If you saw a chart that showed that the rate of left-handedness had significantly increased around a certain time, you might reach very different conclusions based on whether or not you understood the mechanism behind it.
First: “the map is not the territory.” When you’re looking at any kind of chart, data, etc., always remember that you’re looking at what was recorded, not what necessarily was. If you need a clearer example: Imagine a survey where the question was: “Have you ever committed a felony and gotten away with it?” 2% of respondents answer “yes.” Would you conclude that only 2% of people had gotten away with felonies? There’s always bias in the way people answer.
Which brings us back to left-handedness. Did people suddenly start being left-handed? Or did being a southpaw stop being stigmatized so people were more comfortable reporting that they were?
Don’t assume the mechanism is what it appears to be. People’s incentives to tell the truth are always a factor.
The Opportunity Machine 