Preface

Statistics Canada places great importance on the management of data quality to ensure that its statistical outputs satisfy user needs. The variety of measures it uses to manage the various dimensions of quality is described in the Agency’s Quality Assurance Framework. Within this overall framework, the design and implementation of the programs that acquire statistical data, whether through sample surveys, censuses, or the use of administrative records, clearly occupy a pivotal role in the management of data quality. The adoption of sound survey methodology provides the basis for producing statistical data that satisfy the information requirements of users.

In the domain of survey methodology there is no overall professional code of practice, or set of standards, that defines good practice for all situations. Survey methodology is rather a collection of practices, backed by a combination of theory and empirical evaluation, among which practitioners have to make sensible choices in the context of particular applications. These choices must attempt to balance the often competing objectives of relevance, accuracy, timeliness, cost, and reporting burden.

The present document consolidates a set of experiences and conclusions which, individually and in their particular contexts, have been judged to represent “good practice.” They will not all apply equally to every data acquisition process. Their applicability and importance will have to be carefully considered in light of the particular requirements and constraints of individual programs. This document must therefore be used with considerable professional care and judgement.

While the guidelines provided in this document are no substitute for informed expertise and judgement on the part of survey design staff, the underlying concern for quality must pervade all our activities. All statistical program managers are responsible for ensuring that Statistics Canada’s concern for quality is given appropriate weight in the design and implementation of statistical methods and procedures under their control.

The first edition of Quality Guidelines appeared in 1985. A subsequent edition with minor revisions was published in 1987. Significant developments in survey methodology since that time, for example in the integration of computer technology and electronic communications into data collection and processing, have led to changes in the flow of survey operations as well as to new approaches to the storage and dissemination of data. These advances in survey methodology motivated the issuing of a third version of the Guidelines in 1998. The present document has been significantly updated from the previous edition to reflect further advances in survey methodology over the past five years.

Thanks are due to the many Statistics Canada experts who have contributed to the preparation of these Guidelines over many years. The guidance of the Methods and Standards Committee and the comments of the Advisory Committee on Statistical Methods helped to make this a better document.

Ivan P. Fellegi
Chief Statistician