We again fitted a generalized linear model to the binary outcome indicating the promotion decision and accounted for possible correlation between cases in the same department within the same institution by implementing the method of GEE to obtain improved standard errors for all model parameters. Most of the explanatory variables in the promotion decision model were the same as those used earlier in the tenure decision model. Here, however, we considered the percentage of women among tenured associate professors to be the “promotion pool.”
None of the variables included in the model appeared to be associated with the probability that a candidate was promoted to full professor. In particular, there were no significant differences (p = 0.607) in the probability of promotion to full professor due to gender of the candidate, after accounting for the other potentially important factors. Therefore, it appears that once proposed for promotion to full professor, women and men fare about the same across all types of institutions and departments.
There are several reasons our analyses may produce different results than those reported previously in the literature. First, the studies use different samples; for example, faculty at different types of institutions or in different fields. A more interesting difference is conceptual. While the committee’s survey focused on the tenure and promotion decision, most prior studies focus on whether or not an individual is tenured or promoted by a particular point in time. To better illustrate this distinction, one can compare the results of a 2006 Pennsylvania State University study of faculty cohorts entering the tenure track between 1990 and 1998 As Table 5-6 shows, 7 years after entering the tenure track, the percentage of men who were tenured professors at Penn State was higher than the percentage of women who were tenured.
However, in a second analysis, Dooris, Guidos, and Miley (2006) examined the outcome of reviews of faculty who were evaluated 6 years after being hired. As seen in Table 5-7, the rates for men and women were not significantly different at the sixth year review, with observed values of 90 percent for men and 87 percent for women (p = 0.69). The differences between these two ways of examining the data may be due to the departure from Penn State by some faculty, who never came up for review. Alternately, the results may reflect that some faculty took leave, delaying the tenure decision for them beyond the sixth year.
These two foci—tenure status after a specific time period and tenure decisions—correspond to different but partially overlapping groups of faculty. In the committee’s study, the denominator included any faculty who came up for a tenure or promotion decision. In the other studies briefly surveyed above, the denominator included both individual faculty who came up for a decision and tenure-track faculty who have not yet reached that point. A second reason is that men and women may spend different amounts of time at each rank. This topic is discussed below.