Down with Reviews
2010-4-23 21:24 | the-scientist.com | the-scientist
Review articles simply don’t deserve all the citations they receive.
There has been a proliferation of review articles and review journals over the last decade, and it is easy to see why. Biologists find them useful to keep up with increasingly complex science, and publishers find them an easy way to increase the impact factor of journals. Unfortunately, the reasons why review articles are highly cited are mostly unrelated to their educational value. Instead, it seems mostly due to their increasing use as a surrogate for the primary literature.
Many years ago, the privilege of writing review articles was usually reserved for the most senior of scientists, who would describe the most significant findings in their fields from the previous decade. These would appear in specialized publications, such as the Annual Review of Biochemistry. The topics were specialized and were only chosen for review if sufficient progress had been made to justify a conceptual update. I found these types of review articles to be great reference works, and they provided a foundation for my own knowledge in biology.
As I advanced in my career, biology not only became more specialized, but it started to progress faster. This forced me to work harder to keep up with my own area of research, but caused severe problems when I started to teach. For example, when I taught advanced courses in cancer and cell biology, I could easily cover research in my own area of specialization. However, it was almost impossible to evaluate and explain the primary literature in tangential areas of biology. I simply did not have the necessary expertise.Related Articles
Is Peer Review Broken?
Critics Rip Cell Paper
Bringing Order to Authorship
Classic review articles were rarely more useful than textbooks and were usually too dated for an advanced course. Fortunately, the problem of understanding unfamiliar areas of research was solved by the introduction of review journals, such as Current Opinion in Cell Biology, which would review an entire field of research at yearly intervals. These journals were a godsend for teaching because they not only gave a high-level description of the most important advances in a broad range of fields, they also provided detailed evaluations of the primary literature.
Although I used this new type of review article for my classes, I rarely paid much attention to those that addressed research in my own area. I already was familiar with the papers they discussed and usually found myself disagreeing with the author’s interpretation or choice of “important” papers. Rather than subject myself to the irritation, I normally chose to skip reading them. I am sure that my scientific colleagues had similar feelings toward my own reviews.
If review articles are mostly used for teaching and learning about unfamiliar areas of research, then why are they so highly cited? Their use in teaching certainly does not influence citation rates, yet it is their citation rates that are responsible for their proliferation in almost all primary research journals.
Personally, I feel that their high citation rate is due to their being used as a surrogate for the primary literature. Instead of referring to the original article that made a discovery, many authors instead cite a review that mentions the paper. Even better (or worse), a single review article can be used as a substitute for a half dozen original papers. Thus, the high citation number of a review article is mostly a reflection of the aggregate number of papers it points to.
It is a very unfortunate situation when review articles are used as citation surrogates. It robs scientists of the citations that they often need to advance their careers and obscures the credit for scientific discoveries. Many trendy journals (who also promote review journals) encourage this type of citation abuse by placing severe limits on article size and refuse to exempt references from the word count. The easiest way to shrink your article is to cite a couple of reviews rather than a dozen papers.
I believe that the only valid instances to cite a review article in an original paper are cases in which they introduce original concepts or when they provide historical context. Cited discoveries and hypotheses should always refer to the original literature, even if this means spending days tracking down the source. Review articles should go back to being used as an educational resource instead of a way to avoid the hard work of scholarly discussion of the literature.
Steven Wiley is Lead Biologist for the Environmental Molecular Sciences Laboratory at Pacific Northwest National Laboratory.