Framing quality management

With a little bit of experience or research, we can judge the quality of the clothing we buy, the food we eat or the car we drive. However, when it comes to data, quality is much harder to measure. After all, there is no comparison shopping when Canada only has one national statistical office. Nor can data users see a survey being conducted the way restaurant patrons can watch a chef assemble a sushi roll. Instead, for quality of data, users depend on the standards of the organization that produces it.

In April, Statistics Canada released the third edition of its Quality Assurance Framework (QAF). Over 12 chapters, this document describes the principles and strategies in place to ensure that StatCan produces high-quality data.

New edition

Jack Singleton, a senior methodologist in StatCan's International Cooperation and Corporate Statistical Methods Division, explains how the new edition fits into the work of the agency. "Its purpose is to give guidance. It's to ensure that individuals and program managers are using the most up-to-date tools, protocols and good practices that are in place to help them do their job." It also gives an insider's view to our data users, so they can be assured that the data they use is produced in a responsible and professional manner.

The QAF is a useful tool for program managers, particularly when designing new programs or redesigning existing ones. For instance, a manager who is creating a new survey may consult the chapters on managing input data and relations with data providers, allocating and managing resources, and managing relations with data users and stakeholders. Senior managers also refer to it when doing corporate planning exercises, identifying strategic investments for the agency and evaluating potential risks.

The framework is also available as a general reference for anyone working in official statistics—inside or outside of Statistics Canada. Internationally, Statistics Canada is recognized for its leadership in quality management. Within Canada or abroad, organizations producing statistical information can use the QAF to raise awareness or to assess or enhance their own practices.

Changing conditions

This edition has been a long time coming: the previous version was released in 2002. At that time, StatCan was at the forefront of international agencies in creating an overarching governance document for quality management. However, in the past 15 years, the conditions under which official statistics are produced have changed remarkably.

Several factors have affected the way that StatCan gathers, produces and disseminates data. Examples include: the emergence of electronic collection; the development of new sources of data, including the growing use of administrative data; greater use of the Internet for sharing Statistics Canada data; and public expectations of more timely data at finer levels of detail.

While the agency responded to these changes as they emerged, revising the framework brought StatCan's top quality-management tool up to speed.

There was a second reason for updating the QAF. Just as changes in the outside world have affected the way statistical agencies operate, changing expectations and technologies have led to a new kind of governance document.

The new edition is designed as an online document, which includes hyperlinks and the option of searching by keyword. It will be revised on an ongoing basis, as linked content changes or if specific references are no longer relevant.

The revised framework also has a much broader scope than the previous editions. The second edition focused mainly on StatCan's outputs, and while this content remains in the new version, the framework now features additional sections on the corporate environment and the statistical program, such as management of relations with data providers, data users and stakeholders.

Producing data is a complicated process that takes place largely out of sight. But if there is an equivalent of "checking under the hood" when assessing the quality of StatCan's data, data users would see the new Quality Assurance Framework when they look behind the numbers.