Most consumers, given two identical food samples, express a preference for one, rather than choosing a no-preference option. The stability and potential causes of this seemingly irrational preference were examined across three trials under different conditions, specifically, when the first test pair was identical or different, and when participants were explicitly told that the pairs would often be identical. Choice of no preference typically increased from the first to second trial, especially for groups who saw a pair of different samples on the first trial. The explicit instruction that samples might be the same failed to reduce expressing a preference on the initial trial although it had some effect on later trials. Analysis, by individuals, of sequences of preference or no-preference responses across trials support independence of sequential responses and argue against stable personal traits as a predictor of preference choice.


More research needs to be done to understand the origins and operation of biases in preference tests. When tested under conditions in which the samples differ only slightly, participants tend to avoid the no-preference option. This is potentially important when interpreting the results of preference tests and assigning practical significance to their outcomes. Also, single trial testing may produce somewhat different results from multi-trial testing, the latter allowing for examination of effects of variation in recent experience.