The Reproducibility Crisis Why are we in this crisis?
Reasons for the reproducibility crisis
Some select phenomena that brought us here:
Dichotomy of p-values
Arbitrary .05 significance cut-off
generates a false dichotomy of
’false’ or ’true’ conclusions
Significant effect are not
necessarily biologically relevant
∼ Burnham et al. (2011). AIC model selection and multimodel inference in
behavioral ecology: Some background, observations, and comparisons.
Behavioral Ecology and Sociobiology.
Peer-review shortcommings
Reluctancy to make corrections
No clear guidelines on where to
direct criticism towards
No standard process for data and
code access
∼ Allison (2016). A tragedy of errors. Nature.
Research integrity
Research questions often
formulated post-hoc leading to
multiple testing issue
Sloppy reporting of data handling
procedures
Lack of data and code repository
guidelines
Lack in pre-specification of
research
∼ Cumming (2014). The New Statistics: Why and How. Psychological
Science.
University of Leipzig Excursions into Biostatistics 5 / 14