Back to Annual Meeting Page		American Public Health Association 133rd Annual Meeting & Exposition December 10-14, 2005 Philadelphia, PA

Abstract #106796

Ruth L. Eudy, MSW, PhD, Graduate Program in Health Services Administration, University of Arkansas at Little Rock, 205 Ross Hall, 2801 S. University Avenue, Little Rock, AR 72204, 501-569-8667, reudy@aristotle.net

The issues of differential validity and reliability can pose problems for public health researchers studying populations with differing racial, ethnic, socioeconomic, and cultural subgroups. If reliability and validity of constructs differs between subgroups, then validity and interpretation of outcomes based on these constructs is unclear. For instance, in order to measure the outcomes of interventions for depressive disorders, we must be certain that our depression measures capture the same underlying construct with equal reliability in all subgroups.

Structural equation modeling (SEM) provides an excellent tool for determining differential validity and reliability of scales within large population surveys. Moving beyond traditional factor analysis approaches, SEM allows us to test for differences in factor structure, validity of individual indicators, comparability of measurement error, and differences in construct means between subgroups.

SEM is also useful to determine comparability of the same scales administered as components of different surveys or in different languages. This study analyses data from the 1991 NMIHS, the 1995 NLS and the Hispanic version of the NHANES to compare validity and reliability of the Center for Epidemiological Studies Depression Scale for subgroups of respondents both within and between studies. Significant differences in validity of subscales were found for African American and Caucasian women and for respondents to the Hispanic version of the NHANES compared with respondents to the NMIHS and NLS.

Techniques and handouts presented will be useful tools for researchers in the areas of health disparities and outcomes research, whether applied to mental health or other public health issues.

Learning Objectives:

At the conclusion of this session, participants will
• Recognize and articulate the implications of differential reliability and validity for public health survey research
• Describe the application of structural equation modeling (SEM) techniques for determining differential reliability and validity of outcome measures for subgroups in population-based surveys.
• Acquire detailed handouts with instructions for the analytical techniques described.

Keywords: Statistics, Outcome Measures

Presenting author's disclosure statement:

I wish to disclose that I have NO financial interests or other relationship with the manufactures of commercial products, suppliers of commercial services or commercial supporters.