018).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Gaddis 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Statistical discrimination theory posits that employers make hiring decisions based on their prior knowledge about average productivity levels across different groups.

    “Economists have introduced the idea of statistical discrimination, which posits that employers make hiring decisions based on their prior knowledge about average productivity levels across different groups (e.g., women and other identifiable classes of job applicants) (Bielby and Baron 1986;McPherson, Smith-Lovin, and Cook 2001).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Bielby and Baron 1986 source
    • cites McPherson, Smith-Lovin, and Cook 2001 source

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Only when the multidimensionality of gender and racial stereotypes is considered can we adequately assess the changes and continuities of hiring inequality at the initial screening stage.

    “Only when the multidimensionality of gender and racial stereotypes is considered can we adequately assess the changes and continuities of hiring inequality at the initial screening stage.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The random-effect model was used to account for potential variability across studies due to different geographic areas and occupation categories and variation in applicant characteristics.

    “we used a randomeffect model to account for potential variability across studies that may result from two factors: individual studies targeting different geographic areas and occupation categories and variation across studies in the characteristics of job applicants.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The random-effect model assumes that there is not a single true effect size but rather a distribution of true effect sizes.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Crenshaw (1989:139) first coined the term intersectionality based on her evidence from legal cases and her argument that Black women are legally, policywise, and theoretically erased when scholars and policymakers depend on the single-axis analysis that distorts the experiences of Black women.

    “Crenshaw (1989: 139) first coined the term intersectionality based on her evidence from legal cases and her argument that Black women are legally, policywise, and theoretically erased when scholars and policymakers depend on the”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Crenshaw (1989: 139)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    The occupational skill-level variable is an index of occupational education based on the 1989 General Social Survey and 1990 U.S. Census, used following Yu and Kuo (2017) as a measure of overall skill levels of workers within occupations.

    “The occupational skill-level variable is an index of occupational education, which is a measure of the percentage of occupational incumbents' education level (one or more years of post-secondary education completed) created using the 1989 General Social Survey and the 1990 U.S. Census.Following Yu and Kuo (2017), we use this as a measure of the overall skill levels of workers within occupations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Yu and Kuo (2017) source

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Study design and conductnone given

    Elements of the typical design of field experiments (focusing on entry-level positions or mixing fictitious resumes with real applicants) reduce the risk that an employer will become aware of the experiment.

    “Although elements of the typical design of field experiments (focusing on entry-level positions or mixing fictitious resumes with real applicants) reduce the risk that an employer will become aware of the experiment”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Most audit studies involve sending fictitious resumes for consideration for entry-level positions but do not follow up on the callbacks received.

    “Most audit studies involve sending fictitious resumes for consideration for entry-level positions but do not follow up on the callbacks received (Gaddis 2018).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Gaddis 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationcitation

    Policies that aim to solve issues of gender- and race-based discrimination separately often result in reinforcing existing marginalizations (i.e., Black women) who fall into a crack neither noticed nor protected by the policies.

    “In particular, policies that aim to solve issues of gender-and race-based discrimination (separately) often result in reinforcing existing marginalizations (i.e., Black women) who will fall into the crack that is neither noticed nor protected by the policies.”
    Quote in the paperNot a substantial claim
38 of 38

  • The authors state this as an assumption.

    The random-effect model assumes that there is not a single true effect size but rather a distribution of true effect sizes.

    “The random-effect model assumes that there is not a single true effect size but rather a distribution of true effect sizes.”

  • The paper relies on this without support.

    Gender and race are context-specific social constructs.

    “considering the fact that gender and race are context-specific social constructs.”

  • The paper relies on this without support.

    Fictitious names reliably signal race to employers.

    “racial background was indicated by fictitious names”

  • The paper relies on this without support.

    Callback rates from audit studies are a valid indicator of hiring discrimination at the initial screening stage.

    “Most audit studies involve sending fictitious resumes for consideration for entry-level positions but do not follow up on the callbacks received (Gaddis 2018).”
4 of 4

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Gender-manipulation descriptions conflict

    Minor

    The text divides all studies into 35 percent focused primarily on gender and 65 percent including gender as part of the manipulation. That implies every study manipulated gender, contrary to the note that 24 percent did not.

  • Data-collection date range excludes a year shown in the results

    Minor

    The study-level description says the included experiments were conducted from 2002 through 2021, but the results list 2001 as a year of data collection.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 4

  1. Arrow, Kenneth J. 1973. "The Theory of Discrimination." Pp. 3-33 in Discrimination in Labor Markets. Princeton University Press.
  2. Cotter, David, Joan Hermsen, and Paula England. 2008. "Moms and Jobs: Trends in Mothers' Employment and Which Mothers Stay Home." Pp. 379-86 in, American families: A multicultural reader. New York: Routledge.
  3. Crenshaw, Kimberlé. 1989. "Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Anti-Discrimination Doctrine, Feminist Theory and Anti-Racist Politics." Pp. 139-167 in University of Chicago Legal Forum.
  4. Damaske, Sarah. 2011b. For the Family? How Class and Gender Shape Women's Work. New York: Oxford University Press.

Cited by

Sorted by citations

Related

Authors
Park, S. Y., & Oh, E.
Affiliations
University of Wisconsin–Madison (US)
ORCID
Oh, Eunsil
Venue
Sociological Science
Publisher
Society for Sociological Science
Published
2025
Volume
12
Pages
26-50
ISSN
2330-6696 (ISSN-L)
DOI
10.15195/v12.a2
OpenAlex
W4406006721
Cited by
7
References
42
Citations by year
2026: 7
FWCI
1.81
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Sociological Science (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • University of Wisconsin-Madison
  • Division of Graduate Education
Topics
  • Social Sciences › Sociology and Political Science › Names, Identity, and Discrimination Research
  • Social Sciences › Gender Studies › Labour Market and Migration
  • Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Labor market dynamics and wage inequality
Keywords
gender discrimination in hiring, audit studies, callbacks, job applications, gender composition of occupations, intersectionality, race and gender, female-dominated occupations, Black women, labor market inequality, field experiments

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-11-04.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-11-04.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review