Validation of Subjective Well-Being Measures Using Item Response Theory

Background: Subjective well-being refers to the extent to which a person believes or feels that her life is going well. It is considered as one of the best available proxies for a broader, more canonical form of well-being. For over 30 years, one important distinction in the conceptualization of subjective well-being is the contrast between more affective evaluations of biological emotional reactions and more cognitive evaluations of one’s life in relation to a psychologically self-imposed ideal. More recently, researchers have suggested the addition of harmony in life, comprising behavioral evaluations of how one is doing in a social context. Since measures used to assess subjective well-being are self-reports, often validated only using Classical Test Theory, our aim was to focus on the psychometric properties of the measures using Item Response Theory. Method: A total of 1000 participants responded to the Positive Affect Negative Affect Schedule. At random, half of the participants answered to the Satisfaction with Life Scale or to the Harmony in life Scale. First, we evaluate and provide enough evidence of unidimensionality for each scale. Next, we conducted graded response models to validate the psychometric properties of the subjective well-being scales. Results: All scales showed varied frequency item distribution, high discrimination values (Alphas), and had different difficulty parameters (Beta) on each response options. For example, we identified items that respondents found difficult to endorse at the highest/lowest point of the scales (e.g., “Proud” for positive affect; item 5, “If I could live my life over, I would change almost nothing,” for life satisfaction; and item 3, “I am in harmony,” for harmony in life). In addition, all scales could cover a good portion of the range of subjective well-being (Theta): −2.50 to 2.30 for positive affect, −1.00 to 3.50 for negative affect, −2.40 to 2.50 for life satisfaction, and −2.40 to 2.50 for harmony in life. Importantly, for all scales, there were weak reliability for respondents with extreme latent scores of subjective well-being. Conclusion: The affective component, especially low levels of negative affect, were less accurately measured, while both the cognitive and social component were covered to an equal degree. There was less reliability for respondents with extreme latent scores of subjective well-being. Thus, to improve reliability at the level of the scale, at the item level and at the level of the response scale for each item, we point out specific items that need to be modified or added. Moreover, the data presented here can be used as normative data for each of the subjective well-being constructs.