the Evil Line

Give coffee addicts coffee and then have them do a challenging task. Tell half of them that caffeine is known to improve scores on the task, and tell the other half that it's known to have a negative effect. Independently, half get decaf and half get regular, but they all think they're getting caffeine.

Finally, no-one who got the decaf noticed that it didn't actually contain caffeine, and the volunteer's ratings of their alertness and mood didn't differ between the caffeine and placebo groups. So, this suggests that if you were to secretly replace someone's favorite blend with decaf, they wouldn't notice - although their performance would nevertheless decline.

Caffeine's not just a placebo, it does affect their scores - but the subjects don't notice it.