Introduction

The International Adult Literacy Survey (IALS) was the first-ever comparative survey of adults designed to profile and explore the literacy distributions among participating countries. It was a collaborative effort involving several international organizations, intergovernmental agencies, and national governments. In 2000, a final report was released (Organization for Economic Co-operation and Development and Statistics Canada, 2000), which stated that "by 1998, the survey had covered 10.3 percent of the world population and 51.6 percent of the world GDP" (p.87).

Who are the constituencies that are likely to use the data from the IALS once they have been collected and analyzed? It is expected that many individuals, including researchers, practitioners, and individual citizens within each of the participating countries, will read the survey results and make use of the data for a variety of purposes. Yet, the primary reason for developing and conducting this large-scale international assessment is to provide empirically grounded interpretations upon which to inform policy decisions. This places the IALS in the context of policy research. In their classic volume on this topic, Lerner and Lasswell (1951) argued that the appropriate role for policy research is not to define policy; rather, it is to establish a body of evidence from which informed judgments can be made. Messick (1987) extended this thinking to the area of large-scale assessments and noted that, in order to appropriately fulfill this function, assessments should exhibit three key features: relevance, comparability, and interpretability.

Relevance refers to the capability for measuring diverse background and program information to illuminate context effects and treatment or process differences. Both IALS and ALL developed and administered an extensive questionnaire covering a wide range of issues that will be used to identify characteristics that are correlated with performance and that may differ across a variety of language and cultural backgrounds.

Comparability deals with the capacity to provide data or measures that are commensurable across time periods and across populations of interest. Complex sampling, scaling, and translation procedures are being implemented to help ensure that common metrics will exist across participating countries so that appropriate comparisons can be made between countries and among major subpopulations of interest within a country. These comparisons are important both in this initial survey and in future assessments where new countries may join the survey and want to be placed onto existing scales, or where participating countries may want to measure trends in the distributions of skills among various subpopulations of interest.