Preanalysis plans are great way to increase transparency in science, #openscience. But what do you put in a preanalysis plan (PAP)? Initially @bnbakker and I made a lot of mistakes. We now wrote a memo on how to write PAPs. http://www.hotpolitics.eu/about-the-lab/  A thread w our view [1/8]

PAPs should facilitate the review process. Make it easy for reviewers to check and evaluate your PAP. To that end we do not advice adding lengthy theoretical sections to your PAP. That should be in the paper [2/8]

PAPs should contain clearly testable hypotheses. Seems so obvious…still we often fail to do this [3/8]

PAPs should contain exact descriptions sample, measures, procedures, analysis strategy, etc. If you use new scales specify quantifiable decision criteria. Saying it is “reliable” or “acceptable” isn’t enough [4/8]

Limit yourself with robustness checks! It escalates quickly. You can always explore extensions and checks later on. Preregister your best idea [5/8]

Discuss all the preregistered hypotheses in your article. Also one PAP per paper please, think about the reviewers. [6/8]

You can deviate from PAPs, as long as you are transparent about it. Then it is up to the reviewers to decide. When in doubt you can always publish the analyses with deviations as exploratory analyses. [7/8]

Finally, have fun! In our experience writing PAPs really helps to correct errors in our design. We have felt really silly at times! But our science has much improved (at least we hope #reviewer2 agrees [8/8]