The 131st Annual Meeting (November 15-19, 2003) of APHA

Abstract #67104

Adam C. Carle, MA, PhD, Statistical Research Division, U.S. Census Bureau, Center for Survey Methods Research, Washington, DC 20233-9100, 301-763-1863, adam.c.carle@census.gov

Unlike other sciences, measured variables in the social sciences are often unobservable, latent variables. Attempts to compare the correlates, prevalence, and course of mental illness across groups are dependent on the validity of the measurement instrument. When measuring unobservable variables, the possibility exists that a test may be biased across groups. Measurement bias is present when individuals equivalent on the latent trait, but from different groups, do not have identical probabilities of observed scores. Bias can lead to inaccurate diagnoses, attenuate or accentuate group differences, and drastically affect the validity and reliability of epidemiological studies. As public health research looks toward behavioral and psychological determinants of health, measurement bias becomes particularly salient. It is essential that an instrument’s validity be established before utilizing it in cross group settings. Latent trait methods (LTM’s) offer a powerful method of investigating bias. Unfortunately, few social scientists receive training in LTM’s; find it difficult to understand studies employing them; and issues of measurement invariance are often ignored. This poster results from the student’s comprehensive review of studies using LTM’s to assess bias in measures of psychopathology and has three goals. First, for a general audience, it intends to briefly review LTM’s for assessing bias. Second, it will summarize the findings of studies in the psychological literature that have used these methods. Finally, the results of these studies will be placed in a public health context, addressing why public health research should be concerned about the results of these studies.

Learning Objectives:

• At the end of the session the learner will be able to discuss: measurement bias, latent trait models for assessing measurement bias, and the realtion of measurement bias to public health research.

Keywords: Methodology, Statistics

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.