We investigated whether cognitive process models commonly studied in experimentalpsychology provide a deeper explanation of preferential choicesthan the descriptive random utility models commonly studied in the appliedchoice literature, with a specific focus on choices made under time pressure.In two preferential choice scenarios we used two standard manipulations oftime pressure to assess whether changes in decision time affect subjective valuationsof the features of preferential options, and whether the answer to thisquestion depends on the theoretical lens used to understand the data. Acrossfour experiments, we found that choices between preferential options and thetime taken to make them varied as a function of time pressure, reinforcingand generalising findings from lower-level perceptual decisions to higherlevelapplied choices. Critically, theoretical analyses from the two traditionsled to different psychological conclusions about how people adapted to timepressure. The random utility analyses suggested that time-pressure inducedchanges to choices were the result of changes in subjective valuations for thefeatures of preferential options. However, the cognitive process analyses attributedtime-pressure induced changes to choices to differential informationaccumulation; subjective valuations remained stable across contexts, againreinforcing decades of findings from the perceptual decision literature. Weargue that cognitive process models provide mechanistic explanations of theprocesses underlying decisions for preferential options. Furthermore, conventionaldescriptive models of choice in the applied literature may misattributevariability in choices to the incorrect latent cause, which has implications forout-of-sample prediction in the marketplace.