Discrimination by Black Males Revisited:.: Salary Equity Achieved

I Introduction At the direction of the board of regents, SHBU(1) commissioned a national consulting firm to execute a faculty salary equity study in 1990. The consultants' preliminary model, presented to the faculty in 1991, found significant differences in salaries among faculty race-sex subgroups. Their statistical tests were conducted at the 0.01 level of significance. Subsequently, they added three variables to "more completely specify [the] model" (Faculty Salary Equity Study Faculty Briefing 1992). The added variables measured performance, demand, and administrative experience. Performance was not statistically significant; demand and administrative experience were, so they were included in the consultants' final model. In their final model, the consultants found that none of the race-sex variables were statistically significant at the 0.01 level, but because the Hispanic-Asian female variable was significant at the 0.0161 level, "SHBU determined it merited further action" (Faculty Salary Equity Study Faculty Briefing 1992). Nevertheless, SHBU agreed to adjust salaries for all faculty identified by the consultants' model as having low salaries, striving to raise them to 92.7 percent of market. Adjustments approaching $1 million were made over a two-year period starting in 1993. This study contrasts the consultants' salary model with the model of Riggs and Dwyer and uses the latter to answer an important question: Has SHBU eliminated faculty salary discrimination based on race and gender? II Salary Determinants: The Models The original Riggs-Dwyer study of SHBU, based on data from 1989 to 1990, found evidence of salary discrimination by black males against other race-sex groups. Yet, the consultants found no evidence of discrimination using similar data for the same period. This difference in conclusion is partly explained by the consultants' selection of 0.01 versus Riggs' and Dwyer's selection of 0.05 as the level of significance. We chose 0.05, a more conservative level, because it is more consistent with significant levels in similar studies. The conclusions also vary because of differences in the salary models used to estimate discrimination. Riggs and Dwyer and the consultants included in their models a variable reflecting the market value of faculty members in various ranks and disciplines. The consultants used College and University Personnel Association (CUPA) salary data for SHBU's ten peer institutions, whereas Riggs and Dwyer used market values reported by the Oklahoma State University study of land-grant institutions. The results were comparable for this variable. The consultants included an additional demand measure in their model. Riggs and Dwyer did not include it because "demand," properly constructed, would correlate highly with "market" salaries and hence would be redundant. The variable was highly significant in the consultants' model: each unit on their "1 to Y" demand scale contributed an extra $1,135 annually to salary. Because we were unable to discover the basis of the "1 to 3" rankings, we cannot explain what this variable measures. The most important difference in the models related to administrative experience. In a departure from the model by Riggs and Dwyer, the consultants' final model included a variable reflecting SHBU administrative experience. This highly significant variable contributed from $4,811 to $5,673 to annual salaries depending on one of three administrative classifications. Likewise, our original study found that administrative experience was highly significant in determining salary; it added an average of $4,815 annually. We also found that black males had a great advantage over other race-sex groups in gaining SHBU administrative experience, even though they demonstrated no evidence of superior qualifications. This rendered administrative experience a vehicle for salary discrimination, so we excluded it from our final model. …