Greater federal support and new methodology for gauging higher education R&D results in major gains for sector

Statistics Canada has implemented the first phase of a major overhaul of the method used for gauging R&D spending in the higher education sector, with its latest survey capturing a much higher level of expenditures than ever before. Data from the new survey shows that higher education institutions performed $3.963 billion in R&D for FY98-99. The amount is relatively equal to the total for FY97-98, but more than 20% higher than the amount captured before the changes were implemented R&D spending of $3.277 billion was reported.

The comparison for FY97-98 shows that R&D expenditures for that period rose $663.1 million or 20.3% after the new methodology was implemented, and historical values could change again, once further methodological enhancements are introduced. A working paper detailing the revisions will be published later this month.

The increased expenditures contained in the new data come largely from previously uncaptured outlays by the universities themselves, and are also the result of increased federal spending in the university research arena, pushing levels back to where they were in the mid-1990s.

Engineering and the natural sciences continue to command the lion's share of higher education R&D, with $3.207 billion in spending for FY98-99. Of that amount, health sciences R&D takes a 47.4% share with expenditures of $1.519 billion. R&D spending in the social sciences and humanities (representing the largest number of academics in the country), is well behind with $756 million or a 19.1% share, down from a 23.2% share in FY97-98. Indeed, the largest gains as a result of StatsCan's new estimation procedures are in the health science and other natural sciences, suggesting that much of the new "found" money is related to medical research. StatsCan began to incorporate R&D from hospitals affiliated with Canadian universities into its surveys last year.

As expected, Ontario and Quebec are home to the largest R&D expenditures, accounting for $2.723 billion or 68.7% of the national total. But Nova Scotia is displaying surprising strength, moving to fifth spot overall compared to seventh last year (R$, December 8/99). Expenditures in the other Atlantic provinces are also up sharply, with PEI's tally doubling year over year.

"Previous surveys missed 20% of higher education research activity. The quality of the estimate has been troubling people for a long time," says Paul McPhie, assistant director of StatsCan's science, innovation and electronic information division. "The most complex part of the work is to test the assumptions and make the appropriate changes."

The new methodology for R&D spending in higher education and health arose from recommendations made by a working group which developed a framework for improving estimations.

The working group worked closely with several organizations including StatsCan's Centre for Education Statistics (CES), the Association of Universities and Colleges of Canada (AUCC) and the Canadian Association of University Business Officers (CAUBO). The higher education survey draws upon CES and CAUBO data, and the latter is being encouraged to improve the reporting by universities of financial information. The working group is also encouraging AUCC and CAUBO to work together to make a distinction between research grants and research contracts, but CAUBO says the task would be too onerous and it isn't convinced there's a good enough reason to initiate such a change.

"Those who want the data have not convinced us of the validity of why they want it," says CAUBO executive director Maurice Cohen. "Besides, the information is in the financial systems of universities but it's an enormous amount of work to get at the distinction."

The change that has already pushed up the values of higher education research is the inclusion of indirect costs associated with sponsored and non-sponsored research, such as instruction, physical plant, computers, administration and student services. Prior to the change, StatsCan simply subtracted sponsored research funding from the total, resulting in hundreds of millions of dollars in indirect costs being missed.

The working group's recommendations are broken down into six areas, and are estimated to cost $350,000 in cash and in-kind to implement over three years. StatsCan is contributing $100,000 and McPhie says Industry Canada has also contributed funding, as has the Canadian Institutes of Health Research. Discussions are also ongoing with the two other granting agencies, which may result in a time-use survey of university faculty by field of science.

The working group also recommended that StatsCan begin collecting R&D conducted by community colleges to be included with higher education. If implemented, that task would fall to StatsCan's CES.

The Association of Canadian Community Colleges will hold a symposium on research in Ottawa on February 23-24.

R$