05 March 2009

Implicit Bias?

I was just playing around some more with the implicit association research site, and got one that gave information about a group of people and one particular member of the group. There were two implicit association tests: one for the individual and one for the group. Watching my own responses, I came to the conclusion that, on the first one, I had trouble on the switchover not due to any actual associations, but because the task was slightly different from the others I'd done. There was only one extra category besides "good" and "bad". On every other test I'd come across, there were two categories to sort through. So on the first switchover, it felt awkward. On the second, it felt perfectly natural, as I now had practice at it. However, the effect was apparently slight:

Your data suggest a slight automatic association of (named individual) with Bad.
Your data suggest a slight automatic association of (named group) with Good.

I'm not including the names as it does not seem appropriate for me to do so. As far as I could tell, they were simply made up strings of random syllables, but perhaps there was some intended significance in the syllabic choices.

Final thoughts: I hope that they don't always start this particular setup with the individual in the same category, as I suspect that most people will find the first changeover awkward and this awkwardness would introduce bias.

No comments: