You are currently viewing the summary.
View Full TextLog in to view the full text
AAAS login provides access to Science for AAAS members, and access to other journals in the Science family to users who have purchased individual subscriptions.
More options
Download and print this article for your personal scholarly, research, and educational use.
Buy a single issue of Science for just $15 USD.
Summary
Over the past 2 years, increased focus on statistical analysis brought on by the era of big data has pushed the issue of reproducibility out of the pages of academic journals and into the popular consciousness (1). Just weeks ago, a paper about the relationship between tissue-specific cancer incidence and stem cell divisions (2) was widely misreported because of misunderstandings about the primary statistical argument in the paper (3). Public pressure has contributed to the massive recent adoption of reproducible research tools, with corresponding improvements in reproducibility. But an analysis can be fully reproducible and still be wrong. Even the most spectacularly irreproducible analyses—like those underlying the ongoing lawsuits (4) over failed genomic signatures for chemotherapy assignment (5)—are ultimately reproducible (6). Once an analysis is reproducible, the key question we want to answer is, “Is this data analysis correct?” We have found that the most frequent failure in data analysis is mistaking the type of question being considered.