Analysing all that data, we noticed that it is fairly common to see many articles being retracted from journals on a yearly basis; this not only happens in Open Access publications, but sometimes in subscriptions too. We want to find out whether this is just a coincidence - something that is occasionally supposed to happen in the research world - or if this reflects the peer review problem of OA.
Scopus provided us with a useful, sortable database with the full list of retracted articles. We refined our research and looked for retracted articles in 2015 and downloaded the list. Then, thanks to Excel and Google Refine we fixed and adjusted our file, coming up with 281 total articles. Finally, we divided them into subscription and Open Access, then grouped by the given reason for the retraction. The categories were: ghostwriting; based on a retracted article; wrong citation, when one or more articles are cited wrong; falsified/fabricated data, when data is found to be made up; missing approval on data, when the article cites data even when its use was not approved; inability to replicate, when an experiment doesn't give the same results twice; authors reason, which is a personal matter and thus hard to understand; error, which includes all sorts of general mistakes, like grammar; duplicate, when said article was already published elsewhere; compromised peer review; plagiarism; wrong research or inconclusive; unknown.
There are three different visualisations.
The first one is a ring with the distribution between Open Access and subscription articles in order to compare the phenomenon in the two publishing formats. The second one is a radar counting the number of articles retracted for each of the fourteen categories. It is possible to filter the results by OA and subscription based publications. The third and last one is a bar chart highlighting differences between the four main publishers involved: BioMed Central, Elsevier, Springer and Wiley take 200 out of 281 total articles.
It is apparent that most articles are retracted because of problems directly linked to peer review, which can be wrong or done poorly, not mentioning the possibility of journals not doing any kind of review despite declaring so. The problem also appears to be linked to Open Access rather than to subscription journals, especially considering that subscriptions articles are way more than OA's. This confirms the peer review is still a huge barrier in Open Access.
