Non-Standard Errors - new publication by Olivier Scaillet
Researchers need to make many decisions when testing hypotheses on a particular sample — pick an appropriate measure, treat outliers, select a statistical model, etc. If researchers are not perfectly aligned on these decisions, their estimates are likely to differ. This potential dispersion in estimates therefore adds uncertainty to an estimate reported by a single team, as other teams might have reported other estimates based on the same data.
In a new study, GFRI’s Professor Olivier Scaillet and his co-authors claim that EGP variation across researchers adds uncertainty — nonstandard errors (NSEs). They study NSEs by letting 164 teams test the same hypotheses on the same data.
NSEs turn out to be sizable, but smaller for more reproducible or higher rated research. Adding peer-review stages reduces NSEs. They further find that this type of uncertainty is underestimated by participants.
The paper was published recently in The Journal of Finance.
A video on the project is available here >
Apr 29, 2024