[BPS Research link]

There's something strange going on here; they're claiming a huge effect size (six times the compliance rate in the letter test) after 15 minutes of gaming, which doesn't seem to pass the giggle test to me. And if you crunch the numbers on compliance, the most likely totals look to me like 6 of 31 vs. 1 of 29, so they must have some huge error bars.

I can't properly analyze this without access to the original paper, but there's definitely a suspicious odor about it. I'd like to have seen a longer-term follow-up, too.

(That said, people temporarily assuming some of the mentality of characters in media does tally with my experience. And the OSU study looks a lot better.)