Using externally collected vignettes to account for reporting heterogeneity in survey self-assessment

Abstract

The anchoring vignette approach has grown in popularity as a method to adjust for reporting heterogeneity in self-reported survey questions, removing bias due to systematic variation in reporting styles across study respondents. The use of anchoring vignettes, however, has been limited to surveys where both self-reports and vignette questions have been included. This diminishes their wider application. We illustrate, using an application to self-reported health in a large household survey, how externally collected vignettes can be used to adjust for reporting heterogeneity in datasets that have not included anchoring vignettes. Given that self-reported measures are an important facet of social, health and economic research, we anticipate the approach described will lead to new applications of the anchoring vignette methodology.