The central hypothesis of appraisal theories posits that individual differences in emotion are due to differences in appraisal, rather than to objective stimuli. Although many studies have investigated this hypothesis, a statistically rigorous approach to verifying this notion has not been explicitly attempted. A multilevel linear model enables this hypothesis to be formally investigated by testing for significant random effects in an appraisal-emotion association at the individual level. The current study induced fear with an experimental and parametric manipulation of danger through depths up to 60 meters in virtual reality. Following each trial, participants (N = 32) reported on their appraisal of the depth and resulting fear feelings. Multilevel analyses focused on the association between virtual depth and danger appraisal, depth and fear intensity, and danger and fear, modelling both the population-level association and individual-level differences (random effects)...