In two prior posts here and here, we discussed why references are important, provided actual examples of reference errors, and emphasized that a shoddy reference list often equates to shoddy work that should not be published.
In the May 3, 2021 post, we explained how to obtain beautifully formatted references from PubMed. In this post, we provide a primer on basic reference know-how.
Here are our best tips:
1. References are required when you refer to or paraphrase someone else’s work, when you directly quote a source, and when you use statistical figures.
For example, the following statements require references:
Immunosuppressive agents have been linked to cancer in a dose-dependent relationship.
A meta-analysis of 33 studies revealed that the prevalence of patients with calcified aortic stenosis and coronary arterial disease was 37%.
2. Primary sources must be cited. Say you’re reading a case report or a literature review that includes a fact you’d like to include in your paper. If the sentence containing that fact has a reference number after it, go to the reference list and obtain the citation for that original paper. Do not cite the case or the review that restated someone else’s work. We see this ALL THE TIME. Primary sources must be cited.
3. References must be current. Medicine changes constantly, so citing a paper from 1986—unless it’s considered a classic—is poor practice. A good rule of thumb is to limit references to papers published within the last 10 years.
4. Most medical journals require references to be called out in chronological (numerical) order in the text. In other words, reference callouts in the text start with number 1, followed by 2, 3, 4, and so on. Don’t skip numbers. It’s not OK to squeeze reference 7 in after you’ve moved on to reference 9. You have to renumber.
Here’s an actual example from a paper we received. Reference 5 is the first reference callout in the paper, followed by an amazingly out-of-sequence reference string:
De novo malignancies in post-transplant patients have been observed and documented in literature. The incidence was found to be between 3-26% (5). In post-liver transplant patients, the risk is 1.4-4.9 times greater than the general population (3, 7, 13-15, 19).
5. Every reference in the reference list must have a callout in the text. We received a paper this week that didn’t have a callout for reference 7. All the other references had callouts, but poor reference 7 did not.
6. References in the reference list must be numbered. We received a paper this week with numerical callouts in the text, but every citation in the reference list had a bullet instead of a number in front of it. This is true. We can’t make this stuff up.