Digging into the study now. It was very small: the subjects were 16 office workers. They analyzed between-subjects rather than within-subjects, which seems a li

Digging into the study now. It was very small: the subjects were 16 office workers. They analyzed between-subjects rather than within-subjects, which seems a little suspicious given that within-subjects is more interesting and given Wansink’s history of p hacking. Finally, the results comparing visible vs non-visible food (the part alluded to in the HBR article when it says “seeing is eating”) had a p-value of 0.03, so they barely reached significance. My gut says this study is not likely to replicate.