The study of economics at HBCUs and PWIs

AbstractThis article’s authors use student transcript data to identify differences in the study of economics among Black students at HBCUs and PWIs. The data show that a higher fraction of Black students at HBCUs initially intend to study economics, relative to those at PWIs (4.0% vs. 1.3% of micro principles enrollees) and persist in the major (9.4% vs. 3.8%). Logit analysis suggests that (1) academically stronger Black students are less likely to persist to an economics degree at both institution types and (2) Black female students at HBCUs are as equally likely to persist to a degree in economics as their male counterparts while those at PWIs are less likely to persist. Additional research is needed to determine the causal factors responsible for these outcomes.Keywords: Economics educationHBCUprinciples of economicsPWIstudent achievementJEL CODE: A2 Disclosure statementNo potential conflict of interest was reported by the authors.Notes1 The study began in 2015 and follows a cohort of freshmen through graduation at 20 treatment schools and at least 30 control schools.2 Both Bayer and Wilcox (Citation2017) and Bayer and Rouse (Citation2016) provide research evidence on “why economists should care about diversity,” including fairness, broadened views of economic policy outcomes, and higher quality research. Former Federal Reserve Chair Janet Yellen (see Yellen [Citation2014] and Chaney [Citation2018]) has similarly emphasized the value of diversity in the discipline and the Fed’s commitment to promoting diversity in the profession.3 Studies discussed here were part of a 2021 ASSA session entitled “Increasing Diversity in Economics: From Students to Professors.”4 See appendix A for a description of Historically Black Colleges and Universities (HBCUs). Hispanic-Serving Institutions (HSIs) are defined by the U.S. Department of Education as an institution of higher education that meets particular eligibility criteria and “has an enrollment of undergraduate full-time equivalent students that is at least 25 percent Hispanic students at the end of the award year immediately preceding the date of application” (U.S. Department of Education Citationn.d.)5 Clewell, de Cohen, and Tsui (Citation2010) examine the role that the NSF’s HBCU-UP (Historically Black Colleges and Universities Undergraduate Program) grant program has played in building institutional capacity at HBCUs to educate STEM students. In particular, they note the “success of these institutions (HBCUs) in creating a nurturing environment that fosters psychosocial health among African American students, resulting in their satisfaction with and integration into the academic environment” (p. 7).6 While we focus on HBCUs in this study, data in Sharpe and Swinton (Citation2018) illustrate that HSIs also have grown in importance as producers of baccalaureate degrees in economics for Blacks in the last two decades. In 2006, the number of Blacks earning undergraduate economics degrees at HBCUs was more than three times greater than at HSIs; by 2015, Black students at HSIs earned 33 percent more economics degrees than at HBCUs. During this time, the number of HSIs in the United States more than doubled (from 201 to 412), while the number of HBCUs (99) remained the same.7 Here, AEA refers to the American Economic Association, CSWEP refers to the AEA’s Committee on the Status of Women in the Economics Profession, NEA refers to the National Economic Association, and ASHE refers to the American Society of Hispanic Economists.8 While MIDFIELD also includes Florida State University and the University of Colorado, we exclude them from this analysis because both institutions have hurdles to declaring an economics major, which results in no students in the introductory economics courses identified as economics majors. Given that we study the characteristics of students across majors at the time of their microeconomics principles course enrollment, these barriers to entry necessarily result in no observable majors at these institutions. As such, we omit observations from these institutions from the present analysis. Additionally, we omit observations from North Carolina State University due to an extremely small sample size. Finally, we note that our access to the MIDFIELD data predates its official publication and thus is a pre-publication version of the subset of students who took economics without any version code identifiers.9 Most business schools require a combination of courses for admittance, including principles of micro and macroeconomics, financial and managerial accounting, and calculus and statistics.10 These two schools are also included in Bayer and Wilcox’s (Citation2017) list of 12 four-year HBCUs (in their Appendix Table 4) that awarded at least 25 majors in economics to U.S. citizens or permanent residents (i.e., excluding nonresident aliens) in the period 2011–15.11 Moving from table 2 to tables 3 and 4, we have a reduction in sample size of 89 students (22 at HBCUs and 67 at the PWIs). For some years, some of the institutions have no economics majors. These years are dropped in the logit analysis as a result of our inclusion of year-fixed effects. The result is a slight difference in the sample size.12 A student’s declared major is defined as the major at the time that they enrolled in their first microeconomic principles course.13 SAT score is the combined SAT verbal and SAT math scores, or the converted ACT score for those taking only the ACT exam. We also considered an alternative standardized SAT/ACT measure, but results were consistent with those for SAT, so we include only the more common measure here. Graduation rates in table 1 are unrestricted by time and reflect only whether the student is observed graduating at any point in our sample period.14 Cumulative GPA is based on all courses taken prior to the introductory economics course. For students taking their first economic principles course in their first semester (14.0%), this value is missing.15 We include financial accounting in our model because it is a gateway course to the business degree, and all institutions in our sample have a business school. Thus, having completed macro principles and/or financial accounting is a potential indicator of a commitment to the business degree. This is important as Kasper (Citation2008) documents that economics and business are complementary majors; thus, having completed financial accounting suggests a greater commitment to both the economics and business majors.16 We note some differences in significance between tables 3 and 4. According to Greene (Citation2008, 12),An empirical conundrum can arise when doing inference about partial effects rather than coefficients. For any particular variable, wk, the preceding theory does not guarantee that both the estimated coefficient, θk and the associated partial effect, δk will both be “statistically significant,” or statistically insignificant. In the event of a conflict, one is left with the uncomfortable problem of simultaneously rejecting and not rejecting the hypothesis that a variable should appear in the model. Opinions differ on how to proceed. Arguably, the inference should be about θk, not δk, because in the latter case, one is testing a hypothesis about a function of all the coefficients, not just the one of interest.Following Greene, we focus our discussion on the estimated coefficients from table 3. However, we also present marginal effects estimates in table 4 to provide a sense of magnitudes. Furthermore, we note two important points regarding the differences in significance. First, the differences in significance likely stem from a combination of the nonlinearities involved in marginal effects estimation and the use of control measure means to calculate estimates. This is particularly important, considering the significance differences surrounding the micro principles course grade and the interaction between course grade and female. Second, the differences in significance for completion of financial accounting and calculus may appear to be starker than they are. Estimated coefficients for these controls were marginally significant (9% level), and the marginal effects just miss the arbitrary 10 percent cutoff for marginal significance (at 11%).