[astrobase_go]

I think what's needed is a strategy for making actual data analysis an easier and more visible component of the peer-review and publishing process. Submissions are typically sent to three people for review, and the decision to accept or reject is made based on the feedback or criticism given. If the reviewers are busy, overworked professors, they may not have the time to really perform a deep dive into the data and conduct an independent analysis. Furthermore, it's likely these submissions don't include enough data for a proper statistical review.

This is the 21st century, where most if not all reputable journals have an online presence and submission portal. It would be great if authors had to upload their anonymized data sets in a common format (.csv?) that can easily be imported into statistical analysis packages (ex: Minitab, read into R, etc). Journals provide analysis and test recommendations, reviewers run the tests, upload their independent analysis results as part of the review process. The idea here is to produce something like an auditable paper trail.

There has to be some sort of solution for this problem, especially in 2017. Hiding behind shitty (or worse, deceitful) data analysis shouldn't be possible.