Part C

Pilot Study: Summary of Numeracyrelated results and decisions

8. Overview and purposes of the pilot stage

The ALL Pilot study was conducted in 2002 in six countries (and five languages): Belgium (French), Canada (French and English), Italy, Norway, Switzerland (German, French, Italian), and the United States. Large samples responded to test booklets assessing the key skill domains targeted by ALL: Document literacy, Prose literacy, Numeracy, and Problem-solving. The Background Questionnaire (BQ) provided additional information about respondents' experiences with Information Technology and about numerous variables of importance such as employment status, consumption of health services, participation in learning activities, and perceptions of quality of life. The BQ also collected additional personal information that could help to explain or interpret results pertaining to the key variables.

8.1 General goals

The Pilot study was designed to accomplish two goals:

  1. Enable the ALL management and the development teams to evaluate the psychometric characteristics of the item pools created to assess the different skill domains, and provide sufficient data from which to select the items to be included in the test booklets and the BQ that will be used in the Main ALL survey.
  2. Enable participating countries to field-test and evaluate all survey materials and administrative procedures planned for the Main ALL survey, such as contacting respondents, setting home visits, test administration, scoring of responses, data capture, and quality assurance mechanisms.

8.2 Numeracy-specific goals

In addition to the two general goals listed above, the Pilot study aimed to provide information regarding specific issues relevant to Numeracy assessment in ALL:

  • Enable selection of 40 items out of the 81 Numeracy items tested in the Pilot, whose development was described in Part B. The target number of 40 items was determined on the basis of knowing average response times per item in the feasibility studies, and the overall time expected to be available for Numeracy assessment in the Main study.
  • Analyze response patterns and verify that errors are not caused by unclear or problematic stimuli, questions, or instructions, or by inconsistent translation. Also, determine whether the multiple scoring codes created for some items should be maintained for the Main study
  • Collect data on the extent to which respondents used the calculator provided as part of the assessment.
  • Evaluate the relationships between Numeracy scale scores and variables measured by the BQ, including scores on scales related to numeracy practices and to beliefs and attitudes.