Q&A Felicitas Hesselmann: Vague and varied retractions point to weakness in the scientific community
The scientific method is not applied in dealing with aberrant research behaviour.
1 May 2018
It appears as a terse statement — a few hundred words invalidating the claims made in a published paper. Retractions are the prevailing method for correcting gross inaccuracies in the scientific literature, but they are also considered a form of punishment for authors. Sociologist Felicitas Hesselmann at Humboldt University of Berlin investigated the punitive role of retractions by analysing 127 notices published in three bibliographic databases. She spoke to Smriti Mallapaty about her findings.
Why did you take a criminological approach to analysing retractions?
Studying how a community relates to transgressive acts can tell you a lot about its values and norms, as well as the power structures within it. Our group, led by Martin Reinhart at the German Centre for Higher Education Research and Science Studies, wanted to find out how the scientific community detects, reacts to and punishes misbehaviour.
There are many ways that the community addresses misconduct, from formal measures, such as firing a researcher, to informal responses such as gossip. Informal approaches can sometimes be more effective at safeguarding specific norms than official sanctions or punishments. Research is only beginning to look at this type of behaviour.
We decided to focus on retractions because they are public and easily accessible. Over time, I became fascinated with them. Retractions are weird. Scientific language usually employs very exact numerical quantities to refer to things. But retractions are often written in a way that makes it difficult to determine what is actually being said. They use a lot of evasive strategies, and the text often hints at uncertainty and doubt. My analysis investigated the linguistic strategies used by journal editors and authors.
What did you find through analysing retraction notices?
It is very difficult to determine why a paper was retracted — whether it was for misconduct or an honest error. The notices don’t consistently differentiate between the two. There is no firmly established scientific method for investigating scientific misconduct. Notices don’t clarify how the error was detected, how it was investigated, and who is retracting the article.
Blame is also inconsistently meted out in retractions. Sometimes individuals are severely stigmatised, while at other times the notices are mindful of protecting the reputation of implicated researchers. Authors are often blamed, but there is also some responsibility placed on journals. For example, a lot of retractions contain apologies from authors and journals, which is a sign of blame.
Everything is determined on a case-by-case basis. Contrast this with legal trials, where there are very clear rules and very clear expectations on how to assign blame, responsibility and guilt. In the sciences, there are no such clear expectations or rules for assigning blame, which makes it difficult to predict how a case is going to turn out.
What does this analysis say about the scientific community?
The anthropologist, Mary Douglas, led seminal work into how communities and cultures deal with transgressive behaviour. She found that the only communities that have such inconsistent patterns of assigning blame are those that are very weakly organized.
Retractions are a relatively new phenomenon. A lot of the actors involved are unclear about how they are expected to handle these investigations, and are having to figure everything out for the first time. Most of them deal with cases only occasionally, on top of their professional responsibilities as editors, researchers, professors or teachers. This places a lot of responsibility on the individual.
Given the ambiguous nature of retraction notices, do they serve their intended purpose?
That is a tricky question, because what purpose are they intended for? The Committee on Publication Ethics guidelines for retractions specify that notices should clearly state the reasons for retraction, and distinguish between misconduct and error. But based on my observations, the majority of retractions do not clarify the science under question.
Of course, retractions can serve other purposes. They might, for example, be a means of recovering trust in the scientific community, by signalling that journals and universities are doing something about misconduct.
What do scientists feel about the current approach to retractions?
A lot of people feel that there is no due process and outcomes are erratic, some people are harshly blamed and stigmatised while others can continue to do their research. This makes it difficult to foresee what will happen. Some consider it to be an unfair system, which disproportionately blames junior researchers for misconduct, while senior researchers are often allowed to remain anonymous.
Research misconduct has probably always been an issue in science. So why is there no clarity on how to handle it?
The underlying prevalence of misconduct might be very much the same as it was in the past, but more cases are getting detected and sanctioned. All over the world, structures are being created or expanded to deal with misconduct, such as the United States Office of Research Integrity. In Germany, a system of ombudspersons was set up at universities about 20 years ago. We have only just started to talk about and openly deal with research misconduct.