In much of psychology, the experimental stimuli are considered a fixed, rather than random, factor. It is rare than in the creation of models the stimuli are considered random effects even though these same models are then used to generalise over entire populations of stimuli. Quoting Yarkoni:

The clear implication of such findings is that many literatures within psychology are likely to be populated by studies that have spuriously misattributed statistically significant effects to fixed effects of interest when they should actually be attributed to stochastic variation in uninteresting stimulus properties.