2 ms·
Even though you asked everybody to submit answers, you should expect people who expect themselves to do poorly to opt out. So your sample, I suspect, is strong
by tmoertel 14y ago
Even though you asked everybody to submit answers, you should expect people who expect themselves to do poorly to opt out. So your sample, I suspect, is strongly biased toward the answers from people who expected not to do poorly. And that bias is indeed likely to paint a rosy picture of the reality.
- cperciva 14y agoQuite true -- I'll definitely have to be careful in how I interpret the data. I'm hoping that the issue I'm most interested in -- comparing how people do on different questions, and comparing how different people do, will be less strongly biased by the poor sampling method.