The biggest objection I have to self-reported studies and the conclusions drawn from them in articles like this is that, according to research in books like Thi

The biggest objection I have to self-reported studies and the conclusions drawn from them in articles like this is that, according to research in books like Thinking, Fast and Slow, a lot of self-reported facts (including beliefs) are often wrong. Revealed preferences will often show that people believe something different from what they report. And people will fervently believe they told you the right thing. Also, a lot of the findings of these studies end up not being repeatable or have very small effects. I've gotten to the point where I unconsciously put little value in opinion polls and other self-reported data, unless I have a good reason to think otherwise.