Dare to be transparent, manage open research

Was there a particular trigger moment that sensitised you to the topic of Open Science, or did your interest in it develop gradually?

HH: The interest developed rather gradually, partly due to the increasing importance of empirical research and the growing significance of data. Even as a doctoral student, I benefited from open data structures such as those at ZEW, where research data was collected and made available to doctoral students. This was still relatively new at the time, and it was the first time I came into contact with the use of data that I had not collected myself via scientific use files. These files are specially prepared, for example to protect the anonymity of the respondents. Such methods have developed over time and were my first point of contact with open data.

During my doctorate at a research institute at the KU Leuven in Belgium, we collected data for research projects that were to be made available as publicly accessible research data from the outset. This influences the care taken in data preparation in order to meet scientific standards. For me, such open data is an essential part of open science.

Discussion papers are another example of open science in economics. This is particularly noticeable to me at the TU Munich, as this is less common in other disciplines and preprints are often still seen as something new. In some areas of business studies, the practice is less common than in economics. For me, discussion papers were always a great way to get an early insight into current research results, as the publication processes in journals are often very long and the research results can be outdated by then. As a student, I remember that the discussion papers from economic research institutes, including those from the USA, such as the National Bureau of Economic Research, often represented the latest research.

At various universities, such as the Technical University of Munich, I have noticed that disciplines such as engineering and computer science have a much faster publication process, which is why preprints are less important there. It is incomprehensible to me why some journals reject pre-publication in the form of discussion papers. This creates situations in which research results suddenly appear as “submarine publications” without the scientific community having had the opportunity to comment on them beforehand. With discussion papers, on the other hand, there is often direct feedback from the community, which I really appreciate.

Or you get a question that provides new food for thought without being meant as direct criticism.

HH: We often experience this from other authors who point out that a relevant paper has not been cited by them. And often they are right, and we take the opportunity to improve it. However, if the article is already in the journal and no more changes are allowed, this opportunity is lost. So it’s a kind of peer review process that is not only based on the two or three reviewers of a journal, but also on a broader exchange in the scientific community.

This is precisely why I see preprints and the entire preprint culture as part of open science – a practice that is now spreading to other disciplines. I came into contact with this early on and I think its importance will continue to grow. This contrasts with the trend of large interdisciplinary journals that require research to be completely new and not even presented at a conference – such as Science and Nature. I find that strange, because it would make sense for the scientific community to know about the research in advance and be able to discuss it.

Yes, at least at a conference you should present research results for discussion and then improve the paper further instead of just presenting and then leaving.

HH: Exactly, I also find it strange. We organised a workshop some time ago. During the submission process, we once had a participant who only uploaded one graphic. When we asked if the wrong file had been uploaded, he replied that he couldn’t submit an extended abstract because he wanted to submit the paper to a particular journal later and it would be considered pre-published if it was uploaded to a conference tool. We then had to tell him that we could not accept this because our workshop policy requires full papers or longer abstracts to be peer-reviewed. The authors were not particularly happy about this, but it is in line with our policy. We then also discussed the topic critically during the workshop because we believe that science benefits when results can be discussed in advance before they are finally published.

Let’s talk about your personal experiences: Have you experienced concrete benefits from Open Science practices during your research activities?

HH: Yes, definitely. I have had particularly good experiences with discussion papers, as they lead to real discussions – their actual purpose. They offer more than a conference presentation and provide valuable feedback. I find it problematic when a discussion paper appears at the same time as the publication in the journal, as it is often too late to point out new work or exchange ideas. Discussion papers have always been important to me, and I realise that they give my research more visibility and quicker attention.

As far as the data is concerned, sharing has benefited me in several ways. Firstly, it increases the visibility of my research, as my work is cited more frequently. Secondly, it has led to new collaborations, for example when researchers carry out replication studies or build on my data and follow up with their own surveys or analyses. One specific example is a co-author who contacted me after reading my paper and had an idea for an extension for which he needed access to the data. This resulted in two joint follow-up papers. Of course, sharing the data often requires additional work as it needs to be documented more carefully. But overall I have had very good experiences, both in collaborating with others and in furthering my own research.

You mentioned that you publish discussion papers and preprints when possible. Do you also make sure to utilise secondary publication rights for journal publications and make the publisher’s version available in open access? Or does this go through your library?

HH: The topic is relatively new to me. Until recently, it wasn’t of great importance to me, as I assumed that the latest version of an article would be available as a discussion paper anyway. Therefore, I didn’t initially attach much importance to the fact that the publisher’s version is available for free on the journal website. However, with time and the increasing pressure from publishers to promote open access more strongly, I have changed my position on this. With traditional publishers charging for open access, I have recently made more use of it to ensure that my research is fully and accessibly available – even if a discussion paper version still exists. One disadvantage of Discussion Papers is that citations are often not recorded in the Web of Science or Scopus, which affects citation metrics. This might be a reason for others to reconsider this strategy, especially if you are orientated towards such metrics. For my own career, however, this is less relevant and I try not to let it influence me too much.

Back to the research data: Do you publish them at ZEW?

HH: For example. We have also published research datasets on Zenodo. In this case, we uploaded a replication dataset. For another paper, we made the complete scripts available on GitHub, with the note that replication is possible. However, the underlying data was commercial data, which is why we were not allowed to share it. However, if you have the appropriate licence, you can reproduce the results with the source codes and scripts provided. So there was no way to upload the data directly. We pointed out that a separate licence is required for the database in order to carry out the replication.

Do you use collaborative platforms when working with researchers from other disciplines in the field of innovation?

HH: Yes, we use Git, for example. We also use GitLab, which is provided by national data centres, especially when it comes to confidential data and it is important to know where the server is located. We also use various tools for internal teamwork, not only for analysing, but also for writing papers together. For example, we use Overleaf, a writing programme that enables simultaneous editing of documents, similar to Google Docs, but specifically for scientific papers. Overleaf is a very helpful innovation for us that we use on a daily basis. There are also campus licences for it, which has made collaboration much easier, especially when working remotely.

Do you have experience in interdisciplinary collaborations with a pre-agreed data management plan?

HH: A data management plan is an essential part of the project design. A few years ago, for example, we conducted a survey involving one person from Japan, one from the UK and myself. When we were designing the research, we thought carefully about how we would handle the data later on. For example, we decided whether the responses of the Japanese participants should be included in the joint data set or whether we should keep them separate. Ultimately, we made a conscious decision to merge the data and informed the respondents in advance that the results would later be made openly available as a research dataset. We also obtained consent to share the data in anonymised form. It is important to clarify this in the survey design so that we know from the outset what will happen with the data later.

You are also active in teaching. What is your policy on open educational resources? Do you usematerials from other colleagues and do you make your own scripts, teaching concepts or slides available for reuse?

HH: Yes, for example, we use the CORE-Econ material in the basic economics modules for macroeconomics and microeconomics in the Bachelor’s programme, an open access project that I have been using since the beta version [URL: https://www.core-econ.org]. A big advantage for students is that they don’t have to buy a textbook – they have free access. Another plus point is that the material is regularly updated. I have also adapted the slides accordingly every year and passed this feedback on to the organising consortium. I’ve been using CORE Econ since 2017 and think it’s great. We also use the question pools provided, for example for e-tests and quizzes, and develop our own multiple-choice questions, which we play back and share with colleagues. Within the university, we have opened up the content across all locations so that everyone can use the same materials. When I receive requests from other universities or from the community, I also share my materials, especially in the field of innovation economics. This is often an exchange of content, in which my materials or those of others serve as a basis, but are usually customised.

At Master’s level, teaching is often more individualised, while the CORE-Econ project has helped to establish modern teaching at Bachelor’s level. Students appreciate the variety of materials, including the linked videos, and find it more interesting than traditional textbooks. They also find visiting the university library less attractive as they prefer to use digital materials. We initially provided some printed copies in the library because there were requests, but quickly realised that these were hardly used as the online content is updated annually and the print version is then no longer completely up-to-date.

If you were to give younger scientists tips for entering the field of open science, what would they be?

HH: A key tip is not to be afraid of transparency. For empirical researchers in particular, it is important to work so carefully from the outset that the results can be replicated at any time. This means making data and scripts available in a transparent manner. Errors in research not only damage credibility, but can also damage one’s own reputation. Open science also offers the opportunity to develop and scrutinise your own scientific standards at an early stage: how do I generate and process data, and how do I ensure that my work is reproducible? Of course, there are also risks for young researchers: Great transparency can lead to more experienced scientists utilising the disclosed methods and results more quickly. You should bear in mind that there are risks if, for example, you publish and share data before the associated paper has been finalised and published. If you explain in detail how you did something at a conference and at the same time release a GitHub repository with all the code, it can happen that others implement and publish the ideas more quickly. Hence my advice to young researchers: Don’t be afraid of open science, but you should manage the publication process strategically. Think carefully about what you share and when, and make sure you have a certain level of security. This can be done, for example, via discussion papers or archives with a time stamp that document your own work. This makes it clear who worked on a topic first.

Thank you very much!

The interview was conducted on 26 August 2024 by Dr Doreen Siegfried.

About Prof Dr Hanna Hottenrott:

Contact: https://www.zew.de/team/hho and

https://www.ep.mgt.tum.de/eoi/team/prof-dr-hanna-hottenrott/

ORCID-ID: https://orcid.org/0000-0002-1584-8106

LinkedIn: https://www.linkedin.com/in/hanna-hottenrott-81120085

ResearchGate: https://www.researchgate.net/profile/Hanna_Hottenrott