You are here

Data and reproducibility

Journals should include policies on data availability and encourage the use of reporting guidelines and registration of clinical trials and other study designs according to standard practice in their discipline

Our core practices

Core practices are the policies and practices journals and publishers need, to reach the highest standards in publication ethics. We include cases with advice, guidance for day-to-day practice, education modules and events on topical issues, to support journals and publishers fulfil their policies.

The incomplete systematic review


A systematic review on the effectiveness of a comparatively new group of drugs was submitted. The review had originally been for an independent body, so the submission was an abridged version. A reviewer pointed out that the review made no reference to a Cochrane review and the trials it cited, which had been published some four months before submission of the paper to the journal.


A paper which discloses confidential material


In March 2000 author A submitted a research letter to journal X, on behalf of a national screening programme. He also submitted a commissioned editorial to journal Y, relating to the same subject. At the same time, author A sent copies of both articles to B, a recognised authority on the subject. He made it clear that they were confidential and in press and asked for some information on a test used by B which he could include in the editorial.


The wrong standard deviations, the over stringent selection criteria, and the overt attempt at advertising


A randomised controlled trial raised three aspects of concern: 1. The participants’ physical characteristics at entry to the study were listed in a table. For the two groups—intervention and control—one physical characteristic was given as a mean ± the standard deviations (SDs). However, the SDs for both groups were much smaller than they should have been. 2. The inclusion criteria were unusual. These excluded half of the eligible population. 3.


The study that may or may not already have been published


A study purported to have been stimulated by a systematic review that had already been published in the journal. The new study included 15 patients who had been treated in one arm of a study and 15 who had been treated in another arm. The peer reviewers noticed that the original systematic review included 31 patients from the same authors. The editor contacted the authors asking them to make clear whether this was a new study or a presentation of existing data.


The single author, randomised controlled trial


After a randomised controlled trial from a single author had been published, a letter was received in which the correspondent suggested that the original trial might be fraudulent. Firstly, the writer claimed that it was highly unlikely that just one author could perform a prospective, randomised, double blind, placebo controlled trial, especially in a small district hospital. The correspondent was also worried that there was no mention of other standard treatments.


The overlapping papers with conflicting data


Three papers concerning one hospital problem had been submitted to three different journals. Before publication the three editors of the journals became aware of the three different papers and the substantial overlap between them. The three editors communicated with each other and realised that they had four concerns: 1. There was very considerable overlap among the three papers. There didn’t seem to be any justification for publishing three papers rather than one or two. 2.


Author dispute concerning ownership of data


A paper submitted to Journal X was reviewed and rejected with the recommendation that it be submitted to a more clinical journal. The paper was duly submitted to Journal Y. The authorship was A, B, C, D and E, with E being the corresponding author linking together two research groups in different cities, but in the same country. Journal Y sent the paper to reviewers and, after discussion, their decision was to open negotiations.


Misconduct on a massive scale?


Almost five years ago two outsiders approached an editor suggesting that a large series of papers from a particular researcher, including some published in high profile journals, might be fraudulent. Those contacting the editor thought it possible that the patients described in the studies had never existed at all. Round about the same time a few papers from this author were circulating in the journal’s peer review system.


The results that were too good to believe


A study made it a long way through the peer review process before one of the statistical advisors said that the results seemed “too good to be true.” The authors were asked to send in the original data, which the statistician analysed. He remained very concerned about the data. The authors were notified and the journal asked the university to investigate. Has the editor done the right thing?


An anonymous letter in response to qualitative research


Some two months after publishing a piece of qualitative research about health behaviour in an ethnic minority group, an anonymous letter suggested that the work might be fraudulent. The letter was in very poor English, but made two main points. Firstly, the original study did not make clear how many women were included, and secondly, the anonymous respondent could not understand who could have done the interviews.