On Citing Practices
How authors, journals, and indexes can deal with citations better
[Pluto Series] #0 — Academia, Structurally Fxxked Up
[Pluto Series] #1 — Research, the Knowledge Creating Industry
[Pluto Series] #2 — Academia, Publishing, and Scholarly Communication
[Pluto Series] #3 — Publish, but really Perish?
[Pluto Series] #4 — Publish or Perish, and Lost in Vain
[Pluto Series] #5 — On Where they Publish
[Pluto Series] #6 — On Number of Publications
[Pluto Series] #7 — On Citation Fundamentals
[Pluto Series] #8 — On Citing Practices
[Pluto Series] #9 — On Tracking Citations
[Pluto Series] #10 — On Peer Reviews
[Pluto Series] #11 — Ending the Series
Delving into the core discussions of the series, first two points were addressed around the number of publications and where those publications are published, as evaluation criteria for research outputs. Beginning the discussion on the third topic, last post introduced the definition, roles, and some characteristics of citations. This post will be focused on how several stakeholders in the system can do better regarding the “creation” of citations. (i.e. citing practices). The discussion is important not only because citations are used for aggregate metrics in assessing research performances but also because citations are per se acts of giving credits to prior studies.
Like any other discussions in this series, citing practices also span multiple stakeholders in the communication system of academia, and have complex dynamics among them. Some aspects can be put for specific players, but many others would be more like general discourse for academia as a whole. Key roles in the system of citations are taken by individual researchers, journals and their publishers and editors, and citation indexes/databases. This post would address some points as such.
For Individuals
Individual researchers are involved in the citation practices mostly as authors or reviewers. Their roles as reviewers will be discussed in a separate post about peer review, but simply put reviewers need to be just as meticulous when dealing with citations as they would do with the manuscripts.
- Cite every information essentially relevant but not original in my work
- Understand what citations do not just for my work but also for the bigger system of scientific literature
- Be critical about how journals, editors and reviewers deal with citations, and speak up for improper practices
Cite your Conscience
As authors, the only option open for individual researchers is choosing “what to cite.” Regarding these choices by authors, some points have been addressed for long such as that they need to cite primary sources, or that they need to cite any critical information that is not their own. Googling “what should I cite” gives you a lot of library guides, and they’re mostly more or less different expressions for the basic roles of citations described in the last post. That is, when authors are confused what they have to cite, the best guideline is probably to follow the basic roles of citations. Think of which publications have mostly contributed to your work, which publications are required for your peers to validate your claims, and which publications would fit for your readers’ further explorations.
Citation is highly Contextual
Many of those “library guides” also include when authors don’t need to cite, or Common Knowledge. According to what citation does, anything that is not original in a work needs citations, but practically authors do not cite every non-original information in their works. These not-original but also not-cited pieces of information are deemed common knowledge. The MIT academic integrity handbook describes that you should take into account what your audiences might know already and whether they might ask where you obtained specific information. Likewise, whether a given piece of information is common knowledge depends on the audience, or the commons (i.e. community), it is given to. Or shortly, it depends on the context.
Because the notion of common knowledge is very contextual, it is near impossible to have a strict, universal ground rules on “what needs to be cited.” And the ones who best understand what needs to be cited is possibly the authors and their peers, who understands the context. As such, again, individual researchers are asked to act honest and comply to the basic definition and roles of citations, both as authors and as peers (i.e. reviewers and readers).
Beyond these explicit roles like authors and reviewers, researchers as a member of their expertise community need to understand how their community, journals, editors and other members are dealing with citations. When their journals and editors are providing certain requirements, guidelines, or some norms, individual academics need to be critical in taking those, comply if legitimate but speak up when they think otherwise.
For Jounals
Journals, their editors and publishers have the most influences in the publishing process. Their policies and behaviors more than affect citing practices and rather set the standard in the community. They are thus required more changes for better citing practices.
- improve the styling by which publications and their authors are cited
- improve the way cited items are uniquely identified
- embrace more information in citations such as contexts, retrieval, or details of cited information
- treat citations as separable, structured, and updatable resources
Styling
Styling is about how citations are rendered in written texts. Two most notable cases are the in-text citations and endnote references. As the series is about incentives and evaluation in academia, and citations are used as evaluation specifically on aggregate levels, the issues regarding styling may look peripheral. Styling of citations in many journals has strict, explicit, and manifest rules, and authors need to just follow it. Moreover, when citations are aggregated to be used as metrics, their styles do not really have any effect. But still, it is important to note that citations are intrinsically act of giving credits to prior publications for sourced information. Thus how citations are rendered in texts should also comply with the definitions and roles of citation.
Citations are essentially relations between publications. The attributions of sourced information are given to “prior works”, but not to their authors. As such, when academics are evaluated with citation metrics, their impacts are substituted with the numeric counts, which doesn’t directly mean their impact itself. However, if you look at the major styling guides of journals for citations and endnote references, they are rendered in a quite different manner. In most of journals, citations are expressed with all the author names when they’re less than four, and when there’re more than three, “et al.” is used (which is latin for “and others”) and the name of the first author is the only one presented.
This way of expressing the authorship of cited works is bad in that the citations are not given to the authors but the publication. What’s more, in academic publishing there’s no agreed rules for authorship and contributorship*. In a lot of STEM journals, readers would perceive the first authors as the most contributing ones, and the last authors as the principal investigators. But there are disciplines where this doesn’t hold, and most importantly, again, there is no consented norms for this. Journals need to find better ways both in registering the contributions of authors in a publication, and in rendering the citations in texts accordingly.
(*See this Wikipedia for detailed explanations)
(** or CRediT, Contributor Roles Taxonomy, by CASRAI)
Unique Identification
Another important aspect in rendering citations is unique identification of source publications. It is specifically important in the information structure, but it is also as much important in how they are rendered because a lot of researchers are doing research also on pen-and-papers. And of course the scholarly communication can’t never be completely digitized.
Recently, Digital Object Identifiers (DOIs) are working as a decent way to uniquely identify digital objects. As most of journals are registering their publications with DOIs, they work as a great tool for citations. Still, not all journals are registering DOIs, and there should be more discussion on “what is citable.” As such, journals need to continuously improve the way cited works are uniquely identified. Considerations should be given in that i) it can’t be completely digitized, and that ii) more types of items may be open to citations.
Citation Contexts
In many previous studies, it is shown that there are diverse contexts in citations. Most prevalent method was analyzing the sentences where citations are used (i.e. in-text attributions) or the places in the publication where citations are put. Citations were categorized with their contexts, in the perspectives of “cited works”, in binary (confirming-negating) and in ternary (corroborative-oppositional-corrective), and in the perspectives of “citing works” with their manifest roles.
Unlike, currently citations are simple links between citing works and cited works. That is why the aforementioned studies have analyzed the full-texts of publications to extract their contexts. According to Initiative for Open Citations (I4OC), citations are more powerful when they are separable and structured. This is more significant considering that citations are used as evaluative metrics in aggregate level. Separable and structured citations with more contextual information would enable use of those aggregate level metrics with more diversification of evaluations. For example, they can make it possible to track down only the “positive” citations for evaluations.*
(*This part of citation context addresses what “citing work does to cited work”)
Details of Cited Information and Retrieval
Other contextual, detailed information that could be embraced in citations include the details of cited information and details of retrieval. Currently, the details about cited information are usually limited to the place of the information from the source publications, which is usually the pages where they appear. More specific details could be their data types (e.g. direct quote on sentences, datasets, methodology, hypothesis, etc.), details about place of the source information (e.g. in the introduction, body, conclusions, etc.), or the context of source information* (e.g. supportive, negating, juxtaposition to citing work).
(*this part of citation context addresses what “cited information does to citing work”)
For the details of retrieval, currently information on both the brick-and-mortar publishing and the online publishing are given. The former is given within the forms of [journal titles, publishing year, volume, issue, page] and the latter with DOIs or URLs. The online forms can’t substitute the offline information since the publishing cannot be completely substituted with digital methods, and sometimes the DOIs and URLs fail to sustain for long periods. (i.e. they’re not permanent)
The combination of offline and online information for retrieval is thus very powerful in supporting the readers. However, the offline format needs continuous discourse for improvements because not all the publications fit into those formats and more diverse forms of “cited items” may require more diverse formats. Maybe the best details about retrieval of cited information could be “how the citing authors actually obtained the resource”.
Updatable, Separable, Structured
Another yet great policy that journals can have about citations is making them updatable. Just as the publications may be revised or retracted, so does citations. Because the norms around how citations should be practiced can’t be exhaustively manifested, the conformity of citations to those even contextual and vague norms can only be maintained by social validation. This is why peer reviews also include comments on the bibliographic references of the publication being reviewed.*
(*This has also been pointed out by Eugene Garfield in his 1964 paper.)
This necessity of social validation as well means that citations can go wrong. Citations may also require changes due to the changes in the cited publication. An updatable citation would further make them robust when the community can join forces to build more sound citation relations. As such, making citations updatable, separable, and structured will empower them and allow to be “more informing” data sources in the whole research information scene.
For Indexes
Citation indexes and databases collect information on publications and citations between them from journals and publishers, enabling the aggregate level analytics over it. As evaluation with citation metrics are given on aggregate levels, the influence of indexes is significant in the system. Indeed, many citation related metrics used in research assessment cannot exist without defining the underlying citation index, and a lot of times they are omitted and just the names of the metrics are given alone. Most of the times, the indexes are either Web of Science or SCOPUS. More recent ones include Google Scholar or CrossRef.
Indexes have roles in “tracking citations” rather than “generating citations”, so this may be more fit to the next post which will discuss the perspectives on tracking how publications are “cited”. However, just as journals and publishers do to authors, the policies set by indexes may have significant influence on journals. Therefore, indexes should be clear and transparent on what, why, and how they index certain publications and their citations. Moreover, they need to constantly communicate with the research communities for better ways of indexing. Because indexes better understand the macro level aspects of academic publications and their links, they may proactively speak up to the learned societies for changes. But meanwhile they need to as well listen to the societies for their requirements.
One very important thing required for indexes is that they be clear and transparent about their policies on “refereeing”. As mentioned above, refereeing (i.e. peer review) takes a significant role in sustaining the norms and the conformity of citations to them in journals. This has also been pointed out by the father of citation indexing, Eugene Garfield, in his 1964 paper*. Indexes should put efforts in checking that their indexed journals are conforming to their policies about refereeing, and communicate relevant discourses on refereeing with journals and communities.
(*p.530)
Evolving Social Construct
To use citations for evaluative metrics, it must be created well, collected well, and used well. In creating the citations, individual researchers, journals and their publisher and editors have most significant roles, thus being required most significant efforts to work better. However, they should not be asked for all the responsibilities just because they are most central in their roles as generators of citations. Similarly, indexes have most roles in collecting them, but cannot as well take all the responsibilities in the system.
Most importantly, in short, the stakeholders in the overall system need to closely work together and communicate to make a better system. Just like how modern science evolves knowledge.
In the upcoming post, focusing on Citation Indexes, discussions will be addressed around how citations are tracked and used as metrics. Please CLAP, SHARE & COMMENT for better discussions!
[Pluto Series] #0 — Academia, Structurally Fxxked Up
[Pluto Series] #1 — Research, the Knowledge Creating Industry
[Pluto Series] #2 — Academia, Publishing, and Scholarly Communication
[Pluto Series] #3 — Publish, but really Perish?
[Pluto Series] #4 — Publish or Perish, and Lost in Vain
[Pluto Series] #5 — On Where they Publish
[Pluto Series] #6 — On Number of Publications
[Pluto Series] #7 — On Citation Fundamentals
[Pluto Series] #8 — On Citing Practices
[Pluto Series] #9 — On Tracking Citations
[Pluto Series] #10 — On Peer Reviews
[Pluto Series] #11 — Ending the Series
