[Rdap] [RDAP] How much data should researchers *actually* publicly deposit?

Daureen Nesdill daureen.nesdill at utah.edu
Fri Sep 7 12:30:37 EDT 2018


I feel change is coming
NSF is sponsoring a 2-day workshop in October, Workshop on Accelerating Public Access to Research Data
Also sponsored by APLU and AAU. The workshop is by invitation only.

Daureen Nesdill
U of Utah
From: rdap-bounces at kunverj.com [mailto:rdap-bounces at kunverj.com] On Behalf Of Amanda L Whitmire
Sent: Friday, September 07, 2018 10:01 AM
To: Research Data Access and Preservation <rdap at kunverj.com>
Subject: Re: [Rdap] [RDAP] How much data should researchers *actually* publicly deposit?

Hi all,

Sebastian makes a really important point: NSF culture and enforcement around data is ad hoc, and intentionally so. The higher levels of NSF administration recognize the slow progress toward open data, but remain committed to supporting communities of practice. They maintain a pragmatic approach that recognizes that what works for researchers will vary across disciplines. It’s not convenient for us, but it makes sense.

Anyhow - what I really came here to share is an example of how attitudes toward enforcement are changing in one corner of the NSF. In the new BIO DEB (Division of Environmental Biology) solicitation, there is a section on data dissemination & accessibility https://debblog.nsfbio.com/2018/08/23/the-new-deb-core-programs-solicitation-is-here/. It reads:

"DEB is ratcheting up expectations of data archiving and accessibility, as is generally the case across NSF. Our new solicitation makes clear that PIs who have had prior support within the last five years must provide details on how data have been permanently archived and made publicly available. This information should go in the “Results from Prior NSF” section, following the format described in the PAPPG (II.C.2.d.iii). Likewise, when submitting Annual and Final Reports, BIO PIs will be required to “include information about progress made in data management and sharing of research products (e.g., identifier or accession numbers for data sets…and other types of data sharing and dissemination).”

It will be increasingly important to craft thoughtful and thorough Data Management Plans when submitting new proposals. Reviewers and Program Officers will pay particular attention to how data and specimens are stored and when and how they will become publicly assessable. For context and guidance, see http://www.nsf.gov/bio/biodmp.jsp.”

A sign of some progress!

Best,
amanda


<><    <><    <><    <><    <><    <><    <><    <><    <><
Amanda L. Whitmire, Ph.D.
Head Librarian & Bibliographer, Harold A. Miller Library
Assistant to the Director, Hopkins Marine Station
120 Ocean View Blvd, Pacific Grove, CA 93950

http://orcid.org/0000-0003-2429-8879
thalassa at stanford.edu<mailto:thalassa at stanford.edu>
831.655.6228 | @AWhitTwit
--
I observe email-free evenings and weekends.




On Sep 7, 2018, at 5:46 AM, Jonathan Petters <jpetters at vt.edu<mailto:jpetters at vt.edu>> wrote:

I'll add that I usually tell NSF proposal writers to say in their DMP that they publicly share what data they're willing and able to. In many cases I can get them to agree that at least sharing the data displayed in figures/plots would be useful to their colleagues, and some are willing to go further in sharing raw data and processing scripts.

To follow NSF's policy, I agree with Sebastian that if

  *   the peer reviewers of your proposal are alright with your DMP (and what you say you will share),
  *   the program officer is alright with your DMP, and
  *   you eventually share what you say you will,
You're currently good with NSF. And really, the third bullet is a maybe at this time; it's not clear NSF is really checking DMP compliance at present. But they may in the future, and all the funding agencies say that non-compliance with a DMP may count against a prospective PI in future funding proposals.

-Jon

On Thu, Sep 6, 2018 at 8:33 PM, Sebastian Karcher <skarcher at maxwell.syr.edu<mailto:skarcher at maxwell.syr.edu>> wrote:
Hi Kate and others,

I think it’s important to remember that “The NSF” doesn’t really exist for this purpose. While yes, NSF has one overarching data policy (though directorates have slightly differing versions), in effect this is determined by a) the program officer and b) the (anonymous) selection committee
Assessing whether there is enough data sharing is basically a three stage process:
1) We have seen a number of occasions now where program officers have actually sent the DMP back to the researcher and asked them to include more data sharing before sending the proposal out for review. Typically this happens with relatively short deadlines and is stressful, but it doesn’t, afaik, harm the proposal. It’s pretty clear that how much individual program officers push applicants on this varies a lot.
2) The grant selection committee does have access to the DMP and is actually required, by NSF rules, to comment on it in their recommendation. In many cases, they’ll just glance at it and sign off, but these are, typically peers, perhaps even competitors of the grant applicant and I cannot imagine them looking kindly on obvious attempts to withhold data “to maintain a competitive edge”. I.e. not sharing “enough” data in the eyes of their fellow scientists can absolutely sink a grant application.
3) Finally, the program officer reviews and has to sign off on the grant reports. We have strong anecdotal evidence that there is currently very little follow through on promises to share data, but there is, of course, no guarantee that is going to stay that well, and obviously, not having your grant report accepted is a big problem. (Needless to say, it’s also deeply unethical to make promises to share data that one does not intend to keep).

I obviously have a bias against researchers who want to share as little data as possible, so take this with a grain of salt, but I do think that the above makes clear that trying to skirt by with a minimum of data sharing for an NSF grant is a risky strategy. I’d try to discourage it best I could even I didn’t believe it was also an ethically questionable strategy.

All best
Sebastian



From: rdap-bounces at kunverj.com<mailto:rdap-bounces at kunverj.com> [mailto:rdap-bounces at kunverj.com<mailto:rdap-bounces at kunverj.com>] On Behalf Of Kate Barron
Sent: Thursday, September 6, 2018 4:01 PM
To: rdap at mail.kunverj.com<mailto:rdap at mail.kunverj.com>
Subject: [Rdap] [RDAP] How much data should researchers *actually* publicly deposit?

Hi RDAP,

When a project is funded by a federal grant, how much raw data are researchers expected to make accessible? Must they truly make everything and anything publicly available, or are brief extracts and/or summary statistics acceptable (in other words, what is the minimum sharing requirement)? What practices have you noticed among your colleagues/constituents?

I am specifically looking at NSF grants, for which guidelines are (intentionally?) vague. I am working with a faculty member who is hesitant to share all of their data, as they want to maintain a competitive edge.

Thanks for your input,
Kate
--
Kate Barron
Data Services Librarian
Dr. Martin Luther King, Jr. Library
San José State University
kate.barron at sjsu.edu<mailto:kate.barron at sjsu.edu>
408-808-2038<tel:(408)%20808-2038>

Explore Data Services<https://libguides.sjsu.edu/dataservices> at SJSU King Library!

_______________________________________________
RDAP mailing list
RDAP at mail.kunverj.com<mailto:RDAP at mail.kunverj.com>
http://mail.kunverj.com/mailman/listinfo/rdap



--
Jonathan Petters Ph.D.
Data Management Consultant and Curation Services Coordinator
Data Services, University Libraries
Virginia Tech
(540) 232-8682
https://www.lib.vt.edu/research-learning/ResearchDataManagementAndCuration.html
ORCID: 0000-0002-0853-5814
_______________________________________________
RDAP mailing list
RDAP at mail.kunverj.com<mailto:RDAP at mail.kunverj.com>
http://mail.kunverj.com/mailman/listinfo/rdap

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kunverj.com/pipermail/rdap/attachments/20180907/98ef474b/attachment.html>


More information about the RDAP mailing list