Subjective Successful Aging: Measurement Invariance Across 12 Years

Abstract

Background and objectives: Debates about how to define successful aging have dominated gerontology for over 60 years. Regardless of how successful aging is conceptualized, in order to accurately understand how the construct changes over time and how it differs between people of varying ages, successful aging must be measured with instruments that are valid, reliable, and have measurement invariance. These analyses focus on subjective successful aging and examine the extent to which a reliable, valid, 3-item scale has measurement invariance across 12 years for individuals aged 50-86.

Research design and methods: We analyzed 5 waves of data collected from a panel of 5,688 community-dwelling people aged 50-74 when recruited in 2006. We tested measurement invariance using the standard 4 nested steps,, introducing increasing parameter constraints at each step. Analyses were conducted using Mplus 7.

Results: Analyses revealed that the 3-item scale measuring subjective successful aging has adequate measurement invariance across time. We demonstrated that the scale has configural, metric, and scalar invariance by most standard metrics. Only residual invariance was not supported. However, because residuals are not part of the latent factor, invariance of the item residuals is inconsequential to interpretation of latent mean differences.

Discussion and implications: Findings provide the foundation needed for researchers to examine change in subjective successful aging over time, differences in subjective successful aging between people of varying ages, and predictors of subjective successful aging, confident that the scale has adequate measurement invariance.

Keywords: Configural invariance; Confirmatory factor analysis; Metric invariance; Scalar invariance.