Judging a source

Not every source is the same kind of thing, and a review that cites the wrong kind quietly loses its footing.

A primary source reports original research: the authors collected or analyzed data and are telling you what they found. This is the evidence your review is actually built on.

A secondary source digests other people's primary work: a literature review, a systematic review, a meta-analysis, a textbook. These are invaluable for orienting yourself and for finding primary sources (chase their reference lists), but citing a review for a specific finding, when you could cite the study that found it, puts a layer of someone else's interpretation between you and the evidence.

Then there are sources that are not research at all: editorials, opinion pieces, news articles, press releases. They can point you toward research, but they are not evidence. The move is simple: build on primary sources, use secondary sources to navigate, and keep opinion out of the evidence base.

Once you know what kind of source you have, you have to decide whether to trust it. Three questions carry most of the weight.

Was it peer reviewed? Peer review is imperfect, but it means at least some independent experts checked the work before publication. A preprint or a blog post has skipped that step.

Where was it published? Legitimate journals have editorial boards, transparent processes, and a track record. Predatory journals mimic the look of real ones but will publish almost anything for a fee, with little or no real review. Their tells: aggressive email solicitations, promises of review in days, fake or padded editorial boards, and names that imitate established journals.

Has it held up? Papers are sometimes retracted for error or misconduct. A retracted paper can keep collecting citations for years from people who never checked, so before you lean on a key source, make sure it has not been withdrawn.