Open Science can improve the quality of research
Professor Gerald Schneider talks about his experience with Open Science
Three key learnings:
- A reproducible and robust research finding can enable sound discourse with policy and improve policy in the long run.
- Open Science can improve the quality of research and render policy consultants less vulnerable.
- Open Science makes life as a scientist more interesting because it enables exciting and deep discussions with researchers from other disciplines.
You are a great advocate of evidence-based politics. What is your experience of the interaction between science and politics?
GS: Of course there are ministries with sufficient numbers of trained staff who can verify the robustness of research and who know the reputation of individual researchers and institutes. But in general the interaction between politics and science is difficult. Politicians find it hard to go against their own preconceived notions or their own clientele. If a researcher presents a finding that doesn’t fit the political reasoning it can get rejected. In general there are ministries and policy areas which are more science-oriented, among them development and economic policy. But with a reproducible and robust research finding you can force less science-oriented politicians into a discourse; and the relevant bodies can no longer hide so well. Ultimately such discourses can change policy in the long run.
Which benefits of Open Science do you experience yourself?
GS: For me, Open Science surely brings indirect gains in reputation because I am cited more often. In some respects, Open Science also makes me a better scientist and less vulnerable. And my life has become much more interesting because I can have better discussions with colleagues and scientists from other disciplines.
Do you see this citation benefit only for publications or also when you publish research data?
GS: Research data articles get very well cited if they are done well. Sometimes even to an excessive degree. Of course it is not the same intellectual achievement as developing an original idea. But it gives visibility.
In your opinion: which way of storing, sharing and reproducing data is well suited to scientific practice?
GS: The Dataverse option is surely the way to go. The Dataverse project makes it easier to provide data to others and enables easier replication of other people’s work. Researchers, journals, data creators, publishers and also affiliated institutions gain scholarly reputation and visibility on the web. The problem with Dataverse is that some institutions can have a monopoly position.
When did you first become sensitised to Open Science?
GS: In the late 1990s, SAGE publishers contacted me because they wanted me to edit the journal European Union Politics. I was then, as I still am, in close contact with my colleague Nils Petter Gleditsch who served as editor of Journal of Peace Research. The JPR introduced a replication policy fairly early. At the time it consisted in telling authors to submit the data they had used to the journal and to publish them. I thought this was innovative and we adopted this for the journal European Union Politics. The journal has benefited from clear criteria that nobody can hide behind. The replication standards have been further developed since then, of course. We recalculate all stated results before we accept a publication.
How have you, as editor of the journal European Union Politics, developed this Open Science policy during the last years?
GS: The development has been controversial in parts. The American Political Science Association did not wish to go so far and to transform so quickly. The drivers here were researchers who work qualitatively. Their argument was that hermeneutic research stages could no longer be completely reproduced. That is partly true, of course, but to a certain extent it was a purely conservative defense. Back then, there was an initiative of qualitative researchers in the USA who wanted to adopt and even build on the standards that we had in quantitative political science and which were driven by Nils Petter Gladitsch, some other colleagues and me. But it just produced a backlash in the USA. At the time I was president of the European Political Science Association and I wrote that an American association cannot simply impose how policies in individual journals or associations somewhere in Lithuania, Germany or elsewhere should be framed. This kind of imperialism has died out since. The standards are generally accepted. There is still an outspoken minority against them. But the important journals have adopted such policies and use transparency as a competitive argument.
How did this Open Science policy affect your journal?
GS: There has definitely been an increase in the number of submissions and this of course affects the quality of the journal. In the meantime, several studies have been published that show that articles which provide replication data and replication algorithms are cited more often. Our policy can be read in the Notes for Contributors. All important journals now have an Open Science policy. Some journals have gone even further, something I don’t agree with. For instance, the American Journal of Political Science asked that reusers of data provide exact protocols of their analysis and the corresponding algorithms. For me, a precise indication of the source in the usual sense is sufficient. The excessive robustness checks which journals demand because of extremely critical reviews consume too much research time. It would be more productive to conduct debates about research findings in reactions to important findings, rather than in endless annexes which nobody reads.
Have you become more strict as an editor over the course of years?
GS: Data must be uploaded unless there are legal reasons for not doing so. I do not deviate from this, but generally we must and should apply stringent standards because there is pressure from the public.
Is societal impact an incentive for researchers?
GS: Yes, but there are lots of different things that make people tick. My experience has often been that when economists apply for a job the only criterion is where they have published, in top journals or not. On the other hand you have a lot of people who want to shape or influence policy. In my opinion, both matter. I want to be taken seriously as a scientist. But at the same time I want some of my findings to illuminate political or social debate and to contribute to reforms, if possible. The fact that in job application processes the impact factor plays such a large role is of course owed to conservative incentive structures in our business. The creation and provision of datasets should be compensated in similar ways, maybe not exactly like an article in a top journal, but still as an intellectual contribution and service for the entire discipline.
Where do you see the responsibility of economists towards economic research and towards society?
GS: We need transparent researchers who document in detail how they arrived at a certain research finding and a specific statement, so that it is possible to comprehend which interpretations agree with the statements and where the authors are possibly speculating or affected by prejudice. This applies not only to researchers from universities and research institutes, but also to applied research as it is done in ministries.
What is your estimate of the future importance of Open Science?
GS: It will become Normal Science. And that’s a legitimate demand. But everyone also needs to be clear that Open Science needs infrastructures which cost money. Infrastructures for publishing and for managing and distributing research data. There must be no data piracy.
The interview was conducted on March 29, 2021.
About Professor Gerald Schneider
Gerald Schneider has been Professor of International Politics at the University of Constance since 1997 and editor of the journal European Union Politics since 2000. His main areas of research are decision making in the European Union, the causes and consequences of armed violence, trade, migration and financial crises. Gerald Schneider was president of the European Politcal Science Association from 2013 to 2015, and vice-president of the International Studies Association from 2003 to 2004.