Generally speaking, mathematical statisticians are responsible for survey methodology and related research activities. Their work is crucial to Statistics Canada: the costs of survey operations, the quality of survey outputs and the reduction of respondent burden are heavily dependent on the methodology used.
The primary role of most of the mathematical statisticians is to provide methodological services for Statistics Canada surveys. This service is oriented to our internal clients and requires mathematical statisticians to acquire some knowledge of the client area in order to determine and meet its methodological needs. To this end, mathematical statisticians are usually grouped into divisions and sections according to the clients served.
Statistics Canada surveys are normally developed through multidisciplinary project teams. Such teams consist of four main groups: subject-matter (e.g., economists, sociologists, geographers), survey operations, systems development and survey methodology (mathematical statisticians).
Within a team, the mathematical statistician brings his or her expertise, experience and a critical, analytical mind to the area of statistical and survey methods. In the multidisciplinary project structure, he or she is primarily a service provider. The mathematical statistician is responsible for providing methodology services for designing, implementing and evaluating statistical methods, mainly in the following areas: survey frame creation and maintenance, sample design, questionnaire design, data collection methods, edit and imputation of data, treatment of non-response, estimation of parameters of interest and their variances, data analysis, measurement of non-sampling errors, protection of confidentiality, along with quality control/assurance and time series analysis.
The work of a mathematical statistician is highly diversified and requires creativity and adaptability. Individual mathematical statisticians usually work on several surveys or survey activities at the same time. Each survey is a new challenge; ready-made solutions in the literature can rarely be directly applied. To design and implement effective, scientifically sound methodology, a mathematical statistician must always maintain a good balance of skills in analytical and empirical research and operational work. The majority of mathematical statisticians provide survey methodological support services while some of them are also involved in research.
Like any organization striving to be the best in its field, Statistics Canada places great emphasis on research. The organization has a long and rich tradition of research in the area of statistical methods. In particular, methodological research focuses on cutting-edge ways to produce reliable statistics at lower cost.
Statistics Canada provides the environment and support that enable mathematical statisticians to deal with all types of research problems, whether they are associated with a specific survey or are more general in nature.
Mathematical statisticians regularly conduct theoretical studies and empirical simulations to support the methodological services they provide to project teams. They also carry out research on a variety of subjects related to statistical methods, such as spatial and temporal estimation, non-sampling variance, outlier detection, benchmarking, interpolation and calendarization, longitudinal data analysis, and time series. Initially, research could make up only a small part of the duties of a mathematical statistician. However, it could become a major part of his/her duties depending on his/her experience and interests.
Results of mathematical statisticians’ research projects may be presented at relevant conferences and published in technical journals. One of these is Statistics Canada’s own internationally renowned journal, Survey Methodology.
To foster and promote research, Statistics Canada has established an external fellowship program. In addition, Statistics Canada is actively involved in joint research projects with universities and with statistical agencies in other countries.