Canada’s universities and the loss of UCASS data: Scrambling for an alternative

In early 2005, the Ryerson University Faculty Association (RFA) was in the last stages of a massive contract arbitration. Faculty salaries were the last major issue to be resolved. The stakes were high— faculty had already gone two years without a contract and two more years of increases were to be determined. The arbitration award would dictate salary increases for four years, accounting for more than ten per cent of a typical faculty member’s career. The administration argued that Ryerson faculty were already well compensated and earned substantially more than similarly ranked faculty at other Ontario universities. The faculty association argued that the administration’s salary comparisons were flawed and should not be used as the basis for the arbitration award. The faculty association also argued that controlling only for rank—as the administration had done— was not sufficient to produce good salary comparisons. Faculty salaries increase with experience, even after controlling for rank, so one must control simultaneously for both rank and experience to produce valid comparisons. The salary comparisons at Ryerson were also complicated by several different faculty career paths, a consequence of Ryerson’s transition from a polytechnic institution to a university. The faculty association contended that Ryerson faculty were actually earning substantially less than comparable faculty at other Ontario universities after controlling for the appropriate factors.

What is important here is not the outcome, but the nature of the process that underscored the RFA arbitration. This process was able to function because both sides had access to data on faculty salaries and other faculty characteristics that they used to make reasoned, evidence-based arguments about the salary increases that faculty should or should not receive. This kind of process, however, will be much more difficult to follow in the future. In early 2012, Statistics Canada announced that it had discontinued the University and College Academic Staff System (UCASS) survey, which up until then was the most accurate and comprehensive set of data on faculty in Canada. Both the Ryerson administration and faculty association had used the UCASS data to make their arguments in 2005. Future negotiations will no longer be able to rely on this crucial data set. The last round of data released from the UCASS survey covered the 2010-11 academic year and are therefore already out of date for any current or future negotiations.

UCASS was an incredibly important resource for the entire university sector, providing reliable and comprehensive data that made crucial research and analysis possible. The UCASS survey required all Canadian universities to submit data on their full time faculty by October of every year. The institutions were required to submit, among other things, data on the age, rank, gender, subject taught, year of appointment, and years since last promotion for every full-time faculty member. It also included information on faculty salaries and any administrative stipends paid. Statistics Canada worked with the universities to reconcile differences in definitions or faculty classifications across universities and resolve any other difficulties that might compromise the quality of the data. The result was a dataset that could be used to produce meaningful profiles of faculty characteristics and salaries that allowed for useful comparisons across schools. Perhaps most importantly, the fact that the survey was conducted by Statistics Canada provided two important benefits. First, it meant that universities were legally required to submit accurate and complete faculty data. Second, it meant that the output was seen as impartial by both faculty and administrators because the data were collected by a neutral third party.

While the data collected by Statistics Canada for the UCASS survey were indeed crucially important and used extensively, in practice the system was not perfect. For example, many universities failed to meet the October submission deadline, often by several months, which meant their data were unavailable and any attempted salary comparisons suffered as a result. Further, the official Statistics Canada hard-copy publication of the data, Salaries and Salary Scales of Full-time Teaching Staff at Canadian Universities was not itself terribly helpful for collective bargaining or other research on faculty. Average and median faculty salaries were reported by rank, but there was no information on levels of experience or any other faculty characteristics, and the publication included a confusing set of categories and exclusions. In practice, faculty associations needed to work with the raw dataset; something that was not available publicly and often exceeded the in-house capabilities of faculty association bargaining teams.

FeliceMartinelloThe Canadian Association of University Teachers (CAUT) and the Ontario Confederation of University Faculty Associations (OCUFA), however, filled in some of the gaps left by the Salary Scales publication and provided useful data for collective bargaining. Both organizations obtained custom compilations of the Statistics Canada data and made them available to member faculty associations. The CAUT compilation included data on faculty age, rank, responsibilities, and average and median salaries for Canadian universities. The OCUFA compilation reported similar breakdowns of the data but only for faculty at Ontario universities. More recent OCUFA compilations also included tables covering subject taught, promotions, and detailed measures of experience. Both the OCUFA and CAUT compilations were aggregated and randomized to maintain the confidentiality of faculty. This made analysis of the data more challenging, but it was still possible to extract good information from their data tables.

University administrations would also obtain their own custom compilations of the UCASS data from Statistics Canada. Unfortunately, their compilations often specified different parameters, inclusions, and exclusions than either the OCUFA or CAUT releases. The differences between the compilations, combined with the randomization and aggregation of the OCUFA and CAUT runs, often led to discrepancies about salary comparisons. But, despite these differences, the data provided an important, mutually agreed-upon factual basis that was used to resolve conflicts at the bargaining table.

Beyond the context of collective bargaining, the UCASS data were also used by a variety of researchers interested in Canadian higher education. Researchers can access the raw, unaggregated data on individual faculty members, for the years that the survey was conducted, by submitting specific project proposals to Statistics Canada. If the project is approved, researchers can only access the data from specific Statistics Canada sites and all output has to be vetted by Statistics Canada staff to maintain confidentiality. The unaggregated data on individual faculty members allow more detailed and complex examinations of all aspects of faculty covered by the dataset. For example, data on the age profiles of faculty have been used to predict faculty retirements. The retirement predictions are then paired with future enrolment projections to consider faculty requirements and future hiring needs. Other work has used the UCASS data to estimate the effects of the changes in mandatory retirement laws on faculty retirement decisions.

The data also allow a close examination of the status of women within academe. Several papers compare the salaries of men and women faculty and track the changes in the salary differential over time. See , for example, Warman, Woolley and Worswick’s 2010 paper on the evolution of gendered pay differentials at Canadian universities. Another line of inquiry compares differences in time to promotion for men and women faculty, as in Stewart, Ornstein and Drakich (2009). My own work used the UCASS data from OCUFA and CAUT to look at the relationship between university revenues and faculty salaries and the effects of unionization versus special plan arrangements in Ontario.

Now that the UCASS survey has been cancelled, however, we will no longer be able to track the progress of women academics, predict faculty age distributions and retirements, or examine the composition of faculty and faculty salaries by subject, age, gender, or rank. It will be much more difficult for faculty associations to identify groups within their memberships who require extra attention (e.g. junior faculty) and to present compelling arguments, both to the university administrations and to the other faculty association members, that special provisions are required to help them (e.g. lump sum versus percentage salary increases). More importantly for collective bargaining, there will no longer be a mutually agreed-upon pool of information about faculty characteristics and salaries that may help identify settlements that are acceptable both to university administrations and faculty associations. University administrations will surely argue that faculty are already well compensated compared to faculty at other schools. Faculty associations will likewise argue that faculty should receive salary increases. Without the data, however, the arguments will be far less credible. Neither group will be able to convince the other side without a comprehensive and credible source of evidence. As a result, it will be much more difficult to make progress in collective bargaining.

Without the UCASS data set, there are no other comprehensive sources of data on faculty in Canada. For Ontario universities and faculty associations, salaries published under the Public Sector Salary Disclosure Act provide a starting point for faculty salary data, but this source has significant limitations. The names and salaries of all broader public sector employees—including those at universities— who earn $100,000 or more are posted on a provincial government website commonly known as the Sunshine List. The data can be downloaded and parsed to make salary comparisons, but it requires a significant amount of effort. Some researchers have used data from the Sunshine List to study trends in administrators’ salaries; determinants of faculty salaries and their relation to the migration of faculty; and whether salary disclosure discouraged the growth of senior administrators’ salaries, or encouraged their growth by enabling more comparisons across institutions. The fundamental difficulty is, of course, that the disclosure only reports salary data for faculty who earn $100,000 or more, which leaves many faculty uncounted and unavailable for analysis. Researchers have struggled with this limitation. Given current salaries, nearly half of Ontario’s faculty are excluded. This is particularly true for sessional or contract faculty. For collective bargaining purposes, it is possible to make an assumption about the distribution of salaries based on earlier data and impute the missing salaries. But I doubt that the results would be convincing at the bargaining table.

A preferable option would be for all of the interested parties—primarily faculty associations and administrators across Canada—to get together and create a dataset of faculty characteristics and salaries that is roughly equivalent to UCASS. There are clearly benefits for both university administrations and faculty associations and, in theory, it should not be difficult to do. All universities use the same information on faculty members for their everyday human resources responsibilities, so the data are already easily accessible to universities and in common formats. In fact, most administrations already provide copies of these data to their faculty associations on a regular basis. Confidentiality would not be a serious obstacle since, again, the data are already provided to faculty associations and more than half of Ontario faculty already have their salaries published under the Public Sector Salary Disclosure Act. The data could easily be aggregated in a manner similar to Statistics Canada’s process, to preserve confidentiality if or when it is required. Aggregated data does make analysis more difficult since it contains less information than the raw data, but it would still be very useful to university administrations and faculty associations for salary comparisons and other analyses.

Data on faculty characteristics and salaries will only be useful if both university administrations and faculty associations believe it is credible and trustworthy. With Statistics Canada no longer collecting the data, both university administrations and faculty associations need to be included as equal partners in the creation, maintenance, and dissemination of the data set in order to promote confidence and trust. In Ontario, a partnership between the Council of Ontario Universities (COU), which represents university administrations, and OCUFA, which represents faculty associations, could be established to collect, process, and disseminate data. Given the existing institutional structures and functions it would not be difficult nor expensive. It would also create benefits at the bargaining table, and in our overall ability to understand higher education in Ontario.

In a better world, Statistics Canada would be given the mandate and funding required to continue conducting the UCASS data survey, the long-form census, and the other discontinued surveys that will be sorely missed. UCASS provided the best vehicle for impartial, high-quality data on faculty. In its absence, faculty and administrators are left to piece together a second-best solution.

Felice Martinello is a Professor of Economics at Brock University.



Stewart, P., Ornstein, M., Drakich, J. (2009). Gender and Promotion at Canadian Universities. Canadian Review of Sociology  46 (1),  59–85.

Warman, C., Woolley, F., & Worswick, C. (2010). The evolution of male-female earnings differentials in Canadian universities,1970–2001. Canadian Journal of Economics  43(1), 347–72.