Which biosocial characteristics predict the accuracy of self-reported height and weight among adults? A comparison of data from Understanding Society Wave 1 and Wave 2 (Nurse Visit)

Presenter: Rasha Alfawaz, University of Leeds

Author: Rasha Alfawaz

Co-author(s): Graham Law, Eleanor Scott and George Ellison

Height and weight are important anthropometric correlates of many biosocial phenomena. However, few large scale social studies have the capacity to collect objective measurements of these variables, and many rely instead on self-reports which are vulnerable to systematic reporting error. The aim of this study was to assess the relative importance of 17 biosocial characteristics as predictors of the difference between self-reported and measured height and weight, in n=10811 respondents with complete data from both Wave 1, and objectively measured height and weight in the Wave 2 Nurse Visit of Understanding Society. Each of these analyses used multivariable linear regression to adjust for confounders identified using a directed acyclic graph. After adjustment for confounders, and for variation in the time interval between Wave 1 and the Wave 2 nurse visit, eight of the biosocial characteristics (gender, age, marital status, parenthood, physical health, education, employment and sporting activity) were significantly associated with the difference in self-reported and measured height; while five additional variables (smoking, alcohol consumption and three dietary characteristics) were also associated with the difference in self-reported and measured weight. Due to the interval between Wave 1 and the Wave 2 Nurse Visit, the difference between self-reported and measured height and weight will partly reflect genuine growth/shrinkage or weight gain/loss, respectively. But given this will have been less pronounced for height than for weight, it seems likely that the eight biosocial characteristics associated with the difference (between self-reported and measured variables) in both height and weight are those with the greatest potential impact on reporting accuracy.