Sampling theories suggest that quantitative judgments are comparison-based and, therefore, fundamentally relative, and susceptible to continuous revision based on information sampled from our environment. Body satisfaction and weight judgment in women are a domain of clinical relevance in this context, as they have been shown to be related to unfavourable social comparisons.
To predict sampling-based influences on judgment, not only the central tendency of the comparison sample, but also its variability, need to be considered.
Following a simple updating approach, lower variation should lead to longer-lasting effects of bias in prior distributions than higher variation, that is, to assimilation effects. However, judgments may also be influenced by expectancy violations: In a narrow distribution, strong deviations from the mean are rare, and may therefore invite more extreme judgments, that is, produce contrast effects.
We investigated whether the distribution of potential social comparison standards in the domain of weight could function as an “experimentally induced prior”. In a continuous sampling task, we presented schematic figures of female bodies which varied systematically in weight. During the first block, participants observed one of six distributions which differed in the means and standard deviations of the stimuli’s body-mass-index. In a second block, stimuli were drawn from the same wide normal distribution with a conventionally midweight mean.
In line with expectancy violation, we observed a contrast effect: Body-size judgments were more strongly biased away from the prior distribution if it was narrow than if it was wide.