Editor Jordan Anaya
James A Heathers, Jordan Anaya, Tim van der Zee, Nicholas JL Brown
http://dx.doi.org/10.7287/peerj.preprints.26968v1
Jun 16, 2018 - This paper presents a simple algorithm for reconstructing integer data from summary statistics (mean, SD, sample size). The algorithm is heuristic, and is not guaranteed to find a solution if there is one. This technique complements CORVIDS, which is not only guaranteed to find a...
Statistical infarction: A postmortem of the Cornell Food and Brand Lab pizza publications (2017)
Jordan Anaya, Tim van der Zee, Nick Brown
http://dx.doi.org/10.7287/peerj.preprints.3025v1
Jun 16, 2018 - Despite making numerous excuses for why the pizza data could not be shared, the data was eventually posted publicly. Given the interest in these papers I felt it was necessary to perform an independent reanalysis, and it was also instructive to see whether the techniques used in &...
Tim van der Zee, Jordan Anaya, Nicholas J L Brown
http://dx.doi.org/10.7287/peerj.preprints.2748v1
Jun 16, 2018 - This article is a good paradigm for how to perform a detailed investigation of suspected errors/misconduct. After the senior author announced on his own blog how much p-hacking was involved in the generation of these articles we were interested to take a look at them, since in our ...
http://www.ncbi.nlm.nih.gov/pubmed/23982243
May 13, 2018 - This article describes two case where fabricated data was suspected based upon the summary statistics alone, and eventually confirmed. There's an important distinction to make between the technique in this article and granularity testing. Granularity testing seeks to identify erro...
7 comments on PubPeer • http://www.ncbi.nlm.nih.gov/pubmed/28580651
May 12, 2018 - The method presented in this paper is not new, but received a lot of attention because it was applied to a large number of clinical trials and implied some of them could possibly be fabricated. I believe the Carlisle method has multiple problems, which results in both false positiv...
Complete recovery of values in Diophantine systems (CORVIDS) (2018)
Sean Wilner, Katherine Wood, Daniel J. Simons
http://dx.doi.org/10.31234/osf.io/7shr8
May 11, 2018 - This article presents a deterministic method for reconstructing discrete, bounded data. Previously the only methods for doing so were by brute force which quickly becomes untenable as sample size or scale increases, or by a heuristic method (SPRITE) which only gives 1 of potential...
http://dx.doi.org/10.7287/peerj.preprints.2400v1
May 10, 2018 - In this publication I extended granularity testing to variances, and as a result standard deviations and standard errors as well. Extending granularity testing to variances was nontrivial, and in fact wasn't thought possible. Essentially I identified that variances of discrete dat...
Nicholas J L Brown, James A J Heathers
http://dx.doi.org/10.7287/peerj.preprints.2064v1
May 10, 2018 - This article introduces the concept of granularity testing, which is the fact that a statistic of discrete data will also be discrete, and thus only be able to take certain values. The implication of this is that summary statistics can be checked to be mathematically possible if t...
2 comments on PubPeer • http://dx.doi.org/10.1101/179135
May 12, 2018 - This paper was sent for review at PeerJ, and I was one of the reviewers. The paper is apparently still under review, but I took the unprecedented step of publicly posting the reviews: https://medium.com/@OmnesRes/i-knew-biorxiv-wouldnt-post-my-peer-review-and-that-s-the-problem-cc...
Seeking and reporting apparent research misconduct: errors and integrity (2018)
E. D. Kharasch, T. T. Houle
http://dx.doi.org/10.1111/anae.14147
May 12, 2018 - This publication takes issue with Carlisle's analysis of 5,087 trials, and the accompanying editorial. A reply by the authors of the editorial was posted: https://onlinelibrary.wiley.com/doi/full/10.1111/anae.14170 Essentially, we have a scientific dispute and I'm here to resolve i...