Assessing validity and bias of within-person variability in affect and personality

Within-person variability in affect and personality have both been linked to well-being. Variability in negative affect is considered a particularly important individual difference, given that it is a core component of Neuroticism from the Big Five personality traits. Researchers typically measure individual differences in variability either by asking people to report how variable they perceive themselves to be or to give multiple reports on the construct and calculating a within-person standard deviation that is adjusted for confounding by the person-level mean. The two measures are weakly correlated with one another and this lack of convergent validity is concerning because the links of variability with well-being depend on which measure researchers use. One hypothesis is that people’s repeated ratings may be biased by response styles. In a large (N = 399) 7-day study with up to 5 measurements per day, we (1) examined the validity of different measures of variability in affect and personality, and (2) tested the response styles hypothesis by correlating variability in repeated ratings of affect and personality with variability in repeated ratings of a theoretically unrelated construct (i.e., features of images). The measures of variability lacked sufficient convergent validity to be used interchangeably, and only 2 of 10 correlations testing the response styles hypothesis were statistically significant and weakly positive, providing weak evidence supporting the hypothesis. Therefore, the lack of convergent validity between the different measures leaves ambiguity in our understanding of the links between variability and well-being, and this is not explained by response styles bias.