Learning Objectives: Describe two kinds of psychometric validity and how they are important to tobacco disparity reserach Describe how confirmatory factor analysis can be used to establish that an instrument is measureing the same thing in the same way in different racial/ethnic groups. Demonstrate the calculation of item bias statistics from confirmatory factor analysis results.
Abstract: Audience:Tobacco researchers interested in racial/ethnic disparities in tobacco use, especially those involved in research design, data analysis, and instrument selection, will find this workshop helpful.
Key Points:When self-report instruments are used in tobacco disparities research, it is often necessary to validate the instruments for use in specific racial/ethnic groups before making substantive comparisons. Statistics commonly used to evaluate the performance of self-report instruments such as internal consistency alpha and item-to-total correlations are based on classic true score theory and intended for use in studies of a single population or ethnic group. These methods cannot establish cross-cultural equivalence. The classic true score model is being supplanted by the more general common factor model. This presentation will show how the common-factor model (CFM) in conjunction with multi-group modeling with means structures is being used to test the psychometric equivalence of instruments across different racial/ethnic groups. Instruments from the investigators’ currently funded study of smoking behavior in Arab American youth will be used to motivate the presentation and to provide substantive interpretation.
Educational Experience:Worksheets will give participants the opportunity to calculate example results at two times during the session.
Benefits:These methods are being increasingly used in health and behavioral research yet they are not well understood. This workshop will increase understanding of the common factor model and how to evaluate psychometric equivalence.
Back to Poster Session
Back to Evaluation and Surveillance
Back to The 2003 National Conference on Tobacco or Health (December 10-12, 2003)