Thursday, March 09, 2006

Naive Views on Probability

There are people who study human perception of probabilities. We are pretty bad at it, actually. That's why we cannot make random numbers. We try to balance things out. If we simulate a coin flip, we never have enough heads or tails in a row. Along these lines, we tend to "see" patterns where they do not exist.

In the lab, there always seems to be one experimenter who always gets stood up. Conversely, there also seems to be one experimenter who could not buy a no-show for a million dollars. You might think that we never want a no-show, and in the abstract, you'd be right. But in the middle of an experiment, when you've run 3 in a row and 8 in 3 days, you're just burned out. You'd kill to have a free two hours.

On the other end, sometimes you need just 5 more subjects to complete your experiment. Usually in this case, you have 7 people scheduled, and you sweat it out.

During my dissertation experiment, I was always getting stood up. However, my good friend and colleague, Sung Kyoung Lee, never had a no-show. We'd laugh about it. We formed naive views about probability. When the rare no-show happened to SK, we dismissed it as a fluke. Instead, that fluke was evidence of probability in action.

The same thing is happening in the Communication & Cognition lab. One experimenter is fuming about no-shows, whereas two others have perfect attendance records. In the end, it will even itself out, although surely there is some relationship between the times of day when these experimenters are available and when flaky people tend to sign up.

In the meantime, we will continue to blame it on the cruel hand of fate :)


Post a Comment

<< Home