Item Types


Would it be possible to get two additional item types?

-Dataset/Data Source

  • I second the request for a "data set" item type. The ability to cite the data used in research is an essential part of the move toward verifiable and open research. People are starting to publish data sets online and to cite them in publications. Funding bodies are starting to recognise data publication and data citation alongside (if not equivalent to) publication and citation of research articles, books, book chapters etc. As this cultural shift takes place, tools like Zotero need to keep up and support it.

    Ian Barnes
    Australian National Data Service
  • +1 on Dataset.
    -1 on Syllabus (just use Document)
  • Added to ticket.

    Are datasets like databases (see examples below)? If yes, you could use document item type to cite them. Or does the reference have to include other information? If yes, it would be helpful if someone could list necessary fields and a sample formatted reference in a common format, such as APA or Chicago.

    CMS 17.358 Scientific databases

    In the sciences especially, it has become customary to cite databases as follows: list, at a minimum, in this order, the name of the database, the URL, a descriptive phrase or record locator (such as a data marker or accession number) indicating the part of the database being cited or explaining the nature of the reference, and finally an access date. In bibliographies or reference lists, list under the name of the database.

    N: 1. NASA/IPAC Extragalactic Database, (object name IRAS F00400+4059; accessed August 1, 2001).
    5. GenBank, (for RP11-322N14 BAC [accession number AC017046]; accessed August 6, 2001).
    7. Unified Database (Bioinformatics Unit and Genome Center, Weizmann Institute of Science), (for mapping data on candidate genes; accessed July 29, 2001).

    B/R: NASA/IPAC Extragalactic Database. (object name IRAS F00400+4059; accessed August 1, 2001).
    GenBank. (for RP11-322N14 BAC [accession number AC017046]; accessed August 6, 2001).
    Unified Database. Bioinformatics Unit and Genome Center, Weizmann Institute of Science. (for mapping data on candidate genes; accessed July 29, 2001).

    T: (NASA/IPAC Extragalactic Database [object name IRAS F00400+4059])
    (Unified Database)
    (Genbank [accession number AC017046])
  • Hi,
    Can you tell me if adding the item type, "data set," is in the works? If so, any idea when it will be available?
  • This is planned. See and

    Your comments on what else you might need to use the new type effectively are very much welcome at this stage, so we get this right.
  • You asked for comments re: what might be needed to use a new "dataset" item type effectively. The Data-PASS (Data Preservation Alliance for the Social Sciences) partners recently posted a page on the importance of citing research data. The page includes recommendations for the basic elements that should make up a dataset citation. Hopefully it will help with Zotero's creation of a "dataset" item type:

    Data-PASS is supported by an award from the Library of Congress through its National Digital Information and Preservation Program (NDIIPP), and is led by the Inter-university Consortium for Political and Social Research (ICPSR) at the University of Michigan, the Roper Center for Public Opinion Research at the University of Connecticut, the Howard W. Odum Institute at the University of North Carolina-Chapel Hill, the Henry A. Murray Research Archive, a member of the Institute for Quantitative Social Science at Harvard University, the National Archives and Records Administration, and the Harvard-MIT Data Center, also a member of the Institute for Quantitative Social Science at Harvard University.

    Elizabeth Moss
  • I'm still not sure what exactly we need as additional fields. This suggests that version would be a good idea - that seems to make sense to me - popular datasets in the social sciences tend to have multiple versions. Digital identifier - I'm not sure I understand if this should include identifiers beyond DOI and URL - a checksum, really?
    I guess the other thing would be format - always feel the insistence on writing [computer file] in these citations is a bit silly - what else is it going to be? A pile of punch cards? I also don't get the sense that any style actually requires that - it seems more like a byproduct of an old ICPSR practice - but it would be easy enough to do using a medium field.
    The last question is how to deal with distributor and producer distinctions.
    Here's the issue thread
    here's the rudimentary wiki page

    more sustained input from actual experts would certainly be helpful.
  • I would recommend starting with the minimal required elements of the DataCite Metadata Schema ( DataCite is the primary service for minting DOI for datasets.

    I will say that the database examples provided above are often very different than static datasets. It is probably easier to have two different item types for static data and dynamic data.
  • I think we're set on what's needed on the dataset citations (note the above is 10 years old -- so this was before Force 11 Data Citation Guidlines, and the implementation guidelines for data repositories & publishers as published in Scientific Data).

    Here are the current guidelines for Zotero: . CSL citation styles already cover them.

    I agree databases & datasets are somewhat distinct, but I doubt we actually want/need a separate database item type. Between webpage and dataset, most citation/referencing scenarios seem covered, and multiplying item types create all sorts of usability issues.
  • Dear Zotero developers,

    I’m a data archivist at FORS – Swiss Center of Expertise for the Social Sciences. I take the opportunity of participating in the Swiss Open Research Data Hackathon ( to write you about this idea of developing a new item type for "Datasets". We are now developing an app to import metadata from data archives' catalogs in reference management software like Zotero. The idea is to help researchers cite correctly the data they use. In this respect, it is a pity that Zotero has not yet developed a specific item type for "Datasets".

    Indeed, with the recent years' development in social sciences (but probably in other domains too), there is an urgent need for a specific item type for "Datasets" in Zotero. Reading the different discussions on this topic, the main argument against this proposition is that the item type "Document" could be used instead. I see different downsides of using the item type "Document" for datasets:

    1. The "Document" item type does not automatically include a persistent identifier (PID; e.g., a DOI). Thus, the PID (DOI) is only mentioned in the Extra field. However, the PID (DOI) is more and more recognized as a mandatory element when citing datasets. This way datasets can be tracked and data creators/authors acknowledged for their work.

    2. A dataset is rather a collection of documents, including data file(s), documentation file(s), and metadata, than a single document.

    3. Bibliographic styles could develop (if not already) specific ways to cite documents and datasets, so the distinction is important.

    Finally, other important elements to cite a dataset, apart from authors, title, year of publication and publisher, are: dataset version and info on the data type (often mentioned in square brackets [Dataset] or [Data file]). It could also be convenient for researchers to add specific information for datasets (e.g., data collection method, data collection period).

    I hope that you would consider this improvement for the next Zotero version. Please let me know if I can help you somehow!

    Best regards,
    Christina Bornatici
    FORS Data Archive Team

  • Reading the different discussions on this topic, the main argument against this proposition is that the item type "Document" could be used instead.
    No, this is a misunderstanding of the discussions we've had over the last 5+ years on this (again, the beginning of this thread is >10 years old and not really relevant).

    To recap the current state of discussion: Everyone involved agrees that there should be a dataset item type and there will be a dataset item type. That's also made explicit on the page I link to in my last post.

    Zotero has, for technical reasons, not been able to make any updates to any item types over the last years. Using Document is a _workaround_ that enables people and translators to save datasets into Zotero in a way that relevant information will be captured. The DOI in the Extra field is already used in the citation styles and will be merged to a proper field once the dataset item type exists. (That's no different from other item types like Reports and chapters that increasingly use DOIs but don't have them in Zotero) Same for version and dataset type. Datasets entered this way are already treated differently by citation styles where they specify such treatment (see e.g. the APA style).
Sign In or Register to comment.