0
Dependency Accounted for?

I wonder if in your extraction of APA statistics, did you take into account the fact that there may be dependency in the data that could bias (in either direction) the rate of 'bumps' in reporting of p-values?

That is, each article is written by particular authors, therefore all reported statistics within an article are subjected to whatever research philosophies, possible QRPs, etc. that the team of authors bring to it.

Would it not be best to randomly select one APA statistic from each article to create your data? You could then bootstrap multiple extractions from the articles resulting in a potentially more robust test of your hypothesis.

waiting for moderation
1 Answer
0
Accepted answer

Thanks for your question Paul. We looked into the dependency in another paper that uses the same dataset, which you can find here. Your approach is different from the one used there (i.e., the ICC).

waiting for moderation