The problems with bias and reproducibility can be very problematic. When I read through papers I have the tendency to view the research as “correct”, even though there’s the chance that there are flaws within the article. Analyzing each paper to the depth required to thoroughly go over the material can be a huge time sink, especially if the paper covers topics, experimental methods, or data analysis in which I am not familiar with. Therefore, I rely on the journals to do their utmost and ensure that published data has met standards that allow me to trust the research.
It is apparent from the various articles that the current method of research review is not working. Since these papers are likely to form some of the foundation of my research in the future, I believe that it is important to to find better ways to reduce the amount of bias and error in publications. Therefore, I was intrigued by the solution offered by PubPeer in which papers can be publicly commented on.
PubPeer’s approach is to provide a public forum where people can anonymously comment on any publication. This opens up the peer review process, although posted after the paper has been accepted, to a substantially large amount of reviewers. In theory, with all these fresh eyes on a paper any potential errors which slipped through the first review can be found and subsequently addressed. While I like the idea, there are a few considerations I think need to be addressed to improve the effectiveness of this process. I would be to interested to learn which papers are getting the bulk of the criticism. As flawed as the current model might be, for submission to most reputable journals, the papers must go through some type of review process, and depending on the journal, which papers. Here it seems that not all papers have an equal chance of getting reviewed, with most likely a small number of high profile papers consuming the highest amount of traffic. This also introduces some bias into the system, as the users chose which paper they wish to review, instead of it being assigned in a random fashion.
Although it does have its own flaws, entities like PubPeer can help to ensure that more extensive reviews, either conducted by the journal or other entities, help to reduce the amount of errors that get published, which leads to more confidence in accepting pertinent research findings.