State-dependent memory effects using caffeine and placebo do not extend to metamemory

The authors examined the impact of caffeine on human memory and predictions of memory (i.e., metamemory). On Day 1, 83 college students drank a sweetened beverage containing either caffeine (4 mg/kg body weight) or a placebo before they studied 40 pairs of words. While the participants studied, they predicted their future memory performance for each word pair. On Day 2, the participants again received caffeine or a placebo before the memory test. The participants who drank the same beverage on both days (either caffeine or a placebo) recalled more word pairs than did those who drank different beverages (caffeine on 1 day and a placebo on the other day). In contrast, memory predictions were more accurate when the beverages did not match on both days. These data provide evidence for state-dependent memory when caffeine is used, but not for state-dependent metamemory. People's memory and their predictions of memory can be influenced in different ways if they drink caffeine before they study or take a test.