Summary: The culture and ethnic backgrounds of respondents can affect measurement in social science research, but these effects have been largely ignored in scale development and evaluation. Instruments are often used in populations that differ from those on which the scale was developed. Cross-cultural use of measures without investigation into the applicability of the measure across cultures poses a problematic measurement issue. In this investigation, the performance of the PANAS (Watson, Clark, & Tellegen, 1988), a widely-used measure of positive and negative affect, was examined in a large community sample of ethnically diverse, older, rural participants. Tests of scale invariance across multiple groups (defined by ethnicity and gender) were conducted using multiple-group confirmatory factor analysis and a sequence of nested tests. These tests revealed equivalence of a correlated two-factor structure with correlated errors. Partial invariance was found among factor loadings, intercepts, and other structural parameters. Evidence for convergent validity included relationships between the PANAS scales and demographics, health, and psychosocial functioning that were consistent with those found in previous research. Overall, the PANAS was found to be relatively invariant across Black and White men and women and showed high reliability and validity across both the standard and shortened versions. Future research should address cross-validation of these analyses to assess the stability of these findings and explore psychometrically appropriate approaches for probing the results of invariance tests.