After I found out about the panel discussion on Good Scientific Practice at the University of Cologne via Twitter, I joined yesterday to watch the discussion as it was closely related to my thesis’ topic.
The panel was filled with five professors and one junior professors from different faculties1, whose positions were related to “good vs bad science” in some way.
Asking about causes and effects of bad science, in terms of breaching rules of good scientific practice, is very complex and there is only little theoretical or empirical research on this matter. Thus, ninety minutes are very short to discuss this topic anyway.
However, the panel instantly reached a consensus, that there are problems, but everyone seemed anxious to note that these are only rare, individual cases. A meta-scientific discussion on structural reasons at the university was not achieved and probably not desired.
As one could expect from the setting2 there was a focus on graduate students and their scientific research. A lot was talked about those junior scientists while some in the panel reminisced about the “good old times”. It was stated that the requirements for a doctorate did not change over the last years – only the circumstances may have changed, e.g. through the digital world, in contrast to 30 years ago.
There was no real discursive discussion, despite occasional attempts to find a counterposition.
At the University of Cologne many graduate students participate in graduate colleges, which is a big difference to many graduate students in Bonn and my personal experience. This led to some discussion on these Graduate Colleges and institutions such as a Thesis Advisory Committee, that I am actually unfamiliar with, among the Panel.
The core topics, in my opinion, such as reproducibility, transparent reporting of methods, publishing (raw) data or p-hacking were only briefly mentioned. Everyone agreed that students should learn from early on, what it means to conduct research cleanly and truthfully. However, specific measures – apart from using software to detect plagiarism already in seminar papers – to teach and support undergraduate and graduate students were not proposed by anyone. In fact, journals and publishers were called to introduce stricter rules for publishing.
The event nearly became entertaining, but in the general consensus quite pale.
However, there was one issue, that I consider important, both for my own work, but also for science in general:
Science often means failing and defiance. […] The brilliant idea does not lie on the street – and if so, than it wasn’t that brilliant.
Science likewise is to follow paths, that render unfruitful. This is part of the process and can also be a part of a dissertation. Supervisors should support their students in their work, even if this may not lead to a publication in a highly ranked journal. Research endeavours like this will most likely not lead to a publication at all, but a dissertation should also contain “non-findings” like this. I consider this an important statement, but I presume that this is not a widespread mindset.
In the open round of questions after the discussion, mostly Professors and Post-Docs participated. The last question, however, coming from a graduate student, was notable and highlighted the dependency of students to their supervisors: if a supervisor decides an experiment has to be performed in a particular way, you may not be able to express concerns. A culture of bad science might be continued that way. I guess this is quite the common occurrence.
The answer of the panel to this only considered the institutions where students could, maybe, eventually have an effect on the supervisor. An idea discussed earlier, splitting supervisors and examinations, was not further regarded in this – probably because everyone was longing fur the buffet.