…with his survey of economists.
Fortunately, unlike a lot of surveys, he gives us some data on how the survey was done.
All individuals on the AEA’s opt-in list were asked to participate in the survey. Therefore the respondents were not “sampled” in a statistical sense. Those who chose to participate were included among the respondents.
So the respondents were self selected. This means the survey is worthless.
With that said, it is helpful to understand the rules for statistical significance that would apply to random samples of similar sizes, as a guideline in understanding the importance of differences among various subgroups of the data.
In other words, “we know the survey is worthless, but here’s some statistical mumbo jumbo to make it look better.”
The bottom line is that this survey only tells us what the people surveyed thought, and can’t be relied on as a indicator of how the set of all economists think.