This review was solicited as an autobiography. The “problems” in my title have two meanings. First, they were professional difficulties caused by my decision to study oviposition preferences of butterflies that were not susceptible to traditional preference-testing designs. Until I provided video, my claim that the butterflies duplicate natural post-alighting host-assessment behavior when placed on hosts by hand was not credible, and the preference-testing technique that I had developed elicited skepticism, anger, and derision. The second meaning of “problems” is scientific. Insect preference comes with complex dimensionality that interacts with host acceptability. Part Two of this review describes how my group's work in this area has revealed unexpected axes of variation in plant–insect interactions—axes capable of frustrating attempts to derive unequivocal conclusions from apparently sensible experimental designs. The possibility that these complexities are lurking should be kept in mind as preference and performance experiments are devised.