At the beginning of this instructional resource we identified clarity, conciseness, accuracy, and integrity as essential elements of scientific writing. Unfortunately, the latter two concepts are sometimes overlooked with certain citation practices. Consider what can happen in the following scenario. A researcher needs to conduct a literature review for a manuscript that she will be submitting for publication to a biomedical journal. A literature search yields several useful abstracts and the researcher proceeds to track down the various journal articles. Unfortunately, one key article is not available online. It is not carried by her institution’s library, nor is it available at nearby libraries as it has been published as a technical report in an obscure nontraditional journal with limited circulation. Pressed for time, the researcher decides, instead, to rely on material from the abstract for the literature review and includes the journal article citation in the reference section. However, nowhere in the paper does she reveal that she relied on the abstract and not on the actual journal article.
Another variation of this problem occurs when the researcher cites the published version of the paper, but actually relies on the contents of an earlier version that was published in the proceedings of a conference, or the preliminary version that had been distributed at the conference presentation itself or a pre-print server. These behaviors violate the requisites of accuracy and integrity.
The main problem with relying on versions other than the published paper is that elements of these earlier versions may be different from their counterparts in the published version of the paper. Such changes are typically the result of the peer review process, editorial changes, or errors that are spotted and corrected by the author between the time the paper is presented at a conference and the time that it is subsequently published. In some cases, the published version will contain additional data and/or interpretations that are substantially different or perhaps even contrary to those of earlier versions. For example, a conference paper describing experimental data may, in its published form, contain additional data from a new experimental condition or new statistical analyses that were carried out in response to referees’ suggestions.
Data from the new condition can place the earlier data in a new perspective possibly leading to somewhat different interpretations. With respect to abstracts, relying on such summaries can be problematic because abstracts typically may not provide sufficient details about the paper’s relevance (i.e., Taylor, 2002). In addition, because of their condensed form, abstracts cannot provide essential details about a study’s methodology, and results. Moreover, it should be noted that in some databases there may be instances in which individuals other than the author/s of the journal article write the article’s abstract. As a result, subtle misrepresentations are likely more common with these abstracts. Writing guidelines, such as the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, discourage the use of abstracts as references.
Lastly, given the rise in retractions and corrections, authors must ensure that the evidence cited is up-to-date. Regrettably, a significant number of retracted papers continue to be cited in the literature (see Ferguson, 2015). In addition, the large number of corrections, errata, etc., issued each year suggests that a substantial number of uncorrected results and/or interpretations are being cited as legitimate evidence and both of these outcomes contribute to the further contamination of the scientific record.