Recent decisions about the mandatory long-form census are not as much of an unforeseen development as many seem to suggest. They are part of a general trend in government away from the kind of data gathering that can be critical to good policy-making.
The last time this trend made news was three years ago, with the Paillé report on opinion research practices in the federal government. To be clear, the research that was the subject of that report has nothing to do with the census. It typically takes the form of standard telephone opinion polls, focusing on issues related to public policy.
The Paillé report suggested that government spending on opinion research would benefit from more co-ordination and restraint. These are sensible suggestions, and the Harper government responded decisively. According to annual reports by Public Works and Government Services Canada, the total number of opinion research projects managed by Health Canada was 105 in 2003-04 and only 22 by 2008-09. For Human Resources Development Canada, it dropped from 63 to eight in the same period; for Environment Canada, 27 to one. In the policy world, this is a colossal change.
Significant policy change can be laudable, of course, and even be a sign of responsive decision-making. But it's difficult to know whether this was the case where opinion polling was concerned. There was relatively little public debate on the matter and, like the census decision, this shift in the government's data-gathering efforts occurred without much serious discussion.
Where opinion polls are concerned, more debate may not have helped. It's too easy to paint a picture of manipulative and partisan polling firms, particularly given that media coverage of polling focuses on voter intentions. Polling firms conduct serious policy research, however, and the content of government-sponsored polls is more mundane but also more important. The census, for example, is used to help identify the regions that require a given policy; opinion research is used to see how citizens feel about those policies.
Let's consider the two most common arguments for greatly reduced spending on opinion research: to save money, and because we learn little from polling.
Saving money is not a good argument in its own right - a government shouldn't save money just for the sake of saving money, but rather consider various options for saving money and find ways to spend less without doing too much damage to the quality of policy-making and governance.
The suggestion that we learn little from opinion research is simply wrong. Polls can (and do) measure how many citizens have family doctors; whether it's easy or practical to use public transit to get to work; whether Canadians support more foreign aid, or stronger sentencing for criminals; or how we feel about our experiences at national parks or with Revenue Canada. All of these can be important inputs into the policy-making process.
The census is clearly on a different scale and of greater importance than opinion research. But the justification for reductions in one are not very different from the justification in reductions in the other. And regardless of which body of research we're concerned with, 2006 may be the last year for which Canadian governments have a thorough view of what their citizens want or need.
Dropping the long-form census is a mistake, but this decision is part of a larger trend. Data gathering is a central component of policy-making, policy implementation and policy evaluation. Governments, and Canadians, should think seriously before making changes that will have a pronounced effect on governments' capacity to deliver good public policy.
Stuart Soroka is an associate professor of political science at McGill University and director of the Canadian Opinion Research Archive at Queen's University.Report Typo/Error
Follow us on Twitter: