Retention and Promotion of Male and Female Faculty members at UIUC

Jane W. Loeb and Susan Greendorfer

In the early 1990s, the Chancellor's Committee on the Status of Women (CCSW) at UIUC asked to review existing data on the progression of women and men through the faculty ranks on the Urbana campus. This study was motivated by an informal observation that the campus, similar to higher education in general, seems to make noticeably better and faster progress in hiring women at the entry level than in "integrating " the senior ranks by gender. This observation suggests that women may be leaving the campus and/or higher education more frequently than men, or that women may experience a lower rate or a slower speed of promotion through the academic ranks. It seems this local observation is quite reflective of the literature on gender differences in academic progress and rewards. In his early evaluation of the equity of academic recognition and rewards accorded to men and women, Jonathan Cole (1979) found that female faculty in biology, chemistry, psychology, and sociology held significantly lower academic rank than did males after controlling for publications and citations. Further, discrepancies in rank tended to be worse in more prestigious departments (Cole, 1979; Long, Allison, & McGinnis, 1993; Sonnert, 1995). Similarly, in their study comparing the rank distributions of men and women over time, Bentley and Blackburn (1992) found no evidence of improvement, concluding, "the revolving door is still spinning"(p. 701).

In the initial faculty retention study performed for CCSW Loeb (1991) traced the later rank of faculty hired between Fall 1976 and Fall 1985 as assistant professors as well as those hired or promoted to the rank of associate professor during the same years. In order to control for possible differences among disciplines, equal numbers of men and women were selected for study from the departments that had hired or promoted individuals of both sexes during the study period. The overall retention rates of men and women were very similar, as was their progress through the ranks. Where differences were found, men were favored. However, the differences were not significant, which suggested that these differences might not be stable if the study were repeated on another cohort.

In the late 1990s, the UIUC Chancellor's Committee on the Status of Women asked for an update and extension of this initial study in order to see whether or not the small differences that favored men were persistent over time and to permit a longer term follow up of the associate professor cohort in order to assess the final, long-term rank attained by these men and women. Additional purposes of this updated study were to collect survey data from faculty who left the campus and to investigate their perceptions of the working conditions and climate they experienced while at UIUC. The data provided here represent an analysis of the results of this extended study and present comparable analyses from the earlier study for comparison.

Method

Samples and Institutional Data: The initial study involved two groups of faculty. An assistant professor cohort consisted of most female faculty members who were first hired as assistant professors between Fall, 1976 and Fall, 1985. This time period allowed for the completion of the six-year probationary period by the end of Spring, 1991 for those hired in Fall of 1985. For each female assistant professor, a male was randomly selected from among those hired as assistant professor in the same department during the same time period. The associate professor cohort consisted of faculty hired as or promoted to associate professor during the same ten-year period, Fall 1976 through Fall 1985. Men were randomly selected from among those new associate professors in the same departments as the women, parallel to the sampling process used for assistant professors. Data concerning degree date, experience prior to UIUC hire, initial tenure code, progression through the ranks, leaves taken, death, resignation or retirement, were collected from campus personnel records. Except for tenure status of assistant professors, data were traced through June 21, 1991. Tenure status of assistant professors was checked as late as August 21, 1991 in order to capture tenure decisions made during the 1990-1991academic year.

For the current study, the same method was used to define a new cohort of assistant professors hired from Fall 1986 through Fall 1992. Both these assistant professors and the initial group of associate professors were followed through the 1998-99 academic year. The full set of information was captured from campus records as before, with one exception. The fact of death, resignation or retirement had been recorded in the typewritten cumulative personnel record that was discontinued recently without replacement. Thus, information concerning death, retirement and resignation during recent years was gleaned from the Payroll File and was available for very few of the faculty members who left the Urbana campus. Since some members of the associate professor cohort are now old enough to be experiencing death or retirement, it is a methodological weakness that we were unable to correct, as we could not ascertain which and how many of these faculty who recently left UIUC had died or retired.

Follow-up Survey: A questionnaire was designed to probe perceptions of: (a) the support provided for research and teaching, (b) the department's evaluation of the individual's research and how it compared to the individual's salary and his or her own perceptions of research productivity; (c) teaching and service loads, (d) the kind, usefulness and frequency of feedback received on progress toward promotion, tenure, and career success, (e) reasons for leaving; the work environment, and (f) information about the person's current job. For the open-ended questions pertaining to reasons that might have contributed to leaving UIUC, categories of types of reasons were developed by reading responses to the questionnaire that is regularly sent by the Office of Instructional Resources to those faculty leaving UIUC for other jobs whose departure is seen by their departments as a loss.

A search of the literature on perceptions of the work environment for female faculty revealed the availability of an Academic Work Environment Scale for Women (AWESW), a questionnaire that was designed and tested at UIC (Riger, Stokes, Raja, and Sullivan, 1997; Stokes, Riger, and Sullivan, 1995). Riger et al. (1997) developed a scale with 35 items and a short-form containing 13 items. Principal components analysis of the 35 items suggested three dimensions, of which the first was dominant and seemed to be a general measure of departmental climate for women. The short form had a coefficient alpha of .94, and a correlation of .97 with the long form. The short form of the AWESW was included in the follow-up questionnaire sent to members of the assistant and associate professor samples who left UIUC.

The follow-up survey was piloted on a small group of UIUC faculty of mixed rank and gender. After making editing changes and corrections, the final form was then mailed, in September of 1999, to faculty who had left the campus. Follow-up postcards were sent to non-respondents in November, 1999. Addresses for this mailing were found in a general, national faculty directory, various disciplinary directories, and through other resources. These included UIUC department and personal contacts, as well as looking for recent publications by individuals to find out their institutional affiliation and then checking the web page for that institution.

The population that was surveyed included all members of the assistant and associate professor samples who had left UIUC and were not known to have retired or died. Of the associate professors who had left, there were 50 males and 52 females. From these numbers, two males and 14 females were known to have died or retired, yielding populations of 48 males and 38 females. Of these, no functional address was available for 10 men and 7 women. Thus, the 19 responses received from men and 19 from women represent 50% and 61%, respectively, of men and women with addresses (or 40% and 50% of the entire population). In addition, 14 female faculty members were surveyed who were in the initial 1991 sample of newly hired or promoted associate professors but had been dropped due to lack of a male hired or promoted in the same department during the same time period. They were included in the hope that their results could shed light on potential response bias in the initial sampling. Unfortunately, only 8 had useable addresses. Six of these responded, too few for the intended use. However, since the balance between the male and female groups within the same department had essentially disappeared through attrition (due either to no address or non-response), these six responses were included with the others in the analysis.

Of the assistant professor sample, 77 men and 73 women had left UIUC, and none were known to have died or retired. No functional address was found for 16 men and 10 women, leaving 61 and 63, respectively, to be surveyed. Of these, 25 men and 21 women responded, yielding response rates of 41% and 33%, respectively, of those who are believed to have received surveys. (These are 32% and 29% of the full group of assistant professors who left UIUC.)

Results

Institutional Data. Table 1 indicates that UIUC has hired considerably more women assistant professors in the 1986-92 period (37.4% of all new assistant professors) than during the 1976-1985 period (19.9%). From one-half to three-quarters of the campus's newly hired or promoted assistant and associate professors are represented in these three groups.

Table 2 indicates that overall about one-half of assistant professors achieve tenure. This table also displays small but consistent differences in rates of tenure attainment favoring men. With 9% and 6% more male than female assistant professors achieving tenure in the two time periods, the overall difference of 7% is statistically significant at the one-tailed .05 level.

Table 3 indicates that approximately half of the two groups of assistant professors remained employed by UIUC in Fall of 1990 or Fall of 1998. There is no consistent gender difference in retention. Table 4 indicates the final academic ranks achieved by these men and women who were retained and shows a small but consistent tendency for women to be over-represented in the lower ranks and not to have achieved the rank of associate professor or higher.

Table 5 tallies the final rank held for faculty hired as assistant professor who had left UIUC by the end of the study period. A consistent tendency for women to leave at lower ranks than men is seen both in the frequencies and in the mean rank, where the mean is based on arbitrary codes of 1, 2, and 3 for the three ranks and on the assumption of equal intervals between ranks. For the two cohorts combined, the difference in mean ranks is significant at the two-tailed .05 level.

In Table 6, the same data are displayed for all assistant professors in the two cohorts. Again, there is a small but consistent tendency for women to hold lower ranks, and the difference in mean ranks for the combined groups is significant at the one-tailed .05 level.

Table 7 indicates how many men and women left before the tenure decision (tenure code of 1 through 5) after denial of tenure (tenure code of 6 and T), or after achieving tenure (tenure code of A). (It is assumed here that individuals who leave after serving at tenure code 6 have left after denial of tenure, but before the award of a T contract.) Consistently, men and women who leave UIUC have experienced denial of tenure at the same rate, but women are more likely to leave before the decision while men are more likely to leave after a favorable decision. For the combined groups, the distributions are significantly different at the two-tailed .05 level.

Tables 8 through 11 describe the associate professor cohort. Table 8 indicates that the number of women achieving full professor status is larger than the number of men, although the difference is not significant. The mean number of years at the associate professor rank favors men, at 6.83 vs. 7.10 years, but the difference does not approach significance.

Table 9 indicates that the number of men and women in this group who were still employed in tenured positions at UIUC in 1998-99 was nearly identical. (Only tenured positions were counted here in order to exclude as retained several individuals who were hired back onto non-tenure-track jobs after retirement.)

Tables 10 and 11 indicate the highest rank achieved by those who were still tenured at UIUC in 1998 (Table 10) and those who had left (Table 11). Of those who were still here in 1998, fewer women than men were full professors, but the difference was not significant. The reverse was true of those who had left: more women than men were full professors. While the overall loss is about the same for men and women, it was more common for men to leave as associate and for women to leave as full professors.

Follow-up Survey. Research. Table 12 displays the final tenure codes of male and female assistant professors who returned the follow-up survey, and indicate a significant difference in the distributions. Although very similar proportions of men and women left while on tenure code 1 through 5, it appears that more women left on terminal status while more men left after receiving tenure. Thus, one would expect the women on the average to report less positive experiences than do the men. Interpretation of the results must take this important gender difference into account. To control for the effect that failure to achieve tenure might have on the results, partial correlations between gender and other variables were calculated with this variable (1=final tenure of 6, T; 0=other) partial led out.

Table 13 describes the assistant professors' views of the research support they received. Although the findings indicate that all differences favor men (that is, men report greater support), most do not approach significance at the .05 level. The one exception is found in the "equipment support" category. The average level of support in the form of equipment reported by men is significantly higher than that reported by women. In an absolute sense, neither group reports strong support, as almost all means for men are close to "3", the midpoint between "none " and "all I needed", while the women's means are mainly a bit below _3". Perceptions of research support seem likely to be associated with one's final tenure status, either as a result of a real difference between support received by those that did and did not achieve tenure or possibly as a result of a tendency to blame the department for the failure to achieve tenure. However, the difference between men's and women's perceptions of their equipment support remained significant when termination was partialled out. In addition, the difference between the two groups in average rating of research support received came close to significance at p=.06.

As seen in Table 14, both men and women displayed a tendency to believe that research support was more readily available to tenured than to untenured faculty. Men generally thought support was more readily available than did women, but the difference was significant only in their answers concerning its availability to tenured faculty. This difference remained significant after terminal status was partialled out. The number of "don't know" responses was also tallied, in the belief that women might be less knowledgeable, on the average, than men concerning the way things work and what they might expect in the way of support. Women averaged more "don't know " responses than men, but the difference was not significant.

Table 15 details the respondents' perceptions of their relative standing within the department on research support, research productivity, final salary, and departmental evaluation of their research. For these scales, responses indicated the individual's perception of his or her position within the department, broken into top 20%, next 20%, middle 20%, etc. For example, a person who believed her productivity was in the top 20% would respond "5". Both groups viewed the research support they received as a bit below the midpoint for the department, based on means below 3.0 (where 3 is defined as support in the middle 20%). On the other hand, both groups saw their productivity as above the midpoint, with men reporting significantly higher self--perceived productivity than did women. Even though they had perceptions of above average productivity, both groups reported that their final salaries were below the middle 20%, with means of 1.7 and 2.2, respectively, for women and men. The question did not ask for a ranking of salary relative to others of the same rank, however, which could account for this result.

Asked to describe the department's evaluation of the individual's research endeavors compared to others, men described themselves as considerably more highly ranked than did women, with means of 4.0 and 2.8, respectively, a significant difference. Both of the significant differences between men and women (department evaluation and self-perception of productivity) remained after terminal status was partialled out. Thus, it seems that these male respondents saw themselves as more productive and as evaluated more highly on their research than did the female respondents, even after the difference in proportion on terminal status was taken into account. In addition, men gave slightly more "don't know" responses than did women, but the difference was not significant.

The findings reported in Table 15 suggested a comparison of individual's perceptions of their research productivity to their perception of their relative salary standing, on the one hand, and their perception of the department's evaluation of their research on the other. Difference scores were calculated for all individuals, and Table 16 portrays the mean of the difference between the individual's perceptions of his or her productivity and his or her relative salary, and the mean of the difference between self-perception of one's relative productivity and the department's evaluation of that productivity. Sex differences were negligible on either variable, so men and women were combined to see whether on the average they felt their relative productivity exceeded their relative salary, and/or whether their relative productivity exceeded the department's evaluation of it. The combined mean difference between individuals' perceptions of productivity and salary standing was 1.583, t=6.589, p<.001. For these 36 individuals, self-rating of productivity significantly exceeded salary standing. This could be a result of their junior standing in their departments, itself normally predictive of relatively low salary standing. The difference between self and department's evaluation of research did not differ from zero, with a mean of .321, t=1.302, p=.204. These men and women did not believe, on the average, that their departments under- (or over-) valued their research.

Table 17 shows that men and women did not differ significantly in their tendency to indicate that the department's evaluation of their research was a factor in their decision to leave UIUC. On the other hand, men (with a mean of 3.0) were neutral about whether the UJUC climate or work environment helped or hindered their productivity, while women (with a mean of 1.8) reported that it hindered them. This difference was significant, and the significance remained after terminal status was partialled out. Thus we see the suggestion of a less comfortable or supportive environment for women than for men.

Tables 18-22 summarize results for associate professors that parallel the assistant professor analyses of Tables 13-17. Table 18 indicates very small differences between men and women in their reports of research support received. In Table 19 we see that all differences in perceptions of research support availability favor men, including the number of "don't knows", which is lower for men than women. But again, none of the differences achieves significance at the two-tailed .05 level. In Table 20 there is no tendency for responses to favor either sex, and none of the differences is significant. This table indicates that unlike the assistant professors, both men and women saw the research support they received as above the midpoint for the department, based on means of 3.5 and 3.6 (where 3 is defined as support in the middle 20%). Similar to the assistant professors, the proportion of "don't know" responses is quite low, and although men used this response a bit more frequently than women, the difference was not significant.

Both gender groups saw their productivity as considerably above the midpoint, with women reporting slightly higher self-perceived productivity than did men at 4.5 and 4.4, respectively. Neither of these small differences is significant, of course. With well above average productivity, both groups nonetheless reported that their final salaries were only somewhat above the middle 20%, with means of 3.6 and 3.1 for women and men, respectively. This finding suggests the possibility that disgruntlement over salary may have played a role in some of their decisions to leave. Women report somewhat lower departmental- than self-evaluations of their research, at 3.9 and 4.5. For men, the two evaluations are essentially equal with a mean departmental evaluation of 4.5 and self-evaluation mean of 4.4.

As one would expect from the results in Table 20, Table 21 indicates that both men and women report sizeable differences between their self-evaluations of relative research productivity and final salary relative to others. Since the means of 1.0 for women and 1.4 for men are not significantly different, the data were pooled to test the hypothesis that the population value of this difference is zero. The combined mean difference was 1.156, t=5.3 57, p<.00l. There is a significant difference between these former associate professors' ranking of their own research productivity and their own salary, again suggesting a possible reason for their having left UIUC. Concerning the difference between their own and their departments' evaluations of their research, the men and women have different views. The women's mean for this difference is .7 and the men's is -.1. For each gender group, the hypothesis that the population mean for this difference is zero was tested. For women, t=3.571, p<.0l. For men, t=-1.468, p=.l64. Thus, is seems that for women (but not for men) the self-evaluation tends to be higher than the perceived departmental evaluation of research. This finding suggests an additional reason for the attrition of women from the ranks and one that does not seem to apply, on the average, to men. Interestingly, however, Table 22 demonstrates that neither men nor women rate the department's evaluation of their research productivity as a particularly important factor in leaving. The mean difference between women and men, 2.4 to 1.9, is not significant. Similarly, and unlike the assistant professors, both men and women were faintly positive about the UIUC climate/work environment, with means of 3.3 for both groups.

Qualitative results. Survey respondents were also asked "Are there other kinds of support you think your department should have provided to facilitate your productivity?" Thirteen women's and fourteen men's responses were categorized into: (a) requests for release time; (b) various direct supports such as equipment, cash, space, etc.; (c) collaboration, colleagueship, or mentoring; (d) administrative issues; and (e) positive statements that no other support should have been provided. Assistant professor results are listed in Table 23. Given the low number of responses in each category, it is difficult to discern patterns. However, it appears that responses for men and women are not very different. The administrative category included complaints that the department head was "unscrupulous and ambitious", that "honesty" was needed, and "Could have showed some interest in my existence--I was hired by a different administration than the one I worked under--new administration had different goals. I felt lost in the shuffle." Additional administrative issues included "I was given no feedback on or evaluation of my research productivity;" "Consistency in leadership/structure", and "mentoring, supportive, non-vindictive." Table 24 shows similar responses from the associate professors, but with fewer requests for more support, a few more positive statements that nothing else was needed, and no complaints about administrative issues.

Follow-up Survey. Teaching and Service. Tables 25-28 outline responses of assistant professors to questions about teaching loads and support and about service activities. As seen in Table 25, women's and men's reported teaching loads were virtually identical, at 3.2 and 3.1 courses per year, respectively. In addition, both groups reported a slight tendency for their loads to increase after the first year, while both rated their own loads as a bit above the middle 20% for their departments, at 3.6 and 3.5. None of these very small gender differences approached statistical significance.

Table 26 summarizes support provided for teaching, with all differences favoring men. Both gender groups rated support received for teaching at about the 3.0 midpoint of the scale or lower. But for TAs, secretarial help, equipment, and the average rating, women's means were below the midpoint while men's means were a bit above it. For graders and professional development, both groups' means were below the midpoint. The gender difference was significant for graders, with women averaging 1.2 and men averaging 2.7. This difference remained significant after terminal status was partialled out. For professional development, the gender difference was 1.9 to 2.5, which was not significant.

Table 27 indicates that for men but not women, support for teaching may have declined after the first few years, with means of 3.4 during the first years and 2.9 afterwards. For women, the means were the same for both periods of time (first few years and after the first few years), 2.6 and 2.6. Neither difference between men and women was significant; although once again, women rated availability of support during the first few years as below the scale midpoint (2.6), while men rated it a bit above the midpoint during that time period (3.4).

Table 28 shows that men and women were virtually identical in their reports of undergraduate advising loads, with both groups reporting they were in the middle 20% of the department at 2.9 and 3.1, respectively. The small gender difference was reversed for graduate advising, at 3.3 for men and 2.8 for women. Neither difference was significant. Finally, women reported a larger number of committee assignments per year at 4.6 compared to men's 4.0; again, this difference was not close to significant.

Tables 29 through 32 provide the same information for associate professors. No gender difference in Table 29 is significant. Unlike the assistant professors, women of higher rank indicate a somewhat higher course load than men, at 3.8 vs. 3.2 courses per year. There is also a tendency for women to indicate their course loads went down after the first year, and for men to indicate theirs went up. Finally, women rate their own teaching load as slightly above the midpoint of the scale, 3.3, while men rate theirs at 2.9. In contrast to the assistant professors (see Table 26), most of the very small differences between male and female associates favor women. Most means are close to 3.0, with professional development the only exception for both sexes. Women have a mean of 2.6 and men a mean of 2.3. No gender difference approaches significance.

Table 31 summarizes responses concerning availability of teaching support and indicates that neither gender group reported a change from the first few years at UIUC to later years at UIUC. On both variables, men and women had means close to 3.0, and the small gender differences were not significant. In Table 32, women's mean position with respect to relative undergraduate advising load is about average (3.2), and men_s is somewhat lower, at 2.5. Women's and men's means are virtually identical for graduate advising and are somewhat above the midpoint at 3.7 and 3.6, respectively. Committee assignments seem heavy, at 6.2 for women and 5.5 for men. None of these differences in reported load is significant, however.

Qualitative results. Table 33 summarizes the categories of responses former assistant and associate professors made to the question, "What other kinds of support should your department have provided for teaching?" Only three associate professors responded to this question, so their responses are summarized by parenthetical entries in the table containing the assistant professors' responses.

Requests for more thoughtful course or load assignments included complaints about: (a) extensive program responsibilities, (b) inadequate credit awarded for an intensive clinical training course, (c) suggestions that junior faculty should not be alone in teaching large undergraduate courses, (d) that junior faculty should be given a grad seminar course, and (e) that consistent assignment of courses to gain expertise rather than assignment of courses not selected by senior faculty would be helpful. Complaints about administrative support included: (a) that it was unavailable, (b) that the head was not interested in teaching, (c) that the department provided no help with the serious administrative details of organizing and supervising a particular course, and (e) that teaching was not valued as highly as scholarship and as such received minimal support or reward. Ideas about development of teaching included: (a) the suggestion of a mentoring program, (b) more sharing of ideas and experiences with colleagues, and (c) co-teaching with experienced senior faculty for new teachers.

Tables 34-37 present assistant professors' responses to questions concerning feedback, mentoring, and the work environment they experienced at UIUC. Although Table 34 indicates that about half of the women and two-thirds of the men had mentors, this difference is not significant. Both groups reported having an average of about one mentor while at UIUC. Table 35 compares experiences with annual reports and attendant feedback and indicates that no gender difference was significant at the .05 level. Most respondents believed tenured faculty were required to submit annual reports (58% of women and 67% of men), and even more believed that non-tenured faculty must do so (80% of women and 71% of men). More women than men (90% and 69%) who responded either "yes" or "no" believed the annual report was a part of their evaluation for merit raises. At the same time more women (50%) than men (30%) indicated that they did not know whether or not the annual report was used for this purpose. While about two-thirds of all assistant professor respondents said they usually received written feedback and 69% of women and 89% of men said they were provided oral feedback about the annual report, 17% of women and 11% of men said they usually received no feedback. Asked to rate the feedback in terms of usefulness, with 5 being "very useful" and 1 being "not at all useful," both men and women saw the feedback concerning the department's demands for tenure or promotion as rather modestly useful, with means of 2.4 for women and 2.7 for men. Similarly, women's mean rating of the usefulness of the feedback concerning the department's perception of their research activities was 2.6 compared to men's mean of 2.9. Both women and men reported receiving more feedback in their current job (5=much more, 1=much less) than at UIUC, with means of 4.0 and 3.8, respectively. Since their means did not differ, men's and women's responses were pooled to test whether the combined population means on these items were simply equal to 3, the midpoint of the scale. The combined means were 2.575 for usefulness of feedback in learning about departmental demands for tenure and promotion, 2.718 for usefulness of feedback in learning about the department's perception of one's research activities, and 3.900 for how much more (or less) feedback is received on the current job is than at UIUC. Only for the last of these did the population mean appear to be other than 3.0 (p<.001). The 95% confidence interval for the population mean for extent to which feedback on current job exceeds that of feedback at UIUC was 3.48 to 4.32. It seems that feedback on the current job is more plentiful than that received at UIUC.

Table 36 provides information about these faculty members' experiences with a third year or other formal pre-tenure review at UIUC. Most (76% of female and 58% of male assistant professors) had such a review. Most of those responding "yes" or "no" indicated they received written (100% for women, 75% for men) and oral (77% and 92%, respectively) feedback. The difference between men and women in the proportion receiving written feedback was significant at the .05 level, and this difference was still significant after terminal status had been partialled out. Thus, more women than men were given written feedback, and this difference was not due only to differences in treatment that could be explained by terminal status (e.g., the department needing to document the person's shortcomings in order to support a later termination seen as likely). No women indicated they received no feedback, although 18% of the men said they received none. Men's ratings of how well the feedback helped them to understand their progress toward promotion and tenure were lower than women's ratings, at 3.6 vs. 2.7, respectively (where 5=not at all well and 1=very well). This difference was close to significance, with p=.067. Similarly, women rated the review and feedback at the middle of the scale (3.0) in terms of how well it helped them understand what they needed to do to improve their chances for promotion and tenure, while men's ratings were less positive (3.5). This difference did not reach significance at the .05 level, however.

Table 37 shows that men were less hindered by family responsibilities than were women at 4.2 compared to 3.4 (with 1=greatly hindered and 5=not hindered at all). This difference was not significant, however. Asked whether there are ways UIUC should help lessen work-family conflict, 80% of women and 40% of men said yes. This difference approached significance with p=.074.

Tables 38-40 compare the responses of male and female associate professors on the same questions concerning mentoring, annual reports and feedback, and the work environment. No questions about the third year review were asked of this group. The proportions that had mentors while at UIUC were 76% of the female associate and 65% of the male associate professors. The percentages of associates who reported having mentors are similar to those of the assistant professors, at 52% and 67%, but if anything the female associates may have had a somewhat more positive experience in this regard than did the female assistant professors. Male associate professors reported somewhat more mentors, at 2.7 on the average compared to 1.7 for females, but this difference was not significant.

Table 39 details the associate professors' experiences with annual reports and feedback. There were no significant gender differences in any of these items. Most men (75%) and women (78%) reported that annual reports were required for tenured faculty and also for non-tenured faculty at 75% and 76%, respectively. Almost all of those responding _yes_ or _no_ indicated that the annual report was a part of their evaluation for merit raises, with 86% of men and 93% of women so reporting. Yet 29% of female associate professors and 18% of male associate professors indicated they didn't know whether or not it was used for this purpose. Of those responding _yes_ or _no__to questions regarding written and oral feedback, 43% of women and 70% of men indicated they usually received written feedback, while 57% and 55% indicated they usually received oral feedback. Another 25% of women and 17% of men indicated that no feedback was usually received. Both sexes found the feedback moderately useful with respect to demands for tenure or promotion, at 3.0 for women and 3.3 for men. It was a bit more useful with respect to the department's perceptions of the individual's research activities, as the mean for each group was 3.6. Both groups tended to receive more feedback on their current jobs than at UIUC, with means of 3.8 for women and 3.6 for men. Since the sexes did not differ, men's and women's responses were pooled to test whether the combined population means on these items were simply equal to 3, the midpoint of the scale. The combined means were 3.196 for usefulness of feedback in learning about departmental demands for tenure and promotion, 3.586 for usefulness of feedback in learning about the department's perception of one's research activities, and 3.700 for how much more feedback they received on the current job than received at UIUC. The population mean of usefulness of feedback for understanding promotion requirements was not significantly different from 3.0. However, its mean usefulness with respect to departmental perceptions of the person's research activities was significantly higher than 3.0, p<.05. Also, these individuals reported receiving significantly more feedback on the current job than at UIUC, p<.001. The 95% confidence interval for the population mean usefulness of feedback relative to one's research is 3.1 to 4.1. For quantity of feedback on current job versus their UIUC position, the comparable interval is 3.3 to 4.1.

Table 40 indicates that neither gender group was hindered very much by family responsibilities, with means of 4.0 and 4.1 for women and men, respectively. Forty-four percent of the female and 25% of the male associate professors indicated there are ways that UIUC should help lessen family-work conflicts. Neither difference is significant for the associate professors. It appears that the former assistant professors are more likely than the associates to believe there are ways UIUC should help lessen family work conflicts (at 80% and 40% of the women and men, respectively).

Qualitative results. Several open-ended questions related to feedback and mentoring and the work environment. All respondents were asked how their departments helped them understand the nature and extent of the requirements for tenure, promotion, and career success. Those who indicated they had a colleague or friend they could call a mentor were also asked to describe what these individuals did to foster their career development. Both assistant and associate professors were asked to comment on how the feedback they received could have been more useful, and assistant professors were asked how they thought their pre-tenure review and feedback could have been more useful. Finally, those assistant and associate professors who indicated there are ways that UIUC should help lessen work-family conflict were asked to comment further. Results are presented in Tables 41-49.

Tables 41 and 42 contain the categories into which responses of assistant and associate professors were grouped when asked about ways their departments helped them understand the requirements for promotion and tenure and career success. A third of the men and half of the women indicated that the department did not help them at all or did a poor job of helping them understand requirements. Several women indicated that, when given, such help was delayed in coming. Most respondents mentioned formal methods of communication, such as annual reviews, third year review, or conversations with an administrator (head or dean), while only a few mentioned informal means. Fewer associate professors indicated that they received no such information or that its quality was poor; however, the numbers were still not zero. Several pointed out they had been tenured, or virtually tenured, upon arrival. Unlike the assistant professors who spoke mainly of formal methods of communication, both male and female associate professors mentioned informal as well as formal means of communication. This was particularly the case for female associate professors. Perhaps, as a rule, informal networks are more open to more senior people and, as a result, assistant professors tend not to receive this sort of help. It is also possible that methods of communication have become more formal over time.

Several respondents, mainly assistant professors, mentioned problems associated with the communication of expectations. These problems included standards that varied within the college or between the department and the campus committee, and standards that shifted over time. For example, one person responded, "Didn't. Mostly by rumor. Constantly changed and was interpreted differently by different individuals or groups. Seen as mysterious and political." Several individuals commented that there were two sets of rules, the written ones and the unstated ones (e.g., "They made the written rules and tenure codes clear. The unstated rules and emphasis, however, were not really made clear--additionally I think they changed over time, and changed with department heads"). In contrast, among the assistant professors who seemed satisfied with their departments' communication of expectations, comments included the following: (a) "Department head sat me in his office and told me directly. There was no mystery about what the criteria were by which I would be judged"; (b) "Yearly feedback plus mid-tenure track review"; (c) "Meetings for new faculty to review P & T procedures"; (d) "1. Yearly personal meeting with department. head. 2. Yearly informational meeting conducted by dept. head and attended by all untenured (tenure-track) faculty"; (e) "Department sponsored _brown bag' lunches for untenured faculty. Individuals received support from their P & T advisors (an appointed committee of 2 tenured faculty). Yearly evaluation meetings with department chair." Given these comments, it appears that departments vary considerably in how they handle the communication of expectations and tenure requirements to new assistant professors.

Tables 43 and 44 provide a categorization of male and female assistant and associate professors' responses relative to what their mentors did to help them. Many respondents at both ranks wrote of general encouragement, support, or advice. Some respondents, especially associate professors indicated that they received "open support in the department". "Checking progress toward promotion and making sure that the requirements were understood" was mentioned by assistant professors, especially by men. Both assistant and associate professors mentioned "help making contacts". Other responses by both assistant and associate professors included reviewing manuscripts and collaborating on research.

Tables 45 and 46 contain comments suggesting ways that feedback on annual reports could be made more useful. Several associate professors indicated that the feedback received was fine; but several other associate professors, along with some female assistant professors, indicated it would have been more useful if there had been some feedback. Some assistant and associate professors, but particularly the assistants, described the feedback as too general, or as mechanical. Respondents also suggested that feedback could have been "more positive, less negative" and "the type that would help the individual develop their career." A few others commented that they received feedback primarily about their job but not their research (librarians), or that they received feedback as a faculty member but not as an administrator. Importantly, some assistant professors stated that feedback would be more useful if linked to promotion and tenure progress or standards, suggesting that for some it had not been. Finally, both assistant and associate professors' responses contained some mention of feedback perceived as dishonest or discrepant. These responses came primarily from associate professors and from women more than men. The following comments serve as examples of this point: "Even if more specific and honest, it probably would not have made a difference since all decisions were based on the personal agenda's (sic) of the power structure which consisted of one man and his chosen few." "After my tenure was denied I heard indirectly that some did not like my work--but I'd earlier been told that these same individuals thought highly of me..." The following are examples from associate professors: "The problem was an increasing gap between the implications of the official feedback (based on objective accomplishments) and the actual pattern of rewards and punishments (based on the chair's sense that you loved and appreciated him)." "The problem was mostly with the unofficial feedback. For example, grad students often told me that their advisors discouraged them from taking courses with me." "Feedback on how I fared compared to college-wide colleagues was through the grapevine rather than directly..." "Most details came from mentors. My performance was stated in a positive manner for salary or negative when I confronted an administrator who was very much out of line and close to racial harassment!"

Tables 47 and 48 present categorizations of the few responses that addressed the question pertaining to ways in which UIUC should help lessen work-family conflict. Among assistant professors, comments were made by both men and women, but among associate professors, it was mainly women who commented. General comments about the need to recognize and support family commitments were made by members of both ranks. "Day care" also was mentioned by respondents at both ranks. "Greater flexibility" was mentioned by several respondents. Fewer comments related to maternity/family leave, economic support, more TA/grader support, and the need for more spousal hiring. Several assistant professors felt expectations are too high. Finally, several associate professors made positive comments (i.e., praise for the spousal hiring program, "UIUC is not bad in this regard", and "My department was extremely supportive of family. We were all young, bearing children, etc. We enjoy birthday parties for colleagues' kids!").

Table 49 categorizes assistant professors' responses relative to how the pre-tenure review and feedback could have been more useful. The most frequent responses included: (a) that such a review and feedback would be useful if it had been done, (b) that an earlier review or more frequent reviews would be helpful, (c) more detail would help, or (d) greater honesty or a less political process would be useful. Representative comments from this last category include the following. "From the committee ... . You are doing extremely well; funding, teaching, research and publications. It is necessary that you get along with the HEAD of the departmentÉ". "The claim was frequently made to me the funding shouldn't matter, except in a peripheral way--however, it clearly is very important, and I might have been better off if someone had just said to me 'Look--the party line is that funding does not matter, but in reality, if you don't have cumulative funding in this general range, you probably won't get tenure.' I don't expect a department to say that on paper, but people informally told me the 'real' tenure expectation before I took my new job (outside UIUC) and I greatly appreciated it; "Not clouded by all the extraneous 'rumor-type' feedback that seemed to be in direct contradiction of it."

Tables 50 and 51 detail the means and standard deviations of male and female assistant professors' responses to the Riger, et al. (1997) questionnaire on climate. These items, as analyzed, are all keyed so that a high response (above the midpoint of 3.0) indicates the respondent sees the treatment of males and females as equal, and a low response indicates the individual disagrees that their treatment is equal. For example, one item reads "Male faculty tend to get more feedback about their performance than female faculty do." As originally keyed, the responses of 5 would have meant "strongly agree" that in this regard men and women are not treated equally. All such scales were switched in direction so that the response of _5" means strongly agree with an "equality" proposition.

On all items, the male assistant professors' mean is higher than the females' mean, and on all but one, this difference is significant. On most items, and on the mean of all items answered, the men's average is above the midpoint of 3.0 while the female's mean is below the midpoint. Thus, female and male assistant professors who responded are quite divergent in the degree to which they view treatment of the two groups as equal. On several items, the difference between the means is at or near two scale points: "In meetings, people pay just as much attention when female faculty speak," "Faculty are serious about treating male and female faculty equally," and "Most faculty would be as comfortable with a female chairperson as with a male chairperson." The average of all items answered was 3.9 for men and 2.5 for women.

In general differences between male and female associate professors were not as large, but for many items the male mean was similar to that for assistant professors while the female mean was somewhat higher for associates. All but three items yielded significant gender differences in response, with the differences in the same direction as those found between male and female assistant professors. The means of all items answered were 4.0 for men, and 3.1 for women.

Tables 52 and 53 display information about the current positions held by the assistant and associate professor groups. Of the assistant professors, one in five men and one in four women were no longer employed in higher education due either to retirement or to taking a job in another industry. The remainder continued as librarians or faculty members. While some of the associate professors had left higher education either for a job in another sector or by retiring, about half of each gender group had attained significant leadership positions in higher education. Respondents included eight deans or former deans, four holding chairs or named professorships, four department executive officers, and three head librarians. Among those still holding positions in higher education, there was no gender difference in the proportion who held traditional professorial positions (41.2% of men, 52.6% of women) when compared with these other positions of leadership (chi-squared=.472, p=.492).

Tables 54 and 55 display the proportions of male and female associate professors who indicated that their current job is tenured or tenure track. About two-thirds of male (66.7%) and female (72.2%) assistant professors were still tenured or on the tenure track, as were approximately three-quarters of former associate professors (76.5% of men and 81.8% of women). Neither of these gender differences was significant.

Tables 56 and 57 provide information about the institutions that currently employ these former UIUC assistant and associate professors. At the time of the survey most male (83.3%) and female (68.4%) assistant professors were employed at research universities, as opposed to teaching institutions or non-higher education organizations. The percentages of males and females employed by research universities did not differ significantly (chi-squared=1.321, p=.250). Of the associate professors, a most were employed at research universities (82% of women and 88% of men). This small gender difference was not tested for significance because of several small expected cell frequencies.

The survey offered five general types of reasons for leaving UIUC, and respondents were asked to comment about their applicability. The reasons provided were economic, family, location/weather, tenure and promotion requirements and "other". Although questions were designed to capture adverse comparisons between UIUC and the other institution, in a few cases, responses were clearly positive statements about UIUC. Tables 58 and 59 show how many indicated that a particular category was a factor in their leaving ("negative"), or for a few respondents, was a positive feature about UIUC (headed "positive"). Of the 21 women and 24 men who were assistant professors, at least half responded to each category as a factor. For men, family issues received a majority response (71%), with other categories receiving comment from about one-half of the men. For women, responses were more equally spread across the five categories, with family, location, tenure/promotion requirements, and "other" given as reasons for leaving by two-thirds to more than three-quarters of those responding.

Among associate professors, about two-thirds of the men commented about "economic issues", "location", and "other", while half to two-thirds of the women made comments about_these same three issues. Family was less an issue to many associate professors, with only 33% of women and 42% of men commenting about this category. Promotion requirements were important to only a few women (13%) and men (16%). Associate professors made more positive comments about UIUC than did assistant professors.

Tables 60 and 61 provide additional detail from those who indicated that economic issues had been a concern. Among the assistant professors, salary level was the primary concern, with 11 of 12 women and 9 of 12 men indicating it was an important reason for leaving UIUC. Still, several of these respondents, men as well as women, indicated it was not the primary reason. A few others mentioned benefits, small, compressed, or non-merit based increases, poor opportunity for advancement, poor spousal opportunities, poor research support, sex discrimination in salaries, and a small community. As with assistant professors, the associates were primarily concerned with salary level, as indicated by 14 of 16 women and 11 of 14 men. Other responses were fairly equally distributed among the same issues mentioned by assistant professors, with the addition of several comments on the negative effects of budget cuts, and the positive effect of a low cost of living in Champaign-Urbana.

Tables 62 and 63 present family issues in more detail. For male assistants, partner's job was the most frequently mentioned issue (10 of 17 men), while female assistants' responses were more diverse. Nevertheless, the most frequent response (6 of 17 women) was partner's job. "Partner located elsewhere" was an additional concern for both sexes, as were "extended family", "self or family wanting to live elsewhere", "seeing C-U as a poor community for kids", "singles' isolation", and "supports at UIUC". The two negative comments about support involved lack of on-campus child-care. The positive comment cited a rollback. Fewer associate professors mentioned family issues, and no single category received more than a few comments.

Tables 64 and 65 detail the geographic factors mentioned by assistant and associate professors as reasons for leaving. Men's responses were scattered over "remoteness", "distance from friends or family", "smallness of the town", "recreational or cultural issues", "scenery", "weather", and "unspecified related factors". For women assistants, weather was surprisingly important, with 9 of 15 responses mentioning it negatively, while 3 mentioned its positive aspects. "Scenery" and "distance from family or friends" also received comment, as did "recreational or cultural opportunities".

Among associate professors, one-third of the men mentioned the small town as a negative, as did one-quarter of the women. "Scenery" was more important to women than men, with seven women and no men mentioning it. Remaining responses from both groups were dispersed over "remoteness", "distance from family or friends", "unattractive aspects of the Midwest", "recreational or cultural opportunities", "weather", and "unspecified other factors".

Tables 66 and 67 contain more detailed responses that contributed to the decision to leave UIUC and that pertained to promotion and tenure issues. Six men and seven women (of 13 and 16, respectively) indicated that the stringent standards or the pace or pressure were an issue. Of these 13 comments, three explicitly stated that job performance rather than research should be the major criterion for librarians' promotion. Additional issues for men and women included unclear expectations and unfair, political, or inconsistent evaluation. A few women cited a lack of support from colleagues, and one man mentioned racism. The response rate was much lower at the associate professor level. Of the women, five of eight said they retained their rank in the move, which was recorded as a positive comment about UIUC. Several other women indicated they were promoted faster elsewhere, as did several men. Only one person cited unfair, political, or inconsistent evaluation, and two people indicated they gave up tenure to move. These latter were also recorded as positive comments about UIUC. In short, not very many of these associate professors had any particular impetus to move because of promotion issues.

Tables 68 and 69 present _other" reasons for leaving UIUC that were listed by the former assistant and associate professors. Half of the female assistant professors who responded indicated their colleagues' characteristics were involved, and about one-third cited administrative problems and climate issues. The first two categories received comment from several men as well. The following are several representative descriptions of difficulties with colleagues: "Greed and self promotion of senior people"; " É very difficult to work with, quite unprofessional, and prone to constant bickering"; and "department offered no collegiality and no mentoring." Most comments about administration involved the individual's department executive officer or boss, in the case of the Library. Typical climate issues included "large department where I didn't fit in"; "I never felt welcome"; and "I never felt like I was accepted." Among the female associate professors, "fit with research interest" or "departmental quality" was the issue most commonly cited, and for male associates "administrative" and "other" concerns were most common. Both men and women included positive comments, indicating that leaving UIUC was only to take an administrative advancement elsewhere. Finally, several people at each level complained of sexism, racism, or homophobia.

Tables 70 and 71 display responses concerning additional practices or characteristics of UIUC that enhanced professional development and productivity. Responses from both male and female assistant professors included a "positive research environment". In addition women also tended to praise specific research programs or support services, such as the Library and availability of funds for travel. This same pattern was evident in the responses by the associate professors as well.

Tables 72 and 73 present more detailed responses pertaining to "additional factors that hindered respondents' productivity at UIUC". Almost two-thirds of the female assistants responding to this item made negative comments about the administration, as did one-third of the males. About a quarter of each group commented about colleagues. About one-third of the female assistant professors felt they had lacked various supports for teaching or research, while a quarter of the men cited problems with workload. There were fewer negative factors cited by the associate professors, and several positively asserted that there had been no negative factors that hindered their productivity. The negative responses were scattered among the same issues cited by the assistant professors.

Tables 74 and 75 contain more detail pertaining to "other comments about experiences at UIUC". Male and female assistant professors cited both positive and negative experiences, and their responses ranged across topics similar to those already mentioned in previous questions. There were quite a few positive comments among associate professors. In fact, the men's remarks were almost all positive. Women, however, mentioned both positive and negative experiences, again, similar in content to earlier responses.

Discussion

The large increase in number of women assistant professors hired is a very positive sign of progress toward gender equity in hiring. From the 1970s to the 1980s the percentage of doctorates earned by women in the U.S. grew from 20.7% to 33.9% (Stephan and Kassis, 1997). Thus, campus hiring at UIUC has reflected the growth of women in the PhD pool. However, the small differences in tenure rates and in final rank attained by women hired as assistant professor suggest continuing difficulties in fully integrating women into the academic workplace of the Urbana campus. Although there has been a small difference in tenure rate favoring men that has been consistent over the two time periods studied, there has been no consistent gender difference in the retention of assistant professors. At the same time, there is a consistent tendency for males to achieve higher ranks than females who are hired initially as assistant professors. In contrast, there is no such tendency among the associate professors studied. The difference between the two professorial rank groups may be due to the fact that of those assistants that are not retained at UIUC, more women leave before getting tenure and promotion, while more of the men leave after attaining the rank of associate or full professor. We do not know how many of the unpromoted assistant professors left because they had been informed their chances for tenure were slim. If this is a primary reason for leaving before tenure is awarded, the small difference in tenure rate noted above could be an underestimate of the gender difference.

Retention of associate professors is about the same for men and women. However, of tenured professors who leave, more women than men are full professors. Of course, some of these losses, of both men and women, are to administrative opportunities elsewhere. Whatever the reasons for the loss, a tendency for female full professors to leave UIUC more frequently than male full professors would tend to lower the proportion of senior professors who are women compared to the proportion of female junior faculty hired.

The follow-up survey was designed to probe reasons for loss of faculty and to compare men's and women's perceptions of UIUC. The respondents are a self-selected sample, and there is no reason to extrapolate the results proportionally to all those who left. Comparing the final ranks of the assistant professor respondents to all those who left, it is clear that an almost identical proportion of female respondents received tenure (5%) as did females who left (7%), but for males, 32% of respondents had achieved tenure here compared to 14% of all those who left. The differences between male and female assistant professors' perceptions of UIUC could be colored by their difference in tenure status, which also reflects a higher rate of termination (i.e., final tenure code of 6 or T) of female compared to male respondents.

Be that as it may, male and female assistant professors did not differ much in their perceptions of support received for teaching or research, or in their teaching, advising or committee loads. Men judged their research productivity and its evaluation by the department to be higher than did the women assistant professors, which is consistent with their higher rate of achieving tenure. On the other hand, women reported that the UIUC climate or work environment hindered their productivity, while men, on the average, did not.

Half to two-thirds of these assistant professors reported they had had a mentor, and there were few differences between men and women in the kinds of annual report and third year review feedback they received. There were also few differences in their perceptions of its usefulness. Unfortunately, usefulness of feedback was not highly rated, and respondents reported receiving more feedback in their current jobs than at UIUC. A fair number of both gender groups described their department's help in understanding promotion and tenure requirements as nil or poor.

Differences were striking between male and female assistant professors' perceptions of the degree to which men and women are treated equally at UIUC. On all of the questions on the Riger et al. (1997) questionnaire, males' mean responses were higher than females', which indicates that men perceived a greater degree of equality. On several items, the male mean was above the midpoint of the scale, while the women's mean was below it, thus indicating that men on the average perceived a degree of equality, while women perceived a degree of inequality.

If there are generalizations that can appropriately be drawn from this small number of qualitative responses, they might be the following: (1) It appears that some of the women assistant professors, more than the men, never felt welcome at UIUC. Interestingly, however, some men as well as women who left reported cronyism, "old boy" power structures, and other such constellations that can make progress and productivity difficult to achieve for those outside the "in-group". (2) Positive aspects of UIUC are often nested within statements describing negative circumstances. For example, even in some comments regarding the rigors of tenure standards, one derives a sense of the excitement and a stimulus of the campus' focus on achievement. (3) A small number of faculty who leave UIUC look back with little pleasure at their experiences here. Serious problems were reported by more assistant than associate professors. These include allegations of unprofessional behavior on the part of colleagues or administrators, sexism, racism, and homophobia. These reports are by their very nature one-sided, and the self-selected nature of the group that responded more than likely provides a disproportionally high number of such responses. Still, it is discomforting to note that some former UIUC employees feel justified in allegations of sexism, racism, homophobia, or simply self-serving and unprofessional behavior on the part of their colleagues and administrators. Although such comments were few in number, their very nature underscores the continual need for cultural diversity, fairness in procedures, care for development of junior colleagues, and perhaps even greater concern for the climate of the work place. The issue of climate applies to men and women and can affect productivity of both.

The associate professor respondents appeared to be a "happier" lot. Of course, all had been promoted and none had been terminated, so they had "made it" within the UIUC system. Still the fact that almost half of the women not retained had been full professors suggests that the follow-up survey might be helpful in identifying some of the factors involved. Male and female associate professors were quite similar to the assistant professors in their reports pertaining to the support they received for research and teaching and their teaching, advising, and committee loads. However, a gender difference did emerge in responses to questions about individual research productivity and the department's evaluation of that research. Both men and women rated their own productivity as well above average. Women, but not men, reported a gap between their own productivity and the department's evaluation of their research.

Both women and men reported having had mentors at UIUC, and there was little difference in their remarks regarding annual report requirements and feedback. Unlike the assistant professors, they tended to find the feedback they received as useful relative to the department's perception of their research. At the same time, however, they were similar to the assistant professors in reporting that they receive more feedback in their current positions. In addition, a few female associate professors spoke of discrepant, dishonest, or unknowledgeable feedback. These associate professors were employed at UIUC in an earlier era, and so these comments may not be applicable to current conditions. On the other hand, similar remarks were made by some recent assistant professors, suggesting that the issue of feedback is one on which this campus could improve.

Gender differences in response to the Riger et al. (1997) questionnaire were not as stark among associate professors as they were among assistant professors. On most items men endorsed an average response suggesting more equal treatment of men and women than did women. On most items, women tended toward answers near the midpoint of the scale, indicating neutrality, while men more actively endorsed a position suggesting a degree of equality. Thus, it seems that these women who left did not, on the average, view the treatment of men and women as strikingly unequal; however, they also tended not to view it as equal.

Similar to the assistant professors, associates who left tended to comment favorably on the research community at UIUC. Unlike the younger group, they offered few descriptions of problems with colleagues or the administration. For both men and women, the additional comments about their experiences at UIUC tended to be positive. Why then, did these senior professors leave? Many commented on economic factors, especially salary level, as did the assistant professor group. Some men and women left for advancement, and women, especially, tended to leave for a department that better suited their research interests.

An additional factor that could account for the loss of senior women might be early retirement. The 1993 National Study of Post-secondary Faculty found that new faculty in higher education in Fall 1992 were about 42 years old, with no noticeable difference for men (42.3) and women (41.8) (Finkelstein, Seal, and Schuster, 1998). On the other hand, results from the same survey indicated that more full-time male than female higher education faculty intended to work until at least age 65 before retiring (57% vs. 44%, respectively) (Chronister, Baldwin, and Conley, 1997). This difference could be a factor influencing the lower proportion of women found at senior levels, compared to the proportion found at lower levels. In short, if men work more years than women, proportions of women would tend to be smaller at higher professorial ranks.

Most assistant and associate professors who left UIUC found tenured or tenure-track positions at other research universities. Among the associates who remained in higher education, about half of the men and women attained positions of leadership such as department executive officer or dean. In this sense, perhaps UIUC served most of these former faculty well.

In sum, the campus seems to lose a few more women than men prior to tenure and also after the rank of full professor is achieved. Women report a "chillier climate", at least as measured by the Riger et al. (1997) questionnaire. Some female assistant professors feel unwelcome, and some of their female colleagues at higher rank leave for a "better fit". These descriptions suggest that retention of female faculty by campus units may require a greater concern and effort toward developing a better working climate.


References

Bentley, R. J., & Blackburn, R. 1. (1992). Two decades of gains for female faculty? Teachers College Record, 93(4), 697-709.

Chronister, J. L., Baldwin, R. 0, & Conley, V. R. (1997). Retirement and other departure plans of instructional faculty and staff in higher education institutions. U.S. Department of Education, National Center for Education Statistics, NCES 98-254.

Cole, J. R. (1979). Fair science: Women in the scientific community. New York: Macmillan.

Finkelstein, M. J., Seal, R., & Schuster, J. H. (1998). New entrants to the fill-time faculty of higher education institutions. U. S. Department of Education, National Center for Education Statistics, NCES 98-252.

Loeb, J. W. (1991). Memorandum to Morton W. Weir, September 25, 1991. (Internal UIUC document)

Long, J. S., Allison, P. D., & McGinnis, R. (1993). Rank advancement in academic careers: Sex differences and the effects of productivity. American Sociological Review, 58(October), 703-722.

Riger, S., Stokes, J., Raja, S., & Sullivan, M. (1997). Measuring perceptions of the work environment for female faculty. Review of Higher Education, 21, 63-78.

Sonnert, G. with the assistance of Holton, 0. (1995). Gender differences in science careers: the project access study New Brunswick, N.J.: Rutgers University Press.

Stephan, P. E. & Kassis, M. M. (1997) The history of women and couples in academe. In M. A. Ferber & J. W. Loeb (Eds), Academic Couples: Problems and Promises (pp. 44-79). Urbana, Illinois: University of Illinois.

Stokes, J., Riger, S., & Sullivan, J. (1995) Measuring perceptions of the working environment for women in corporate settings. Psychology of Women Quarterly, 19, 533-549.