The role of answering behaviours on weight misreporting

Abstract

Biases in self-reported weight are very common among young adults and adults. Although social norms are the most commonly accepted explanation for these misreports, corresponding evidence is scarce and conflict-ridden. An alternative explanation for biases in weight self-reports comes from answering behaviours; non-random rounding, formally an answering behaviour, has been found to play a significant role in several studies of weight misreporting. However, the presumably rich role of answering behaviours has seldom been explored. This study brings a second answering behaviour into the analysis: inconsistency. An inconsistency index was computed as an individual-level score from several questions across waves in the National Health and Nutrition Examination Survey. By regression analysis (N=3480 men and 1856 women) the simultaneous role of inconsistency and of non-random rounding on weight misreporting was explored. Inconsistency was found to be associated with higher self-reported weights. Inconsistent individuals provided significantly different misreports, with women under-reporting 0.23[kg] (0.01-0.45) less and men over-reporting 0.42[kg] (0.02-0.82) more than their consistent counterparts. Inconsistency was found to play a simultaneous and substantially larger role than non-random rounding. This result was clearer among men than it was among women. Although social norms are usually thought to be the central explanation of weight-biased misreports, there are other factors, such as answering behaviours, that might play a more influential role.