The Denver Police recently conducted a study of racial bias using a videogame. The idea was to determine if officers were more likely to fire on assailants with firearms if they were black or white. The game itself is rudimentary and somewhat disturbing. Here's how it works:

The game clearly and weirdly encourages shooting rather than holstering. A Denver radio news program covered the study, and you can read the full report as well (PDF). In the broadcast, you can tell that the police department considers the result a success.

As Josh Fishburn pointed out to me, the game seems to measures reaction time in milliseconds despite being built in Flash and runing at a fairly low framerate. In his experiments with the game, Josh discovered that just mashing the holster button resulted in a measured deviation of around 25ms, which makes one somewhat chuckle at the study's detailed discussion of response times.

Thanks to Devin Monnens for bringing this to my attention.