Also found by the study: people choose the food with the highest expected caloric content when hungry (actually the #1 correlate, ahead of expected tastiness), which seems unsurprising. The paper doesn't seem to have any way of dealing systematically with obvious correlations (or anti-correlations) of perceptions of tastiness, healthiness, and expected caloric content, or the obvious theory that people tend to choose the food that they think has the highest caloric content when they are hungry. The entire paper seems statistically perilous.