Does statistical analysis matter? (2003)

Garnett Picot 2003 (Article based on presentations to the 2002 Canadian Policy Research Awards, Graduate Prize Seminar held in Ottawa October 2002.) Horizons , 6(1): 6-10

Although playing a supporting role in what is largely a political process, public policy is influenced by the manner in which a statistical program develops, and the empirical analysis conducted with the statistical output. However, this leaves many questions. Could public policy be better served if additional empirical research were brought to bear on these issues? Is the communications gap between the statistical and research world on one hand, and the policy development community on the other, too large, resulting in an inefficient use of the information that does exist? Do researchers adequately put the information in the form needed by policy analysts (or the public), or should they? Should policy analysts themselves spend more time and effort seeking out empirical research? These questions go far beyond the goals of this article.

One of the most important contributions a statistical system can make to policy development is to ensure that the basic data on trends and events are reliable, and seen to be so. Otherwise, it is the existence or magnitude of the trend that risks becoming the debate, not what to do about it. Maintaining a statistical system that has the trust of the population and policy analysts is essential.

Assisting with an understanding of why events are occurring is the next priority. Developing an adequate policy response depends on the understanding of the causes. In the area of labour markets and firms, the statistical system in Canada has moved a long way in developing data and analysis that assists in this way. The development of new special data sources (often longitudinal, and often in partnership with policy agencies) and related analytical capacity has been significant. Given this increase in the supply of policy-relevant data, increased analytical capacity with a public policy orientation in universities and research institutes as well as government could lead to considerable gains. Improving access to the newer micro-data bases is another important dimension to increasing policy relevant analysis. The development of the Research Data Centres by Statistics Canada, the university community and the Social Sciences and Humanities Research Council is playing a central role regarding access.

Continued efforts to form links between the policy and statistical/research communities can also produce improved policy-relevant information. Considerable progress has been made along a number of dimensions that could lead to more "evidence-based" policy development.