Add this to the shocked-I-didn’t-know-this category (or Greg’s Reference in the nomenclature I set up here), but remember the old story about experiments in the 1950s on using subliminal ads to get people to buy Coke and popcorn at the theater?
Yeah, they’re BS, according to Brian Dunning, who apparently writes out his Skeptoid rants beforehand. [Link]
Harcourt Assessment, which was known at the time as The Psychological Corporation, invited Vicary to repeat his experiment under controlled conditions. He did, but this time no increases in sales were shown at all. Pressed for an explanation, Vicary confessed that he had falsified the results from his original study. Indeed, five years later in a 1962 interview with Advertising Age, Vicary revealed that he had never even conducted the Fort Lee experiment at all. He had literally made up the entire thing. But of course, by then, it was too late. The headlines had run their course, and to this day it’s a generally accepted fact that flashing brief messages onscreen produces a desired behavior, despite the fact it never happened.
Not that I don’t trust Dunning, but I’d like to see that article. Read the entire thing for some insight on other notes on the difference between subliminal (hint: to be subliminal you can’t actually consciously perceive it) and suggestive ads.  Or listen to his podcasts — they’re brief, but very entertaining.