Measures of self-reported well-being: their affective, cognitive, and other components

Abstract

This investigation begins from the hypothesis that social indicators of perceived well-being — e.g., people's assessment of their own life quality — will, like other attudes, reflect two basic types of influences: affect and cognition. In addition, the indicators were expected to include two other components: unique variance (mainly random measurement error) and correlated measurement error. These ideas are investigated using a structural modeling approach applied to 23 assessments of life-as-a-whole from a national survey of Americans (N=1072) and/or a survey of urban residents in England (N=932). In both sets of data, models that included affective and cognitive factors fit significantly better than more restricted models. Furthermore, as expected, measures of (a) ‘happiness’, ‘fun’, and ‘enjoyment’ tended to be relatively more loaded with affect than were measures of (b) ‘satisfaction’, ‘success’, and ‘meeting needs’; and (c) measures designed to tap both affect and cognition tended to fall between the first two groups. In addition, the results suggest that measures employing relatively many scale points and direct assessments yield more valid indicators of people's evaluations of life-as-a-whole than do measures based on three-point scales or on explicit comparisons with other times or groups. These results contribute to basic knowledge about the nature of life quality assessments, help to explain some previously puzzling relationships with demographic factors such as age and education, and may be useful to designers of future studies of perceived well-being.