Integration of survey instruments
Integrating survey instruments means making them consistent with each other and with other data collection instruments used by a country’s statistical system. This involves using comparable questions, concepts, classifications, etc. Objectives of integration include generating a coherent set of data, comparable over time and across sources when relevant, and ensuring cost-effectiveness of data production.
Challenges will vary depending on whether a country has an established history in the collection of social data through administrative records, censuses, or household surveys, and whether it is relative newcomer to these activities. The former may face challenges of integration and harmonization. The latter may find the job easier, although the lack of experience using household surveys will be a challenge in itself. Nevertheless, both types of agencies will be challenged to develop meta databases to assist in both the collection and dissemination of data. They will be challenged to develop standards that will adhere in every degree possible with international standards. They will be challenged to integrate data from various sources to ensure balanced and comprehensive reporting. And they will be challenged to provide information rather than just basic data.
Some keys to success in meeting these challenges will be the following:
National statistical agencies must carefully guard their political independence. It can be a challenge, but if the statistical outputs are to have credibility and, therefore, utility, agencies must ensure they are free of political interference. One of the best ways to ensure independence of statistical agencies is to develop strong partnerships with major clients, including international agencies, government departments, academics, non-governmental agencies, businesses, and the media.
Relationships with the media are especially important, not only from the perspective gained from listening to media and public perceptions of needed avenues of inquiry, but also from the opportunity to provide the media with not just data but information, i.e., analytical outputs, written in a journalistic style that can be easily pasted into communications vehicles. Information to which the public can relate, such as profiles of geographical regions or particular population groups, are useful. For example, profiles of single-parent families, seniors, children and youth, or agricultural workers can help develop an informed population and provide critical information to decision-makers.
It is crucial that data from various sources are consistent within the agency. Concepts, definitions, naming conventions, and classifications must be consistent from one source to another. Agencies, therefore, must waste no time in developing a standards policy that embraces international standards as much as possible. The policy should ensure conceptual frameworks are used for consolidating statistical information; standard names and definitions are used for population groups, statistical units, concepts, variables, and classifications; and consistent methods of collection and processing are used. The policy must be controlled and enforced at the highest level of the agency.
A meta database must be developed to support the policy. The main function of national statistical agencies is to produce and disseminate statistical data on economic and social conditions in that country. Statistical data take the form of numbers in data files, statistical tables, or in texts such as news releases and articles. These numbers on their own cannot be understood. Explanatory information is called metadata, and it is essential for understanding and interpreting statistical data.
At a minimum, metadata must cover the description of the data. A standard used by Statistics Canada to structure and present this type of metadata is ISO/IEC 11179, “Information Technology – Specification and Standardization of Data Elements.” In statistical terminology, data elements are commonly referred to as variables; this standard, therefore, provides a guideline for structuring and presenting basic information about variables. The process of creating information according to this standard, however, also brings about more consistency and rigor in the conceptualization, naming, and organization of variables for which data are produced. Experience shows that metadata are best documented at the outset of any new survey design or redesign rather than after the fact. When they are part of the process, they actually assist it.
Metadata support three essential statistical activities. First, they support the design and development of new surveys or redevelopment of existing surveys and provide an immediate record of what was done before and might be used again in the future, thereby rendering significant efficiencies and cost savings. Second, metadata provide a platform for developing and maintaining standards and best practices, which also leads to efficiencies, cost savings, and improvements in product. Finally, metadata promote and encourage effective mining of an agency’s data holdings.