Research seminar: Tom D. Stanley (Deakin University, Australia)


Research seminar: Tom D. Stanley (Deakin University, Australia)

We would like to invite you to a research seminar by Tom D. Stanley (Deakin University, Australia). 

The seminar will take place on Friday, April 19, in Room 206 at 12:30.
Paper: Selective and Mis(leading) Economic Journals
Abstract: Over the last several years, we have been conducting large surveys of research across the disciplines and developing the underlying statistical methods needed to measure typical statistical power of published research, research inflation (or exaggeration), and other metrics of publication bias and misleading evidence. In this seminar, I will summarize the findings of several of these meta-research surveys and briefly sketch the methods upon which they are based.
Take away:
- Statistical power is very low across the disciplines but lowest in research published in the leading economics journals.
- The typical empirical effect reported in the leading economic journals is exaggerated by at least 100%. Divide any economic finding by 2, and you will be much closer to the magnitude of the underlying effect being studied.
- At least half of the ‘positive’ findings reported in leading economics journals are ‘falsely positive evidence.’
- Economics research is more exaggerated, on average, than medical research, psychology, or environmental science.

Fortunately, simple changes to journal policies can make economics science better. We have shown this to be the case for mandatory data-sharing policies.
The talk will cover results from the recent papers, including:

If you need a discussion slot with Prof. Stanley, please write to Zuzana Havránková, and she will try to find a suitable time.