Why articles are retracted: a retrospective cross-sectional study of retraction notices at BioMed Central

There is no "Method" section in the abstract. On should be added. Response: The abstract follows BMJ Open guidelines and there is no method section (the study design is reported under the design heading). Objectives A sentence summarizing what the COPE guidelines recommend regarding retraction notices should be added in the objective to make the result section understandable to the non-specialist reader. Response: The abstract has a word limit (300 words) We have provided this in the first line of the introduction on page 3, and expanded this as you suggest below. Results "Although they account for 0.07% of all articles published": It is unclear from the sentence which was the denominator: original research articles? All articles including commentaries letters and editorials? A sentence such as: the rate of retraction is rising should be accompanied by numbers, if not by statistics. Response: Thank you, this was an oversight. Between January 2000 and December 2015, our search identified 134 retraction notices. This accounts for 0.07% of all total articles published (190514 excluding supplements, corrections, retractions and commissioned content). We have added this to the abstract and the results section. Introduction: Page 3 Lines 30 to 35: The COPE recommendations listed here are incomplete. COPE also recommends for example to identify the name of the authors and the title of the retracted article in the retraction notice heading, and also that the notice should be clearly linked to the retracted article in all electronic sources. Response: Yes. We have now added the following sentence: “In addition, retractions should be clearly identifiable; freely available; published promptly and be linked to the original article that is retracted (which should also be identified as a retraction).” In the discussion and conclusions we also clarify that in these respects BioMed Central adhered to COPE guidelines and also discussed issues where relevant information was unfortunately not provided. Line 37: what is the difference between "misconduct" and "fraud"? These terms should be clearly defined since one of the major problem that research on misconduct is facing is the lack of clear definitions! Response: Thank you, the ORI definition of misconduct is fabrication, falsification or plagiarism. We are using fraud to refer to fabrication or falsification. We have clarified the terms in the methods. As you point out, because of ambiguity in the literature about what constitutes misconduct – and whether plagiarism is research misconduct or publishing misconduct we now simply describe the individual reasons for retraction (plagiarism, duplicate publication, image duplication etc) and only broadly classify retractions into three general categories of misconduct, honest error or unclear as defined in group.bmj.com on November 26, 2017 Published by http://bmjopen.bmj.com/ Downloaded from