- Haignere, Lois; Lin, Yanging. (1994). Systemic Gender Differences in Academic Salaries at the University of Manitoba. United University Professions.
- Brown, L. K., Troutt, E. and Prentice, S.. Ten Years After: Sex and Salaries at a Canadian University. Canadian Public Policy. 2011: Ten Years After: Sex and Salaries at a Canadian University | Canadian Public Policy (utpjournals.press) .
- Brown, Laura K.; et al. Regression and selected decomposition results for a 3-period sex and salaries analysis. 2016: salaryregressions.pdf (umanitoba.ca) .
- Brown, L.K., Troutt, E.. Sex and Salaries at a Canadian University: The Song Remains the Same or the Times They Are a Changin'? Canadian Public Policy Vol. 3, Issue 3. 2017: Sex and Salaries at a Canadian University: The Song Remains the Same or the Times They Are a Changin'? | Canadian Public Policy (utpjournals.press) .
- Schirle, Tammy. Study of Gender-based Salary Differentials at the University of Manitoba. January 30, 2019: Microsoft Word - UM-Equity-Jan2019.docx (umanitoba.ca) .
The University of Manitoba (UofM) and the UofM faculty association (UMFA) had conducted an initial analysis and recommended further study of equity in gender pay. The 1994 report is a commissioned report to undergo said analysis. The 2011 report is a follow to the 1994 study, and indicates that as a result of the 1994 study, a monetary payout to female staff was implemented. The goal of the 2011 study is to see whether the pay gap persisted. Further, all of the latter studies are conducted by the same researchers involved in the 2011 study. Thus, there is a great deal of continuity in the reports as time progresses. The 2016 study includes a history of equity studies at the UofM that dates back to 1974. The 2016 study also indicates that the UofM study is the longest running study of the equity pay gap in Canada, spanning 20 years. However, the 2016 study only reports the results of the regressions and does not contribute any analysis. The 2017 report is the analysis of the regressions reported in the 2016 report. These two reports can be considered the same study.
A database of 1270 full-time academic staff was used for the 1994 study, obtained from the faculty association. The data used for the 2011 study are essentially the same as the 1994 study, and a 2003 version of the same data. The datasets include all teaching faculty and librarians. The data exclude administrators above the level of department head, and include faculty on leave, except members on long-term disability leaves or long-term leaves without pay. The 2017 study uses data from of 1,215, 1,099, and 1,180 teaching staff in 1993, 2003 and 2013, respectively.
Multiple regression was used in the 1994 study. The logarithm of annual salary was regressed against gender, rank, time since degree, time since promotion, time since first hire, educational attainment, appointment type, and faculty. Additional analysis was conducted on the rank variable to see if there was bias in the rank variable, odds ratios were calculated. A similar analysis to Blinder-Oaxaca decomposition was also conducted, however the study did not name the method as such.
The 2011 study is meant to be a replication of the 1994 study. Largely the same multiple regression was conducted, however minor tweaks were made to conform to conform to more recent studies of pay equity. The year in rank variable was not available in the 2003 data set and was dropped from the regression. A Blinder-Oaxaca decomposition was also produced. In addition, a Wellington-Blinder-Oaxaca decomposition is also calculated, which is an analysis of the change in the Blinder-Oaxaca coefficients over time, it was used to see if the bias found in 1994 had been reduced by 2003. The 2017 study omits the multiple linear regression, but uses the Blinder-Oaxaca and the Wellington-Blinder-Oaxaca to estimate the gender pay gap as well as the change in coefficients from one time period to the next, i.e. the changing nature of gender pay differences.
The 1994 study estimated a $2,005 average difference favouring men using the multiple regression. The Oaxaca-Blinder decomposition returned a result of a $1,856 lower annual salary for women. Further, the results of odd-ratio test on ranks indicate that women were less likely to attain a higher rank than men at each stage of advancement.
The 2011 paper found that the gap in gender pay remained largely unchanged in magnitude. The study found that women’s overrepresentation at lower-paying ranks, underrepresentation at the higher ranks, and differences in highest degree and experience attained explain much of the wage gap. This is a slight difference from the 1994 study, which indicates that the source of gender pay-equity evolved over time.
The main findings of the final study were that when experience, faculty, type of appointment, and rank, there was no significant difference in gender pay in the 2017 study, which contrasts this study with the two prior studies. The main salary differences in aggregate appear to be driven by the rank variable which suggests that women still face some bias with respect to career advancement.
In a 2019 analysis, the researcher concluded that women tend to dominate in lower ranks, have less seniority, and tend to segregate into different disciplines than men, but that there is no evidence of a systemic gender-based salary differential once key salary determinants are controlled for. The University of Manitoba Faculty Association made a series of recommendations on this report to address issues seen related to employment equity including “The slower rate of promotion to Professor and the lower likelihood of women holding the rank of Professor are likely to generate lower career life-time earnings for women. This would be reflected in annual pension contributions and income in retirement. It would be useful to investigate lifetime earnings and pension accounts to assess if this hypothesis is correct.”