[greygoo222]

This can't explain the data in the article, such as the fact that people underestimate the rate at which other survey responders will report being happy.