Comparing the Well-Being of Older Europeans: Introduction

Subjective self-ratings reported by survey respondents are used frequently in the social sciences; examples of self-assessments are personal health, well-being, work ability, and job satisfaction. Unfortunately, self-ratings have been found to be subject to substantial reporting bias. The key problem is that subjective self-ratings involve respondents’ evaluation of some domain of their own objective reality (such as their health) as well as their subjective thresholds for mapping their evaluation onto the response scale defined in the survey instrument (e.g., ‘‘excellent’’, ‘‘good’’, ‘‘fair’’, or ‘‘poor’’). When these thresholds vary across respondents, their responses are not comparable any more—a phenomenon referred to as ‘‘differential item functioning’’ Much can conceivably be learned from cross-national research using surveys among households and individuals with self-assessments of physical or mental health in various domains, and questions on satisfaction with, for example, household income, jobs or daily activities, or family life and social contacts. If, however, such comparisons suffer from differences across countries and socio-economic groups in the way people answer subjective survey questions, the method of anchoring vignettes is a potentially effective way to solve this problem. Anchoring vignettes are defined as ‘‘short descriptions of a person or a social situation which contain precise references to what are thought to be the most important factors in the decision-making or judgment-making process of respondents’’ (Alexander and Becker, 1978, p. 94). Statistical methods for adjusting self-rates for differential item functioning with the help of anchoring vignettes were developed by King et al. (2004). The basic idea is to get from the same respondent or group of respondents not only a self-rating for some variable of interest, such as health, job characteristics of satisfaction with life, but also ratings for vignette persons whose descriptions keep the levels of that same variable fixed.