What can research institutions do to encourage open science?

Question

The research landscape is complex, and different actors have different options to help with a systemic move towards open science.

Research institutions

• employ researchers
• administrate financial, legal, ethical and other aspects of research
• contribute to research infrastructure
• help communicate research
• do many other things

In this thread, I would like us to explore which options research institutions have to encourage open science in-house as well as in the communities and other stakeholders they interact with.

If you have multiple ideas for concrete actions, I suggest to post them as separate answers (rather than as one monolithic answer), so as to facilitate the discussion and refinement of your ideas.

Comments

Encourage and legitimate the use of ArXiv.

In rather theoretical fields such as high energy physics, the top people directly publish to and cite from the ArXiv whereas publishing to a print or pay journal too is not mandatory.

However, in more "applied" research fields, such as atmospheric science for example, publishing to the ArXiv or citing ArXiv papers is frowned upon.

The negative attitude of more applied research fields with respect to the ArXiv should be changed.

---

Yes - encouraging preprints is an important step towards making more of the research process open. The discussion about that is slowly reaching beyond arXiv-y fields, e.g. as more biologists are beginning to discover the concept (cf. http://asapbio.org/ ).

Any reason you posted this as a comment rather than an answer?

---

Unfortunately, the negative attitude concerning preprints is still way to widespread:

http://meta.physics.stackexchange.com/q/7750/2751

---

I never noticed it was frowned upon in the atmospheric sciences to use ArXiv. I did, I do not think that hurt my career, but it probably also did not help because likely no one noticed.

ArXiv has a category for "atmospheric physics", that is somewhat limited and many studies in the atmospheric sciences do not fit in it. That may be one of many reasons why the adoption has not gotten above a critical mass.

There is now a better place for such manuscripts: EarthArXiv.
https://eartharxiv.org

Answer 1

Change hiring and promotion practices.

The criteria used in hiring and promotion processes should

• be public
• give weight to research outcomes all along the research cycle, especially those that come before formally published narratives.

A good starting point for considering this issue is this commentary on a passage in a job ad, which asked candidates to outline their activities around openness: http://www.nicebread.de/open-science-hiring-practices/ .

Comments

Here is another such example, from the Department of Psychology, University of Toronto Mississauga (UTM): https://web.archive.org/save/_embed/https://utoronto.taleo.net/careersection/10050/jobdetail.ftl?job=1600713&lang=en

"The research statement should include information about how the applicant has or intends to respond to issues of replicability and open science."

---

The above link seems defunct, but a quick search for the quoted string brings about a number of other examples, e.g. http://dbbs.wustl.edu/Resources/Pages/job_details.aspx?JID=2567 (archived at https://web.archive.org/web/20170313155440/http://dbbs.wustl.edu/Resources/Pages/job_details.aspx?JID=2567 ).

---

See also this growing collection of acadmeic job ads that mention open science practices: https://osf.io/7jbnt/

Answer 2

Encourage and legitimate the use of ArXiv.

In rather theoretical fields such as high energy physics, the top people directly publish to and cite from the ArXiv whereas publishing to a print or pay journal too is not mandatory.

However, in more "applied" research fields, such as atmospheric science for example, publishing to the ArXiv or citing ArXiv papers is frowned upon.

The negative attitude of more applied research fields with respect to the ArXiv should be changed.

Comments

I never noticed it was frowned upon in the atmospheric sciences to use ArXiv. I did, I do not think that hurt my career, but it probably also did not help because likely no one noticed.

ArXiv has a category for "atmospheric physics", that is somewhat limited and many studies in the atmospheric sciences do not fit in it. That may be one of many reasons why the adoption has not gotten above a critical mass.

There is now a better place for such manuscripts: EarthArXiv.
https://eartharxiv.org

Answer 3

Encourage experimentation around open science at all levels within the institution - by individual researchers, educators, students as well as research teams and institutional subunits or across all of them - as well as between institutions.

A prime example for an entire institution experimenting with open science is the Montral Neurological Institute, which declared in early 2016 that

Starting this year, any work done there will conform to the principles of the “open- science” movement—all results and data will be made freely available at the time of publication, for example, and the institute will not pursue patents on any of its discoveries. Although some large-scale initiatives like the government-funded Human Genome Project have made all data completely open, MNI will be the first scientific institute to follow that path, Rouleau says.

For a recent update on that, see http://doi.org/10.1038/533S71a, which also contains pointers to some similar initiatives.

On the level of subunits, a legal department could review existing legislation from an open science perspective and make suggestions for sharing-friendly language in contracts, agreements and other legal documents, as demonstrated, e.g., by Creative Commons licenses as well as author addenda to copyright transfer agreements and redesigned material transfer agreements.

 

Answer 4

Put the website and outreach materials under an open license. For example, the legal department of the German Aerospace Center (DLR) observed:

through the use of a CC licence, additional usage rights are granted and thus, a higher degree of dissemination of the published media content owned by DLR will mostly be achieved.

 

Comments

This part is not just about research and not just about PR materials, but about research and non-research activities generally, including about the institution, its people and facilities etc.

For example, profiles of people working at the institution (e.g. https://infocus.nlm.nih.gov/2016/10/04/focus-on-nlm-scientists-kim-pruitt-has-built-a-career-on-passion-and-persistence/ ) can be very helpful for organizers of conferences, for journalists, journal and Wikipedia editors, funding agencies, potential collaborators and others.

Answer 5

Change evaluation practices. This is closely tied to Change hiring and promotion practices but broader. The classical example that is often given in this context is that of the University of Liège, where

depositing papers in the repository was henceforth the sole mechanism for submitting them to be considered when researchers underwent performance review.

While that policy and also the repository (more details here) are focused on providing free-to-read access to scholarly publications, institutions could expand that to repositories of other kinds, e.g. for data, software, specimens, teaching materials and other results of the work of their staff.

Answer 6

Publish data management plans for funded grant proposals.

Answer 7

Directly from this tweet:

Every scientific institution should have the subdomains software.theinstitute.tld and data.theinstitute.tld

On those subdomains, the software or data generated by the institution should be made accessible with minimal barriers. If they are hosted in a recognized repository elsewhere, the respective persistent URLs should be given.

Comments

Please be aware that my tweet contains a typo and that the second domain should be data.theinstitute.tld.

---

Thanks - fixed.

Answer 8

Sign declarations relevant to Open Science and related issues - and act accordingly.

Some examples:

• Budapest, Berlin, Bethesda declarations or similar initiatives on Open Access
• San Francisco Declaration on Research Assessment
• Cape Town declaration on Open Educational Resources
• Bouchout Declaration for Open Biodiversity Knowledge Management
• Statement on data sharing in public health emergencies
• Bermuda Principles
• Panton Principles

Even if you were not part of the original community that drafted a particular statement, signing on to those principles later on (or with some adaptations, e.g. for discipline-specific things) may make sense.

 

Comments

Another one to sign: "Open Access 2020" - https://oa2020.org/ .

---

See also the Commitment to Reseasrch Transparency: http://www.researchtransparency.org/

---

A brief overview of open data-related policies sits at https://cos.io/blog/landscape-open-data-policies/

Answer 9

Publish and maintain a list of currently ongoing research projects, with abstract, contact information and preferably more details.

Answer 10

Publish grant proposals that got funded.

Comments

I would also publish grant proposals that did not get funded. Not so much as a negative example but because some of them are really inspiring.

---

I fully agree that publishing all research proposals has value. Gotta start somewhere, though.

Answer 11

Review the openness of institutional websites:

- to what extent is the content openly licensed, and tagged as such?

- to what extent are these websites built using free and open-source software? Is this documented?

- to what extent is the website content properly archived, e.g. with the Internet Archive? And what about content linked from the website?

- if the robots.txt has restrictions, are these justified?

- do the websites use the Memento protocol?

- how do the websites perform when used from various platforms (Linux, OSX, Windows, ...), devices (desktop, smartphone, tablet, screen readers ...) or locations?

- are job ads, event announcements, notes on construction work and other rather ephemeral pieces of information included in any of the above?

Answer 12

Appoint a contact person for matters of open science, similar to such contact persons for ethical, diversity or open access matters.

Do not assume that they will take on this role as a volunteer in the long run, but make it part of their job description.

This idea came out of the final meeting of the Open Science Fellowship program 2016.

Answer 13

Provide recognition for researchers doing open science or producing open educational resources.

This could be achieved by

• reducing teaching and/ or administrative load
• prizes, awards etc. that recognize openness

This idea was discussed at the final meeting of the Open Science Fellowship program 2016.