Co-pilot approach to data analysis

Errors in statistical analyses are common and are of a varied nature. For example, analyses may not fit the type of data, intended analyses may have not been carried out adequately, results may not have been reported adequately, or results may not have been interpreted adequately. Such errors may be less likely if four, instead of two, eyes are involved in analysis and reporting, the so-called co-pilot approach to data analysis. The purpose of this document is to highly recommend a co-pilot approach at the UvA Psychology department. A co-pilot approach should increase the likelihood that conclusions are supported by the data, thereby facilitating scientific progress.

Published by  PRI

21 September 2017