Christian Léger, SSC President

As part of the International Year of Statistics, it is a pleasure to share the situation of statistics in Canada and some of the exciting projects being developed in the Statistical Society of Canada (SSC). I am grateful to Marie Davidian, ASA President, for the opportunity.

The last time Amstat News readers had a chance to read about the status of statistics in Canada was in March 2004 when Nancy Reid was president-elect. I will describe some of the links between our two organizations, list the activities of the SSC, and go over various aspects of statistics in Canada, before concluding with the future of statistics and some common challenges. Since a number of our members, in Canada and in the United States, are also ASA members, some of this will be familiar, but I hope other readers will learn something interesting.

The SSC and ASA Working Together

The SSC and ASA work together on a number of projects. We have joint chapters in Montréal, Ottawa, and Southern Ontario (they are called regional associations in the SSC). The SSC has representation of the ASA Committee on Women in Statistics and Committee on Meetings. ASA Executive Director Ron Wasserstein attends our Board of Directors meeting that takes place during our annual meeting, which allows useful sharing of information. Moreover, along with the IMS, ENAR, and WNAR, we are official partners of the Joint Statistical Meetings and constitute the Committee of Presidents of Statistical Societies (COPSS), which sponsors some of the most prestigious honors in our profession.

What Is the SSC?

The SSC is the only national organization in Canada devoted to the development of the professional interests of statisticians and probabilists. We have more than 1,000 members, including a good number of practitioners. About a quarter of our members are students, and recruiting and retaining them is a major challenge, as is the case with most organizations.

We now have six sections: Biostatistics, Survey Methods, Business and Industrial Statistics, Probability, the newly created Statistical Education, and Actuarial Science. The latter was created to serve as the home for academic actuarial scientists.

We publish the successful Canadian Journal of Statistics, as well as Liaison, our bilingual newsletter. In fact, we work hard to offer our services in both official languages (English and French).

The main event of the year is our annual meeting, which usually takes place in late May or early June. It attracts more than 500 participants over three days, plus one day of workshops. A meeting of that size is big enough to offer diversity in the scientific program, but cozy enough to allow ample opportunity for interaction among participants. It even includes a banquet! This is the appropriate venue to present our awards, including our most prestigious—the Gold Medal—for outstanding contributions to statistics of probability, but also the Canadian Journal of Statistics Award for the best paper appearing in a volume and the CRM-SSC Prize for excellence in research within 15 years of completing a PhD.

Our community greatly encourages student participation in the annual meeting, and many join the SSC this way. Students are invited to form teams to take part in the Case Studies for Data Analysis competition. They also are encouraged to present their research, and the best presentations receive an award. This year, we introduced the student conference, which took place just before the annual meeting last May in Edmonton, Alberta.

Another important program we have is accreditation. Implemented in 2004, our accreditation program has been successful. So far, the qualifications of professional statistician (P.Stat.) and associate statistician (A.Stat.) have been awarded to 147 and 84 individuals. We worked with the ASA when its accreditation program was introduced in 2009, and we have a reciprocity agreement so that the evaluation of an accredited member of one organization is fast-tracked if accreditation is sought in the other organization.

We also have Census at School, the popular project used in grades 4–12 to stimulate statistical inquiry, in common. Canada joined the international project in 2003, and Statistics Canada ran it until 2012. Due to cuts, Statistics Canada no longer runs it, and the SSC has taken the program over.

The SSC relies heavily on a large network of volunteers to carry out business. This is, of course, also true of big organizations like the ASA, but in our case, we don’t have a large staff to support the work of the volunteers. In fact, while we have contracted with the Canadian Mathematical Society for basic services, we currently do not have dedicated staff. I am grateful to the other members of the executive committee for the tremendous amount of work they accomplish, and we are looking at ways to improve the situation shortly by hiring an executive assistant.

Training

Statistics is healthy in Canada. Our universities, essentially all public, offer good undergraduate and graduate programs. In fact, students from all over may find the best value for their money at Canadian universities, given that tuition is usually lower than in the United States.

The quality of the research in statistics and probability in Canada has been recognized numerous times. For instance, in the last 30 years, seven of the prestigious COPSS award recipients were trained at either the undergraduate or graduate level in Canada. For a country of our size, this is a pretty good track record!

Of course, attracting good students to the field remains a challenge. While the financial crisis has affected Canada, our economy (and banking system) is robust enough that universities have continued to hire. It is worth noting that faculty salaries in Canada are for 12 months and research funding does not pay salary, even in biostatistics, except in rare circumstances.

Job Market

The job market for statisticians remains good in Canada. Most graduates entering the job market have a master’s degree. Job openings in biostatistics are, of course, often advertised, and the financial sector is also a big employer of statisticians. Because of the attractiveness of the job market for master’s-level graduates, recruiting local students for the PhD is often a challenge.

Research Funding

Research funding in Canada is quite different than in the United States, and there have been many important changes in the last few years. The main source of funding for academic statisticians and probabilists is the Discovery Grants program of the Natural Sciences and Engineering Research Council (NSERC). I am sure many of you have been invited to review such proposals. The discovery grants system funds a researcher’s program over five years. It is not a project-based program in that if the researcher eventually discovers a more promising research program, she can switch. It used to be that most active researchers in Canada would have an NSERC grant. On the other hand, they were typically small, with an average annual grant in the neighborhood of $12,000 and few larger grants in the $30,000–$50,000 range (there is no summer pay in these grants; they only cover travel, research, and student support).

The previous system was Markovian: The level of the previous grant was the starting point for the evaluation of the next grant. Five years ago, the system changed. Researchers are now explicitly evaluated on three criteria: the quality of the researcher, the quality of the proposal, and the formation of highly qualified personnel. All these criteria were used in the past, but now a grade is given to each independently. The grades are added, leading to a bin. The funding level of each bin is then determined, depending on the available budget.

So far, so good, except that statisticians and probabilists used to be evaluated in separate evaluation groups and now are all part of a single mathematics and statistics group. This means that the categories corresponding to grades (i.e., strong, very strong, outstanding, exceptional) must be calibrated among all mathematicians and statisticians, rather than simply within these two disciplines. This is generally a difficult task. Moreover, mathematicians often have a good opinion of their work, while statisticians are often critical.

Not surprisingly, the distribution of the bins for the mathematicians is stochastically larger than for statisticians so that in the last few years, there has been an important transfer of funds from statisticians to mathematicians. For instance, 20% of our budget was transferred to mathematicians in 2012. Needless to say, statisticians are very unhappy with the new system.

Funding for biostatisticians comes from NSERC for the methodological research and from the Canadian Institutes of Health Research (CIHR) for the clinical research. Unfortunately, the scarcity of funds makes the funding of methodological research in biostatistics by NSERC problematic, except for the very best researchers.

The advancement of medical research by biostatisticians will not count when evaluated by NSERC, so time devoted to clinical research will weaken a case as opposed to a statistician, so that the net result for many is to be denied funding by NSERC. On the other hand, CIHR rarely funds the advancement of methodological research.

The only bright light in the funding picture is the recent creation of the Canadian Statistical Sciences Institute (CANSSI). Despite the size of its population, Canada has three mathematical institutes like MSRI, one each in Montréal, Toronto, and Vancouver (CRM, Fields Institute, and PIMS), plus the fantastic Banff International Research Station (BIRS) located in the Rockies, just west of Calgary. All had activities in statistics, but despite the presence of a statistician on a scientific advisory committee, decisions were always taken through the prism of mathematicians.

As an outcome of a long-range planning exercise for the mathematical and statistical sciences mandated by NSERC—and chaired by Nancy Reid—CANSSI was created. Its NSERC funding will be modest at first, but the key point is that we will now have an infrastructure to develop research in the statistical sciences in Canada. Having witnessed the impact of the three mathematical institutes on the development of mathematics research in Canada and its rise on the international scene, I have no doubt CANSSI will eventually have the same impact for statistics.

Public Policy

As statisticians, we are all conscious of the importance of good official statistics to devise the very best evidence-based public policy. Canada is very fortunate to count on Statistics Canada, recognized worldwide for its excellence. The strength of having a single organization sharing its know-how among its census, labor, health, household, and agricultural programs—to name a few—is undeniable. Unfortunately, in the last few years, Canadians have witnessed a decline in the quality of their national statistics due to governmental decisions.

Since 1971, the census had mandatory long-form and short-form questionnaires. On the basis that the long-form census was too intrusive into the lives of Canadians (e.g., by asking how many bedrooms your house possessed), the conservative government decided in replace the long-form census with a voluntary national household survey in 2010. To compensate for the expected nonresponse and resulting bias, the government increased the sampling fraction from one in five to one in three households. Not surprisingly, the response rate decreased from somewhere around 95% to 70%, and all for a similar cost. The impact of this decision will be felt for years, as the long-form census played a crucial role in many decisions and served as a frame for other surveys.

In the same vein, the conservative government purposefully ignores a lot of data when developing policies, such as its continuing policy of getting tough on crime, while Statistics Canada data show crime statistics continue to decline.

Finally, most federal government scientists are banned from addressing the media without first going through the communications department. Talks at conferences, even by statisticians, need to be approved. In short, there is much we take for granted in a society, such as evidence-based decisionmaking, yet governments can have a great impact and change the way it does business in a way that is not always for the better.

Future of Statistics and Common Challenges

The future of statistics is of concern to the SSC, as well as for many other statistical organizations throughout the world. Working together on common challenges is important to make real, lasting progress. As we all know, statistics suffers from a branding problem.

In her SSC Presidential Invited Address at the 2013 annual meeting, Sallie Keller, former ASA president, showed a poster concerning Big Data with many terms associated with it and asked what was missing. The word statistics was nowhere to be seen! Hopefully, the International Year of Statistics will increase the awareness of the public regarding the importance of statistics as a field. But it will continue to be important to work together if we are to solve these problems. COPSS is a good forum for that purpose.

Another challenge we have in common with the ASA is accreditation. While it is already a success, much work remains to better publicize the program to employers of statisticians. The offering of services to accredited members is another issue. The number of accredited members is still relatively low, so the pool of people able to work at defining the required services and willing to organize them is relatively small. However, this is crucial for the long-term survival of the program.

Given that many of our services are freely available through our respective websites, and since there is a multiplication of statistical organizations throughout the world (globally organized around geography, ethnicity, or subject matter), it is important for all our organizations to work at developing a continued relevance. Tangible benefits of membership must be emphasized, but developing a strong sense of community is almost as important to our success.

I conclude by wishing you a very good International Year of Statistics. And what better way to celebrate than attending the Joint Statistical Meetings this August in my home city? Je vous souhaite la bienvenue à Montréal!