36: Statistical inconsistencies in published research
Episode description
In episode 34 we covered a blog post that highlighted questionable analytical approaches in psychology. That post mentioned four studies that resulted from this approach, which a team of researchers took a closer look into. Dan and James discuss the statistical inconsistencies that the authors reported in a recent preprint.
Some of the topics covered:
- Trump (of course)
- A summary of the preprint
- The GRIM test to detect inconsistencies
- The researchers that accidently administered the equivalent of 300 cups of coffee to study participants
- How do we prevent inconsistent reporting?
- 21 word solution for research transparency
- Journals mandating statistical inconsistency checks, such as 'statcheck'
Links
The pre-print
https://peerj.com/preprints/2748/
'The grad student that didn't say no' blog post
http://www.brianwansink.com/phd-advice/the-grad-student-who-never-said-no
The caffeine study
http://www.bbc.com/news/uk-england-tyne-38744307
Tobacco and Alcohol Research Group lab handbook (see page 6 for open science practices)
http://www.bris.ac.uk/media-library/sites/expsych/documents/targ/TARG%20Handbook%20161128.pdf
21 word solution
http://spsp.org/sites/default/files/dialogue_26(2).pdf
Facebook page
https://www.facebook.com/everythinghertzpodcast/
Twitter account
https://www.twitter.com/hertzpodcast