SOCIAL SCIENCE RESEARCH IN CANADA AND FEDERAL GOVERNMENT INFORMATION POLICY: THE CASE OF STATISTICS CANADA
Kirsti Elizabeth Nilsen
Dissertation
Doctor of Philosophy, 1997
Faculty of Information Studies
University of Toronto
EXECUTIVE SUMMARY
The broad objective of this research was to examine the extent to which government information policy has an effect on the use and the users of government information. Canadian federal government-wide policies of restraint and cost-recovery of the mid-1980s applied to many government activities including government information production and dissemination. The effects of these de facto information policies on access to government information in Canada were studied focusing on a specific body of government information, that is, statistics produced by Statistics Canada. The research first examined Statistics Canada's response to, and implementation of, the cost-recovery and restraint initiatives of the 1984 to 1986 period, specifically the commercialization of its output. Secondly, the effects of the resulting price increases and format changes on social science research and researchers were investigated. The specific objective was to identify the extent and nature of use of Canadian and other statistics sources by researchers publishing in Canadian social science research journals and any changes over time in that use.
Two policy effects are investigated here. Firstly, the impact on access to Statistics Canada information is considered, in terms of pricing of its products, and the formats in which the information is made available, i.e., the extent to which information is disseminated in electronic formats and is no longer made available in paper. Secondly, impact of the pricing and format changes on use and users is empirically investigated. The research addressed three significant government information policy issues: information technology, economics of government information, public access to and availability of government information.
In the late 1980s, a number of media stories covered Statistics Canada's price increases and the implications for social science researchers and, ultimately, for society as a whole. Academics expressed concerns about price increases and limitation on access. They argued that academics would not be able to afford to do basic research on Canadians and that they would have to buy cheaper U.S. information and divide by ten. This researcher was motivated to explore the effects of these policies on the use of Statistics Canada information by social science researchers, and to determine the extent to which they were driven to change the statistics sources they use to describe the Canadian scene.
A triangulated study was designed to examine the effects of the federal government restraint and cost-recovery initiatives of the mid-1980s. Case study methodology was used to investigate Statistics Canada's response to the federal restraint and cost-recovery initiatives of the mid-1980s. Documentation from Statistics Canada and other government sources was examined, along with non-governmental analyses, and agency officials were interviewed.
Bibliometric methods were used to objectively document policy effects on social science researchers' use of statistical sources. The analysis focused on a sample of articles published from 1982 to 1993 in 21 Canadian social science research journals in Economics, Education, Geography, Political Science and Sociology. Examination of the citations, tables, and text in the sampled articles revealed the extent of use of statistics from Statistics Canada and other governmental and nongovernmental sources, both Canadian and foreign, over a period before and after policy implementation. Time and discipline were the independent variables against which the use of various statistics sources were measured. Finally, a survey of authors of sampled articles supplements the bibliometric findings and provides more subjective data. A questionnaire was sent to 163 Canadian social science academic researchers identified through the bibliometric analysis as having written articles with a Canadian focus or setting in which published statistics were used. Ninety-seven responded, a return rate of 59.5%.
Results of the case study show that Statistics Canada sought to recover costs and achieve greater revenues through higher prices and increasing electronic data dissemination. Implementation of the cost-recovery initiatives resulted immediately in sharply higher prices for the Agency's information products and encouraged a shift to dissemination of information in electronic formats with further cost-recovery in mind. The case study illustrates price increases for print products averaging 267% in 1985, followed by more gradual increases, leading to a total increase between 1982 and 1993 of 438%. Prices for electronic microdata files were less dramatic, rising by 170% between 1982 and 1993, however, the amount of data per file decreased. Some census computer readable formats have been documented as increasing by 9900%. The case study also revealed that some information formerly available in paper became accessible only in electronic formats at much increased prices. Paper products were increasingly designed for generalists. Those with more detailed or specialized needs were expected to pay for electronic access. The Agency argued that it charges dissemination costs only, but there is evidence that supports the conclusion that Statistics Canada is assigning costs to dissemination which belong elsewhere.
360 articles were sampled in the bibliometric research, from a population of 5,414 research articles, Identifiable published statistics were used in 66% of the articles, 19% made no use of statistics, 11% used only self-collected data derived by the author from experiments or other research methods, and 4% used only unidentifiable published statistics. A subset of the sample was identified consisting of 207 articles (57.5% of the sample) which were written with a Canadian focus or setting and which used published statistics. The data provided below refers to these 207 articles.
The bibliometric analysis showed that Statistics Canada was the single most frequently used publisher of statistics, but it was used in only 41% of the 207 articles written with a Canadian focus or setting. The same percentage used all other Canadian federal government sources combined. Provincial government sources were used in 25% of the articles. Nongovernmental sources (e.g. business, foundations, universities, and published books and articles) were used in 71% of the articles. United States government statistics sources were used in only 10% of the articles and American nongovernmental statistics were used in 32% of the articles. Foreign titles were used mainly for comparative purposes and were seldom the major source of statistics on which conclusions were based. The bibliometric data indicate that social scientists range widely to obtain the statistics they use.
There was no significant change over time in use of statistics from any source.
There was a statistically significant decline over the time period in the number of articles using Statistics Canada's paper formats, but no significant corresponding increase in the number of articles using Statistics Canada computer readable products. As noted, the number of paper formats available from the Agency has declined.
The discipline of the journal in which an article is published was found to be a frequent statistically significant predictor of use of statistics in general, and of use of particular sources. Articles published in Geography and Economics journals made greatest use of Statistics Canada's statistics, and those in Political Science journals made least use of them.
Survey respondents were active academic researchers, almost 75% of them spend more than 40% of their time doing research, and they were heavy users of statistics. While most received their highest degrees in Canada, 37% received their highest degree in the United States. Most had used Statistics Canada as a statistics source at some time, but only 41.5% used the Agency's statistics more than 50% of the time. The survey revealed that social science researchers are unhappy with the price increases, but have not changed the statistics sources they use as a result. The movement of statistical information into electronic formats is well received, though more respondents (in 1995) still used paper products more than electronic ones.
Of the 58 respondents with an opinion on the subject, 71% disagreed or strongly disagreed with the statement, "Statistics Canada prices are reasonable." About 20% of all respondents indicated that prices charged had adversely affected their research, was forcing them to use less of the Statistics Canada data, or that price changes had affected the allocation of available funds. Of the 54 respondents with an opinion on the subject, 52% agreed or strongly agreed, with the statement that Canadian researchers in their discipline were using non-Canadian data because they can no longer afford Canadian data.
Most respondents (76%) still obtained statistics in paper format, while 61% used computer readable products, or both formats. Most researchers (65%) indicated they would most prefer to use computer readable files in order to obtain data. Only 1% of researchers prefer to use special tabulations, though 37% normally obtain data in this way. More "younger" researchers (those who began to do research after 1980) had used Statistics Canada computer readable data at some time than those who began to do research earlier.
The research findings support an interpretation that Statistics Canada's pricing and format changes had no effect on these researchers. But, this does not suggest that social scientists are unconcerned about the cost of the data they use. There is ample evidence in the survey that most are unhappy with the prices charged by Statistics Canada. The findings might be explained by a number of factors. Statistics Canada's monopoly on socio-economic statistics might result in inelasticity of demand. Further, the training of social scientists leads to habitual research patterns which might not be easily changed. Changing habitual work patterns could be too costly, in terms of time and energy expended, and researchers might, therefore, find some way to pay for increased data charges. Elasticity in the allocation of research funding might permit researchers to cope with increased prices. Given the fact that this research showed that there has been no change in the extent of use of Statistics Canada materials, these researchers must be acquiring funding somewhere or making other adjustments to their research allocations. About 23% of respondents noted changes in funding allocations because of price increases, some suggesting that they could no longer afford to hire students to work on the data.
The examination of use of paper formats versus use of computer readable formats does show a significant decrease in the number of articles using paper products. It is possible that the use of paper products declined because there are fewer paper products produced, or the decline might be price related. The lack of a corresponding increase in the number of articles using computer readable products suggests that price is more important than format in accounting for the decline in use of paper products.
Within the research data there are some hints of possible other effects which, while not statistically significant, deserve exploration in subsequent research. For example, it is possible that social science researchers might be using Statistics Canada's data with less intensity than they once did, but not exploring in them as widely.
Some social science researchers might be responding to price increases by writing more on foreign topics and publishing in foreign journals. This research did not consider the extent to which researchers wrote on foreign topics and were published abroad. This is a logical next step in considering pricing effects.
This study used a multidisciplinary approach and used traditional methods (case study, bibliometrics, and a survey) to examine policy effects. Studies of the impact of government information policy on access to government information have been few in number. This study adds to that small body of research by focusing on effects of policy on users. However, the study deals only with specific effects of particular government initiatives on a sub-group of government information users over a short period of time. The methodology used here is applicable in other contexts. Further research over longer time periods and with other user groups would build a body of data which can truly inform government information policy. Such research can contribute to future policy development.