Affirmative Action at the University of Louisville: A Case Study

Ben Foster is professor of accountancy at the University of Louisville, Louisville, KY 40292 and president of the Kentucky Association of Scholars; [email protected].

This piece originally appeared in the Spring 2016 Academic Questions (volume 29, number 1).

Fifty years after the passage of the Civil Rights Act of 1964, debate has intensified within the courts and society at large over the necessity of the continued pursuit of “diversity” through affirmative action and equal employment opportunity guidelines. Also, some attention has focused on whether applications of affirmative action and “equal employment opportunity” policies are themselves discriminatory.

In light of the controversy, I examined the application of affirmative action policies in one university because the mechanics of applying such policies are seldom scrutinized.  Affirmative Action/Equal Opportunity reporting involves specific jargon and much micro-statistical detail. Close examination of one case offers us the opportunity to see the peculiar results of these procedures, such as the example of a nursing department with women holding forty-one of its forty-four faculty positions that was judged to have an insufficient number of women.

One issue of contention is how human resources departments classify people into groups. The ethnic category of Hispanic, for example, does not appear to be clearly defined. Does classification as Hispanic require a Spanish surname, ancestors from certain countries, a certain eye color or skin pigment? Likewise, how is someone identified as black, or African American? Must an individual have a certain proportion of African ancestry, or does the one-drop-of-African-blood rule apply? Do human resources departments have a verification process in place? Sophisticated DNA tests are now available to analyze ethnic heritage. Should such testing be required of applicants claiming minority ethnicity before they can be counted toward meeting established diversity goals? The issue has reached high-profile status with such cases as former University of Colorado–Boulder professor Ward Churchill and Senator Elizabeth Warren, formerly of the law schools of the University of Texas, the University of Pennsylvania, and Harvard, both of whom falsely claimed to have Native American ancestry, as well as more recently—and more bizarrely—that of Rachel Dolezal, a white woman and president of the National Association for the Advancement of Colored People chapter in Spokane, Washington, from 2014 until June 15, 2015, when she resigned following allegations that she had lied about her racial identity and other aspects of her biography.

Even if human resources departments have standard definitions of minority group membership and verification processes in place, fundamental problems remain. Asians and Pacific Islanders are typically categorized as a minority group, but why would citizens of China, India, and Japan receive hiring preferences over people from Eastern European countries? Likewise, why would citizens of Spain receive hiring preferences over citizens of Italy, Greece, or France? More generally, why would an American university give preference in recruitment and hiring to immigrants who meet certain racial or ethnic criteria over native-born non-Hispanic white Americans?

These questions will require further debate and study (and probably multiple court decisions) before they are resolved. Some issues have been resolved in the courts, however, and all organizations receiving federal funds are required to comply with certain laws and regulations. Almost all colleges and universities must complete an Equal Employment Opportunity/Affirmative Action (EEO/AA) Program report to show compliance with the nondiscrimination requirements under the Civil Rights Act of 1964. These reports reveal policies used and actions taken by schools in pursuit of diversity goals. To investigate the practical impact of diversity efforts at my own institution, the University of Louisville (UL), I examined the most recent (2012) EEO/AA Program report, prepared by UL’s human resources department, as well as other materials available at the UL website.^[1] What follows are my findings of a discriminatory employment impact of diversity efforts for one university. These findings may be relevant to many other colleges and universities in the United States.

University of Louisville EEO/AA Program

A significant part of the EEO/AA Program report contains an “Incumbency Versus Availability Detail” (IVAD) for employees, by job category. Sam Connally, former vice president for human resources and UL Equal Employment Opportunity officer, met with members of the College of Business Diversity Committee (including me) in November 2014 to explain how UL uses the IVAD, and informed us that the report was necessary to show UL’s compliance with the nondiscrimination requirements under the Civil Rights Act of 1964 and Executive Order (EO) 11246. EO 11246 prohibits federal contractors and subcontractors from discriminating on the basis of race, color, religion, sex, or national origin in employment decisions. For each job group, Title 41 of the Code of Federal Regulations §60-2.15 requires contractors to compare the percentage of women and minorities currently employed to the availability of qualified women and minorities in order to identify groups that have fewer than the reasonably expected number of women and minorities as employees. The difference between actual and expected numbers that creates an “underutilization problem,” however, is not specified in the regulations.[2]

According to consultants who assist organizations with EEO compliance, different quantitative analyses are allowed when investigating underutilization for the AA Program under EO 11246, including the “Whole Person” method used by UL, the “80% Rule,” “Standard Deviation,” and “Exact Binomial.”[3] Regulators appear to prefer the Standard Deviation method.[4] UL’s proposed 2014 EEO/AA Plan states that the university used the Two Standard Deviations test prior to 2010, when the “whole person” method was adopted.[5]

The whole person rule is the strictest test for determining underutilization and was used by UL recently to identify job categories in which female, minority, black, Hispanic, Asian, and American Indian population subgroups were identified as underutilized. For each area separately evaluated, the percentage of women and minorities available in that category is determined, as measured nationally for faculty and many administrative positions and locally for many staff positions. Then the number of employees in the employment category multiplied by the appropriate percentage available is calculated to determine the number of expected women and minority employees. If this expected number minus the actual number of female or minority employees in that area was 1.0 or higher (a “whole person”), the area is considered to underutilize women or members of a minority group. 

For example, as shown in table 1, UL’s 2012 program showed thirty-one total faculty in Category 20C: “Electrical and Computer Engineering.” Of available electrical and computer engineering professors, 20 percent were Asian; 31 times 0.2 indicates that UL is expected to have 6.2 Asian faculty in Category 20C. UL employed only five Asians in Category 20C—1.2 less than expected—and the category was consequently identified as underrepresented by Asian faculty.  

In the EEO/AA Program report, the term “underrepresented” is used interchangeably with the term “underutilized.” I also use these terms interchangeably in this article. Also, the report does not address the opposite situation, where the actual number of incumbent employees in certain groups is more than the expected number of female or minority employees in those groups. Throughout this article I refer to those situations as instances of “overrepresented” women or minorities.

Connally noted that areas identified as underutilized are subject to strict internal scrutiny in search and hiring practices to address presumed discrimination. He indicated that human resources oversees searches and hiring by university units for the job groups and population groups so identified. On page 4 of the executive summary and page 14 of the report, tables present the faculty job groups in which women and minorities are underrepresented in relation to market availability nationwide among faculty job groups. Page 14 states:

Academic Deans, Chairs, and Search Committees should continue to give particular attention to recruitment and outreach strategies that target women and minorities in the foregoing disciplines, as envisioned in the University’s Affirmative Action Plan.

Another UL human resources publication, “University of Louisville Reaffirmation of Commitment to Equal Education & Employment Opportunity,” identifies the university’s effort to show compliance, but goes further than the EEO/AA Program report, stating that members of “groups [that] have historically been and continue to be underrepresented within the University in relation to the availability may be extended preference in initial employment…among substantially equally qualified candidates” (emphasis added).^[6] This statement appears to contradict federal regulations for government contractors, which specifically state that “Placement goals do not provide the contractor with a justification to extend a preference to any individual, select an individual, or adversely affect an individual’s employment status, on the basis of that person’s race, color, religion, sex, or national origin.”^[7]

Comparison to Earlier Report

Around ten years ago I conducted an informal examination of an EEO/AA Program report and its IVAD based on 2003 employment numbers.^[8] UL used the whole person rule for that report. In an August 2004 letter to UL president James Ramsey I described concerns that the information included in the 2003 report indicated inappropriate discriminatory use of the IVAD, noting two glaring problems:

In fact, the current focus on [numerous] categories...with very few employees can lead to reverse discrimination when no attention is given to areas overutilized by women and minorities. Under U of L’s current method, the University could conceivably still find specific job categories that underutilize women and/or minorities even if total U of L employees were 80% women and 50% minority.

Legitimate science requires statistical significance to state that one group is under- or over-represented in a sample. Conducting easy statistical tests generally requires at least 30 observations. Currently, 40 of the 78 job categories used by U of L contain less than 30 employees—7 of these categories have less than 10 employees.^[9]

I recall receiving an e-mail response from President Ramsey later in 2004, copying Human Resources and stating that he essentially agreed with my concerns. In the earlier report, as stated above in the second paragraph of the excerpt of my letter, there was a problem in how underutilization could be found in departments and areas with too few numbers of faculty to be statistically significant.  That problem persists in the 2012 report, where thirty-three of eighty-one job categories used by UL contain less than thirty employees—and six of these categories have fewer than ten employees.

Connally indicated that a job category would no longer be subject to strict internal scrutiny by Human Resources once it is staffed with the expected number/percentage of women and minorities.  In a comparison of my 2003 information to that in the 2012 program, I found that in nineteen job categories, thirty job/population combinations identified as having an underutilized disparity in 2012 were not identified as having an underutilized disparity in 2003. In fact, six of those job/population categories had more than the expected incumbency by substantial amounts in 2003. This 2003 data indicates no history of discrimination in those areas, and therefore no reason to require strict internal scrutiny of hiring practices for those positions by UL human resources administrators.

For example, job category 20V, “Medical,” is tagged for “particular attention” in 2012 because women are identified as having one less faculty member than expected (7). Category 20V was not identified as lacking in women in 2003, when 27 percent of the faculty was female, while 35 percent of the faculty was female in 2012. Category 20P, “History,” is also now noted as underrepresented by women, when it was not in 2003.

The detail in the 2012 program for category 20I, “Dental,” indicates that black and Asian faculty members were underrepresented, while neither group was listed as underrepresented in 2003. Category 21A, “Miscellaneous Education,” too, shows underrepresentation of minority and black faculty members in 2012, but not in 2003.

Granted, between 2003 and 2012 many more women and minorities were probably added to the numbers in various categories, raising the expected hiring percentage, at least partly due to affirmative action in higher education.  However, 2003 data indicate no history of discrimination warranting preferential treatment according to UL’s strict internal scrutiny policy described by Sam Connally.

The most unusual case noted as underrepresented by female faculty in 2012, according to the whole person rule, was 20J, “Nursing.” Nursing was not shown as underrepresented by women in 2003, but was noted as underrepresented by female faculty in 2012, according to the whole person rule, even though forty-one of forty-four faculty members were women. The field of nursing is so overwhelmingly female that the available pool is over 95 percent female. Consequently, over 95 percent of nursing faculty at UL is expected to be female according to the whole person rule, another ironic example of rule-based pursuit of diversity when following U.S. Equal Employment Opportunity Commission guidelines.

Similar to what I noted in the first paragraph of the excerpt from my 2003 letter to President Ramsay, despite overrepresentation of minorities in certain categories or even overall in a department or school, the 2012 report focuses on each individual department or area. Then, flagging departments as underutilized when employees from other groups are not exactly proportional to the number expected, or higher, is another instance of rule-based pursuit of diversity and micromanagement that can lead to discriminatory practices against males and Caucasians. For example, in table 1 Asians are identified as underrepresented in category 20C, “Electrical and Computer Engineering” (5 incumbents - 6.2 expected = -1.2), and 20F, “Miscellaneous Engineering” (3 incumbents - 4.6 expected = -1.6), but are actually “overrepresented” in categories 20A and 20B.

Overall, Asian faculty are right at the level expected in the Speed School of Engineering, (as seen in table 1), and in 2012 female professors are substantially overrepresented (17 incumbents - 8.2 expected = 8.8). A slight underrepresentation of black faculty (specifically noted for categories 20C and 20E) would also be identified at the School of Engineering level.

It is particularly interesting that Asian faculty are identified as underrepresented by two positions in category 20I (Dental), when they are so heavily overrepresented in other categories. For example, in categories 20V (Medical) and 20H (“ICT & Bioinformatics”), Asians are overrepresented by twenty-two and fifteen faculty members, respectively, out of the total number in each area. With no evidence or indication of discrimination against potential Asian faculty and when Asians are so overrepresented in the medical school, why provide preference to Asians applying for jobs as dental faculty over other population groups? But this is the sort of bureaucratic blindness that arises from the unrealistic expectation that the percentage of women and minorities in every category of a university’s workforce must be at least as high as the percentage of available women and minorities.

Table 1 2012 Data for University of Louisville Speed School of Engineering Faculty
		Incumbents
Job Group	Total	Female	Minority	Black	Hispanic	Asian	American Indian
20A	8	2	5	1	1	3	0
20B	8	1	3	0	0	3	0
20C	31	3	6	0	1	5	0
20D*	9	2	3	0	0	2	0
20E	18	4	4	0	1	3	0
20F	23	5	6	2	1	3	0
Total	97	17	27	3	4	19	0

			Expected
Job Group			Female	Minority	Black	Hispanic	Asian	American Indian
20A			0.7	2.3	0.4	0.2	1.6	0.1
20B			0.7	2.3	0.4	0.2	1.6	0.1
20C			2.6	9.0	1.7	0.8	6.2	0.2
20D			0.8	2.6	0.5	0.2	1.8	0.1
20E			1.5	2.5	1.0	0.5	3.6	0.1
20F			1.9	6.6	1.2	0.6	4.6	0.2
Total			8.2	25.3	5.2	2.5	19.4	0.8
Over (under) represented			8.8	1.7	(2.2)	(1.5)	(0.4)	(0.8)
*20D listed one minority faculty member as meeting criteria for two or more categories, but did not specify which categories.

In addition, focusing on each department can lead to the anomalous situation in which the university’s overall numbers show overrepresentation of women and minorities, yet application of the whole person demands furthering preferential hiring of these groups. Tables 2 and 3 provide information on the total workforce at UL that I constructed using 2003 and 2012 data, respectively. Immediately obvious from a comparison of the two tables is the tremendous growth in the number of employees from 4,373 in 2003 to 6,227 in 2012.^[10] (UL implemented a hiring freeze and an early retirement program around 2012.) In both years, well over half of university employees were women, and female employees were overrepresented compared to the expected number by a substantial amount: 62 in 2003 and 246 in 2012. “Minority,” “Black,” and “Asian” contained substantially more than the expected numbers for those groups in both 2012 and 2003. In addition, the overrepresentation of those three categories was even larger in 2012 than in 2003. 

Table 2 2003 Report Analysis of University of Louisville, Total Workforce
	Percentage Actual	Percentage Expected	Total Employees Actual	Total Employees Expected	Excess (Deficiency)
Total Workforce	100.0		4373
Female	55.9	54.4	2443	2380.9	62.1
Minority	20.5	12.1	898	530.2	367.8
Black	13.0	8.4	568	365.4	202.6
Hispanic	1.2	1.5	54	65.8	(11.8)
Asian/Pacific Islander	5.9	1.9	260	84.8	175.2
American Indian	0.4	0.3	16	12.6	3.4
Note: The total excess for minorities is slightly off due to minor errors in the UL report.

Table 3 2012 Equal Employment Opportunity/Affirmative Action Program Report Analysis of University of Louisville, Total Workforce
	Percentage Actual	Percentage Expected	Total Employees Actual	Total Employees Expected	Excess (Deficiency)
Total Workforce	100.0		6227
Female	54.1	50.1	3367	3120.7	246.3
Minority	24.2	18.0	1509	1122.9	386.1
Black	10.8	6.9	674	427.1	246.9
Hispanic	2.1	2.3	128	144.1	(16.1)
Asian/Pacific Islander	9.9	6.0	618	375.8	242.2
American Indian	0.2	0.5	11	33.3	(22.3)

Conclusion: Legality of Personnel Policies and Practices

Most of the issues mentioned in my letter to President Ramsey about the 2003 report appear to be as relevant today, since UL returned to using the whole person rule in 2010. (Evidently, between 2003 and 2010 UL did use a more reasonable method of identifying job groups “underrepresented” by women and minorities.) Current personnel methods and policies lead to the following issues and problems for UL:

1.  Identifying specific groups as underutilized within narrow job categories using the whole person rule leads to a presumption of discrimination, even when overall numbers show acceptable proportions of minorities and women. Preference in hiring members of certain groups over white non-Hispanic males is unfair when overrepresentation by these groups in other job categories is not considered.
2. Offering preference to members of certain groups in job categories where no discrimination was indicated ten years ago appears to violate Titles VI and VII of the Civil Rights Act of 1964.  I have never heard any discussion or seen any evidence that UL has engaged in discrimination against Hispanics or Asians, particularly in faculty hiring. However, in the “University of Louisville Diversity Plan 2011–2015,” these minority groups—Hispanic/Latino, Asian, and American Indian/Alaskan Native—are targeted for increases in representation even in faculty and professional non-faculty job categories where they are calculated as just slightly underrepresented.^[11] 
3. Underrepresentation of Hispanic/Latino, Asian, and American Indian/Alaskan faculty is likely identified because “availability” percentages come from national data. Rather than discrimination, the purported underrepresentation of these groups is more likely a natural result of geographic distribution of populations such as lower concentrations of Hispanic/Latinos and Asians in Kentucky than in California, and lower concentrations of Native Americans in Kentucky than in Oklahoma.

Based on my analysis of UL’s 2003 and 2012 employment data, UL diversity plans/objectives included in documents such as the “2020 Plan”^[12] and the “University of Louisville Diversity Plan 2011–2015” submitted to the Council on Postsecondary Education appear to push the university to act in a discriminatory fashion contrary to the Civil Rights Act of 1964 and the Title VI strict scrutiny standard. Those plans list goals to increase employment percentages or numbers of African American, Hispanic, and Asian employees in many job categories. In particular, given that the overall UL workforce is overrepresented with female, black, and Asian employees, continued use of the current method appears unduly discriminatory. Thus, on their face, these documents indicate that UL goals value certain people more than non-Hispanic whites due to their appearance/ethnicity/ancestry.

No UL administrator has informed me of any plans to address any of the concerns raised in this article. As an interesting endnote to this study, UL President Ramsey was embroiled in a controversy in late 2015 because he and his staff participated in a Halloween luncheon in which everyone donned sombreros and other traditionally Mexican attire. President Ramsey and his staff apologized for the event. However, after protests involving some students, faculty, and staff, the president and his staff pledged to raise resources to:

• recruit more Hispanics for faculty, staff and administrative positions,
• build funds for Hispanic scholarships and financial aid,
• improve facilities to house all diversity support programs, and
• support more fully and equally UL’s programs for underrepresented faculty, staff, and students.