search

biocaddie / prototype_issues

Used to report and track bioCADDIE prototype issues
3 stars 5 forks source link

citation-publication confusion #301

Open altergc opened 7 years ago

altergc commented 7 years ago

There is a confusion in both DataMED and DATS about citations and publications. We need to distinguish among three kinds of citations and publications.

First, datasets should have a citation that points to the dataset itself. ICPSR and other repositories provide citations for their data (including DOIs), and authors who reuse the data are instructed to put the citation in their publications. Let's call this a "dataset citation."

Second, some datasets are described in a publication, which introduces the dataset and describes how it was created. In some fields this publication is cited in place of a citation to the dataset itself. "Data journals" like Scientific Data publish descriptions of datasets. Let's call this a "data description publication."

Third, publications that re-use a dataset have their own citations, which should be linked to the dataset. Let's call these "secondary publications."

Both DATS and DataMED fail to make these distinctions.

DATS has a "Publication" entity, which can be linked to the "Dataset" entity by a "isCitedBy" property. This property implies that the publications are secondary publications.

DataMED has a "Citation" and a "Publication" section. At the moment, secondary publications from ICPSR are being captured under "Citation" while data description publications from Dataverse are being captured under "Publication". Both of these are wrong. See https://datamed.biocaddie.org/display-item.php?repository=0025&id=59d539815152c6518764a859&query=midus and https://datamed.biocaddie.org/display-item.php?repository=0012&id=56d4b805e4b0e644d312e6ef&query=midus for different representations of the same ICPSR dataset.

This situation needs to be corrected in the next versions of DATS and DataMED.

agbeltran commented 7 years ago

For "dataset citation" such as:

"Ryff, Carol, David Almeida, John S. Ayanian, Deborah S. Carr, Paul D. Cleary, Christopher Coe, Richard Davidson, Robert F. Kruger, Margie E. Lachman, Nadine F. Marks, Daniel K. Mroczek, Teresa Seeman, Marsha Mallick Seltzer, Burton H. Singer, Richard P. Sloan, Patricia A. Tun, Maxine Weinstein, and David Williams. Midlife in the United States (MIDUS 2): Milwaukee African American Sample, 2005-2006 (ICPSR 22840). ICPSR22840-v2. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2012-05-21. https://doi.org/10.3886/ICPSR22840.v2"

as this is information that can be derived from the Dataset metadata, we do not include this string in DATS directly, and people can use the Dataset metadata to build the citation text in multiple citation styles. Maybe adding the citation text is a feature request for DataMed (that would rely on the DATS metadata already available)?

DATS v2.2 does distinguish between "primaryPublications" (what you refer to as "data description publication" and "citations" (what you refer to as "secondary publications") and this was included in DATS after request by @jgrethe and the CDT (please see DATS v2.2. spreadsheet and Dataset schema).

altergc commented 7 years ago

Hi Alejandra,

Thanks for your response. I see that I was not looking at the latest version of DATS.

Regarding the data citation, Matthew also pointed out that the citation can be composed from the elements in DATS. It would be good for DataMed to do that.

The change to "primaryPublications" and "citations" does make the distinction that we need. I'm afraid that DataMed is not implementing it in an intuitive way. I think that there are two ways that DataMed could be improved.

First, I suggest changing "Publication" to "Primary Publications" and changing "Citations" to "Publications citing this dataset".

Second, the items listed under "Citations"/"Publications citing this dataset" should be formatted like citations (as Publication is now) and not as a list of elements.