CRM-SSC Prize in Statistics 2018

David Haziza
CRM-SSC Prize in Statistics


The recipient of the CRM-SSC prize in Statistics is professor David Haziza of the Department of Mathematics and Statistics at Université de Montréal. This award recognizes a Canadian or permanent resident of Canada for outstanding research in the Statistical Sciences accomplished during the first fifteen years after earning a doctorate.

David Haziza grew up in Casablanca where he was schooled in the French system. His family moved to Canada when he was 15 and settled in Montréal. He obtained his BSc and his MSc from Université du Québec à Montréal before starting his PhD studies in survey sampling at Carleton University under the supervision of J.N.K. Rao. While completing his PhD degree, which he obtained in 2005, he started working full-time as a methodologist at Statistics Canada in 2000. Juggling a full-time position at Statistics Canada and completing a PhD thesis at the same time was of course very demanding. But it provided David with a unique outlook allowing him to do academic research on topics of crucial importance to practitioners. And to keep this edge in his academic research, he remained as a consultant one day per week when he left Statistics Canada to join Université de Montréal as an Assistant Professor in 2006. Since then, David became a star in the field and his research is having a large impact in the theory and practice of survey sampling.

David works in several areas of survey sampling, including variance estimation, survey sampling methods robust to influential observations, calibration, small area estimation, and design. But most of his contributions address the very important practical problem of missing data. In survey sampling, we distinguish unit nonresponse (when no information is collected on a sample unit) and item nonresponse (when the absence of information is limited to some variables only). For unit nonresponse, weighting is often a strategy of choice, whereas imputation is often the key for item nonresponse. Inference that takes into account imputation is important. It can either be based on an imputation model (for the outcome variable) or a nonresponse model (for the probability of having a response). But given that there may be uncertainty in these models, the availability of robust methods that work provided that at least one of these models is valid is very useful, both from a theoretical and a practical standpoint. This is the basis of the double and multiple robust methods that David has studied.

With more than 45 publications to date, David is a prolific author. Not only has he published in journals specialized in survey sampling such as Survey Methodology, Journal of Survey Statistics and Methodology, and the Journal of Official Statistics, but also in top general methodology journals such as Biometrika, JASA, Statistica Sinica, Scandinavian Journal of Statistics and Statistical Science. This shows that his work is fundamental and influential.

Another sign of excellence is his popularity as an invited speaker in various conferences – in fact, more than 10 a year. Indeed, David has spoken or given workshops in five continents! Especially noteworthy are the one day (or half a day) workshops that he gave in Marrakech, Morocco (2017), Washington, USA (2017), Geneva, Switzerland (2016) and his keynote presentations at the Boston JSM (2014), the Colloque francophone sur les sondages in Dijon, France (2014), the Journées de Méthodologie Statistique, Paris, France (2012) and at the Rao meeting in Kunming, China (2017).
Because of his expertise, David was invited to be part of the 2013-14 SAMSI Program on Computational Methods in Social Sciences. Since 2015 he has lead a CRT of CANSSI. In 2016 he was appointed as a member of the Committee on the Review of the Marine Recreational Information Program organized by the U.S. National Academies of Sciences, Engineering, and Medicine. His expertise and judgment are also sought by journals. He currently serves as Associate Editor of five journals, including JASA, JRSS B, and Scandinavian Journal of Statistics.

David has already received a number of awards in his young career, including the two most prestigious teaching awards at Université de Montréal and the 2018 Cox Award. He also became an ASA Fellow in 2016.

David is married to Mélanie. They have two daughters, Marianne and Aurélie. As for David, if you see him in a corridor, he is either talking about a statistical problem or one of the great political problems of our world.

To conclude, David’s strong research record is based on strong theoretical arguments and motivated by practical issues found in the field. His impact is felt in national statistical organizations in Canada, France, and in the US, as well as in other organizations, such as Westat. He is a popular speaker in conferences and workshops. His scientific leadership in survey sampling is already well established internationally. One of the leading experts supporting the nomination summarized it best when he wrote: “His work in each of the mentioned areas is impressive, but it is the overall scale of his activity in published research, in organizing and presenting at meetings, in directing postgraduate students, in service as Associate Editor, and in other service to the profession that is truly impressive.” Another one wrote that “even at this stage, he is already one of the leading researchers in survey statistics worldwide, continuing a traditional strength of statistical research in Canada”, which makes us all proud.

David will present an overview of his work in a special session at this year’s SSC Annual Meeting at McGill University.

The citation for the award reads: 

“To David Haziza, for outstanding contributions to survey sampling theory and practice, notably, path-breaking methodology for missing data, innovative methods that improved the robustness of estimation and for their impact on the practice of national statistical agencies.”


Thanks to Christian Léger, who was primarily responsible for producing this material, and Dianne Rioux for the photo.