Paper about "Non-Standard Errors" published
From the Institute of Management Research, Utz Weitzel (Project team) and Sascha Füllbrunn, Sven Nolte, and Stefan Zeisberger (Research teams) were involved in the FinCap project paper "Non-Standard Errors". "It was a great experience to be part of this crowd analysis project", said Füllbrunn.
164 research teams tested 6 hypotheses on the EUROSTOXX 50 index futures data with 720 million trade records spanning 17 years of trading. The results show that standard errors (measuring the uncertainty in sample estimates of population parameters) occur in the data generating process, while non-standard errors (research teams make different choices) occur in the evidence generating process. To take such non-standard errors into account, researchers could double standard errors when computing confidence intervals.
The paper "Non-Standard Errors" can be found here.