Economics in Theory and Practice:
The American Economic Association surveys its members, badly
4 Feb 2009 in Information Quality, Amusements
Attendees of the annual American Economics Association annual meeting received an email asking them to participate in a survey:
Every five or six years AEA does a survey of the economists who attend the ASSA meetings. The survey results affect decisions concerning future annual meetings. Please help us out by taking a few minutes to complete the survey.
This "survey" produces information with little or no value.
- The "survey" is actually an attemped census, as no sampling was performed to determine which attendees to survey. A significant fraction of attendees will not respond; hence, the census will have significant nionresponse bias.
- Will nonresponse bias matter? Yes, because respondents and nonrespondents are likely to differ with respect to many of the questions asked.
- Respondents are asked to reveal if they are willing to pay $25 or $50 more per night to attend a future meeting in San Francisco (where the 2009 meeting was held) or Hawaii. The economics literature has ample evidence that the responses to such questions are invalid and unreliable.
- AEA members who did not attend are not surveyed. For that reason, the results are biased even more.
Economists often criticize government agencies when they use poor research methods. This survey would never pass minimum federal standards for survey research.
