Here, we propose that researchers utilize existing databases to research individuals who have been discriminated against. This is in contrast to the use of interviews (Cuadraz and Uttal, 1999), surveys (D'Ancona, 2017), correspondence tests (Zschirnt, 2019), or experimental designs (Baert, 2018). Not that these methodological approaches fail to contribute, rather, it is more a matter of concern that they do not take a sufficiently in-depth look at examining discrimination.

This does not intend to second-guess the methodologies used by other researchers in previously peer-reviewed and published articles, but rather to note that there are certain instances where researchers could have used specific databases but did not. The contributions to discrimination research that previous researchers have made could be enhanced by using available databases, approaching the data collection and analysis through the lens of empirical legal scholarship, and using state-of-the-art analytical tools. However, the previous research is limited by these factors.

Our proposed recommendation is three-pronged: (1). Identifying societal impetus that fosters the need not only for employment discrimination research, but for research on individuals who have been discriminated against. (2). Identifying illustrative publications that have used empirical legal scholarship and these databases to research employment discrimination and/or employment-related concepts, such as whistle-blowing (Lee and Fargher, 2013), bullying (Martin and LaVan, 2010; Richardson et al., 2016a,b), or justice (Best et al., 2011), and (3). Making recommendations regarding testable premises using the recommended databases.

How researchers should conduct employment discrimination research is undoubtedly impacted by societal shifts, and legal rulings are sure to impact it. For example, concerning societal influences, the most monumental of these changes include the recent “Black Lives Matter” movement and the “Me Too” movement. It can be reasonably expected that both of these movements will result in discriminated individuals having more confidence in pursuing their rights. It can also be reasonably expected that more individuals who were retaliated against for pursuing their rights will pursue retaliation claims. Combined, these will probably lead to organizational issues of disparate impact and disparate treatment.

The contemporary political landscape further underscores the urgency of this research approach. The current administration has issued several Executive Orders (EOs) eliminating Diversity, Equity, and Inclusion (DEI) programs across federal agencies and encouraging similar changes in private sector organizations. There are three notable Executive Orders: EO 14173—Ending Illegal Discrimination and Restoring Merit-Based Opportunity (Executive Order 14173, 2025); EO 14151—Ending Radical and Wasteful Government DEI Programs and Preferencing (Executive Order 14151, 2025); and EO 14281—Restoring Equality and Meritocracy (Executive Order 14281, 2025). Despite these restrictions and changes, EEOC enforcement continues in other areas. For instance, Columbia University agreed to a $21 million settlement in January 2025 for alleged harassment against Jewish employees following October 7, 2023, the largest EEOC settlement in nearly 20 years. As a result, individual researchers and organizations may become reluctant to conduct and participate in survey-based research or qualitative research involving interviews or focus groups. Hence, archival legal databases and EEOC case records provide researchers access to discrimination experiences difficult to capture through conventional methods.

In terms of legal aspects, the Supreme Court recently ruled in RG and GR Harris Funeral Homes, Inc. V. Equal Employment Opportunity Commission (2019) that LGBTQ individuals are protected under Title VII of the Civil Rights Act of 1964. Additionally, the EEOC has recently started to recognize the validity of intersectional claims, meaning that individuals who have been discriminated against can file claims on multiple bases [Equal Employment Opportunity Commission (EEOC), 2005]. How researchers will conduct employment discrimination research in future studies will undoubtedly continue to change due to these societal and legal examples.

Empirical legal scholarship refers to “a specific type of empirical research: a model-based approach coupled with a quantitative method” (George, 2006, p. 141). There is a subset of empirical legal scholarship that uses statistical techniques and analyses. This involves the use of studies that “employ data (including systematically coded judicial opinions) that facilitate descriptions of or inferences to a larger sample or population as well as replication by other scholars” (Heise, 2002, p. 821). Therefore, empirical legal scholarship can range from as simple as counting and surveys to more complex empirical analysis, including significance testing, multiple regression, and logit/probit analysis (Cahoy, 2010). The goals of empirical legal research are verifiability, falsifiability, and reproducibility (Bétaille, 2025).

Some authors are starting to recognize the importance and contributions of empirical legal scholarship, albeit in legal journals, and have pointed to the value of automated content analysis to enhance empirical legal scholarship (Zeiler, 2016; Allen and Blackham, 2018; Irvine et al., 2018; Blackham, 2019; Ovádek et al., 2024). We propose that the study of employment discrimination research can be enhanced by empirical legal scholarship and content analysis methodology, cognizant of both the contributions and pitfalls (Allen and Blackham, 2018).

Researchers can use numerous existing databases to research discrimination against individuals who have reported workplace discrimination. As previously indicated, those who take action and file formal complaints of discrimination are rare (Leonard, 1984), and rulings in litigated cases can confirm whether an individual has experienced discrimination beyond claiming it. Researchers with varying levels of data analytics skills can use the databases proposed here, enabling them to research employment discrimination incidents more appropriately. Many of these data sources are readily available for little or no cost. While we describe these databases in more detail below, Table 1 provides some preliminary information about these potential data sources for employment discrimination research. Table 1 provides information describing the data sources, a description of the data contained within the data source, whether the data source is freely accessible, and whether the data source requires a subscription or other related costs. Table 1 is illustrative.

The databases that we propose utilizing include Federally Litigated Cases, Arbitration Cases, Collective Bargaining Agreements, and Employer Codes of Conduct.

Below, we describe examples of research involving employment discrimination that can be examined utilizing the proposed data sources. Table 2 briefly lists these data sources, associated societal impetuses, potential employment discrimination topics/concepts/terms to explore, suggested testable premises, and illustrative examples of content that may be researched from the databases. These suggestions are meant to be examples and are not intended to be comprehensive.

Federal cases are particularly useful to test a variety of hypotheses related to employment discrimination and retaliation. A few studies in the extant literature support this approach. For example, (Best et al. 2011) empirically tested litigated employment discrimination cases using a sampling frame of over 50,000 cases from 1965 to 1999, ultimately selecting a 2% random sample from both District Courts and Appellate courts. They concluded that individuals who file claims based on multiple bases do not fare as well as those filing claims based on a single protected demographic. Future research may examine how this result is likely to change given the increased societal impetus associated with intersectionality.

(Richardson et al. 2016b) researched cases in the federal court system to identify litigation relating to bullying. They identified ninety-three cases from nine U.S. Courts of Appeal and eighty-four U.S. District Courts. Even though there is no federal law to help prevent it, protection can be provided in the court system, if individuals can prove claims such as discrimination, hostile work environment, retaliation, harassment, violations of the Americans with Disabilities Act, infliction of emotional distress, U.S. Constitution claims, disparate treatment, impact and/or discipline, wrongful discharge or termination, or violations under the Family Medical Leave Act.

Researchers can use EEOC cases to test hypotheses related to various types of employment discrimination. Bases for discrimination may include age, sex, race, color, religion, national origin, and disability. Recent studies that have utilized EEOC cases include (Baumle et al., 2020), who conducted a study examining 9,121 sexual orientation or gender identity discrimination charges filed with the EEOC or state or local fair employment practices agencies. They discerned trends in charge filings over time and industry patterns, issues raised, and the charges' outcomes. They specifically analyzed the differences in claim outcomes when retaliation is an issue.

(McMahon et al. 2017) conducted a study utilizing EEOC cases that looked at individuals with learning disabilities (N = 9,480) when compared to other general disabilities, such as physical, behavioral, or sensory disabilities (N = 313,480). They concluded that discrimination issues tend to involve current employees and involve harassment, intimidation, constructive discharge, and discipline. On the positive side, failure to provide reasonable accommodation and unlawful discharge are less commonplace for employees with learning disabilities.

Researchers can use arbitration cases to examine hypotheses related to employment discrimination and procedural justice, as well as accommodations for illnesses and/or disabilities. Past studies that have utilized arbitration cases include (Colvin and Gough, 2015), who conducted a study using the American Arbitration Association (AAA) database to discern (N = 10,335) case characteristics that would lead to a settlement before the arbitration hearing. In a recent study, (Gough 2021) compared employment discrimination outcomes between litigation and arbitration cases.

Researchers can use collective bargaining agreements to examine hypotheses related to employment discrimination and issues related to transfer decisions, promotions, and constructive discharge. For example, (Riccucci and Saldivar 2014) used Westlaw to search for employment discrimination suits filed against police and fire departments due to the failure to promote between 2000 and 2011. While the authors expected to find that women and people of color are filing lawsuits against police and fire departments, they found just the opposite—most of the lawsuits filed against police and fire departments are “reverse discrimination” lawsuits filed by White men.

To examine hypotheses related to employment discrimination and issues related to fair treatment, researchers can use employer codes of conduct. For example, (Mazza and Furlotti 2019) analyzed the codes of ethics of companies listed on the Italian stock exchange to discern content relating to employees as stakeholders. They noted that Equal Employment Opportunity statements were more likely to be included in codes of conduct in companies with low financial distress.

As demonstrated above, there have been some studies using the proposed methodology involving various data sources, which supports how the databases listed would provide experiences of discrimination. However, that an extensive search only identified a paucity of articles that have utilized this approach in over a decade warrants mentioning. This would support our contention that these databases, yet relevant, are being underutilized. This can be potentially advantageous to increasing our understanding of employment discrimination given the increasing amount of discrimination in the workplace related to the different bases of discrimination, harassment, retaliation, mutability/immutability, communications, discipline, and intersectionality.

Utilizing the proposed databases above is particularly attractive given content analysis and the era of AI, big data analytics, and NLP (Waltermann and Leeuw, 2025; Brewster et al., 2014; Tene and Polonetsky, 2013). To illustrate how researchers can implement the analysis using the suggested databases, consider examining disparate impact in the workplace. Disparate impact may exist when a seemingly neutral policy or practice disproportionately negatively impacts protected individuals. This raises the question of whether there are protections for the bases under the law.

To discern how disparate impact has been conceptualized and researched in the literature, researchers should consult prior research on disparate impact. It is not likely that prior research will have used the recommended databases and empirical legal scholarship. This underutilization of existing databases is one of the main points of this manuscript. Researchers could test this type of hypothesis using federally litigated or arbitrated cases. The researchers should be looking for terms that will be in the case content. At this point, researchers should confer with subject matter experts to help formulate the concepts. In addition, some content analysis software enables the use of trend analysis to discern trends in the data that might not be obvious to human experts. Content analysis software provides output in the form of frequencies of terms.

Researchers can use content analysis software to query the cases for the relevant terms. In the content analysis process, it is possible to identify synonyms for the relevant terms that are coded. It is also possible to eliminate terms that do not have their intended meaning in the given context, such as the term “Black” which could be the name of an individual or a company. Then the researcher can identify the variation in the cases to test the premise. It is possible to use a graphic portrayal of the data and multivariate nonparametric statistics to analyze the data. The results of this research could be structured in such a way as to provide changes in human resource management and organizational change processes.

The proposed methodology faces challenges, including the fact that some cases remain unpublished, making it unclear how well the published articles are representative. It has been contended that there are other methodological issues, such as the fact that the write-up of a case is post hoc, there are changes in the law over time, and researchers use inappropriate statistics and data aggregation methods. Additionally, safeguards are needed to protect subjects, especially when databases are shared.

The U.S.-specific nature of this research presents both strengths and limitations. While the federal legal framework and multi-tiered dispute resolution system create comprehensive databases, the findings may have limited generalizability to other national contexts, cultural norms, or discrimination complaint processes. Additionally, the current political climate's impact on DEI programs and discrimination reporting may influence the types and frequencies of cases appearing in these databases, potentially creating temporal bias that researchers must consider when interpreting trends and outcomes.

Published vs. unpublished cases as a concern is not dissimilar from non-respondents in a survey. In fact, this is not a new issue (Siegelman and Donohue, 1990; Swenson, 2004). This concern is more or less accepted as a limitation. Some publications have offered insight into unpublished cases (McAllister, 2019) and some researchers have been able to combine both published and unpublished cases in their analysis (Cooper and Barrett, 1984). There is even a contention that there is discrimination in how decisions are made regarding which cases get published and which cases do not get published (Tillman and Hinkle, 2018).

Ultimately, the methodology involved in this proposal is a form of policy capturing, which has been linked with contributions and limitations regardless of the subject matter. Early research has cautioned using cases for policy capturing (Roehling, 1993). In some regards, this caution can be briefly summarized as follows: The write-ups of the cases are post hoc, the evolving nature of the laws present sampling problems, the statistics involved may be too simple at times and too complex at other times, there can be systematic sampling biases, and there may be incorrect aggregation of the data.

Yet, we believe that these cautions are controllable in the data collection and analysis process. First and foremost, one has to take the position that while not all cases are published, the ones that are published have a greater impact on future litigation and policy-making. Additionally, the cases are frequently reviewed by more than one individual before publication. They frequently have expert witnesses, friends of the courts, and other documents to supplement the judges' rulings. One way of overcoming changes in the law over time is to have a shorter sampling time frame, and ideally, a more recent time frame. Additionally, the more widespread use of nonparametric statistics can overcome some of the statistical issues. Moreover, since (Roehling 1993), authors have contended (and reviewers and editors have agreed) that these are not insurmountable challenges (Clermont and Schwab, 2009; Goldhaber et al., 2016; Johnson et al., 2008; Knapp and Heshizer, 2001; McMullen, 2016; Zschirnt, 2019; Hajnal and Scharle, 2023).

Methodologists have been working to develop techniques to improve research using policy capturing (Aiman-Smith et al., 2002; Karren and Barringer, 2002) and large databases. As with other methodologies, methodological issues are consistently being addressed by methodologists, such as: the handling of missing data (Bonaccio and Dalal, 2010); the issue of socially desirable responses in self-reporting (Tomassetti et al., 2016), the issue of multi-collinearity (Paetzold, 1992), and the ethics in the use of correspondence studies (Zschirnt, 2019). Methodologists have also worked on issues relating to large databases (Ross et al., 2018). These authors have pointed out the challenge of sharing of data sets and still maintaining the privacy of the research participants (Ross et al., 2018).

It should be noted that some of these concerns are not applicable to the methodological approaches being discussed in this manuscript. There are absolutely no concerns regarding social desirability, the ethics of correspondence studies (Zschirnt, 2019) and the protection of subjects. Moreover, the data in these databases are considered archival, and an internal review by an Internal Review Board (IRB) is typically not required. In some of the databases, however, some identifying data is redacted.

First and foremost, our recommendation would be to use the large databases that are readily available, some of which are free, to research employment discrimination. We have identified at least 25 sources of this data type, but surely there are more. Additionally, we have suggested approaches to developing testable hypotheses utilizing the recommended data sources.

Future research should examine how current political and organizational headwinds on DEI programs and labor immigration policies influence discrimination patterns, complaint-filing behaviors, and case outcomes with these databases. Additionally, comparative studies examining discrimination patterns before, during, and after major events could provide valuable insights into how policy environments shape employment discrimination experiences.

Researchers in our discipline should learn from researchers' experiences in other disciplines, especially psychology, sociology, political science, gender studies, and law. It should be noted that some of these publications are in the top journals of their respective disciplines or subdisciplines.