You have a sentence in your draft that needs a citation, and two kinds of paper could sit behind it: a primary study that reports original data, or a review article that surveys a field. The choice is not about which type is "better." It is about what your sentence actually claims. A claim about a specific measured result needs the study that measured it. A claim about the state of a field, or about where a research gap sits, is usually better served by a review that has already mapped the terrain.
This article walks through that decision as a workflow you can run on a real chapter or manuscript, using a thesis, coursework, or January project-start scenario. It covers how to diagnose what your claim needs, how to compare both source types on the same criteria, where each one fails, and how to combine them into a defensible citation set.
What is your claim actually asking for?
Before you search, write the sentence you want to support in plain language and label its type. Most academic claims fall into a few shapes:
- A specific empirical result: a measured effect, a reported rate, an observed relationship in a defined sample.
- A methodological or design claim: how a procedure was carried out or why a design was chosen.
- A field-level or gap claim: what has been studied broadly, what remains open, or how a body of work has developed.
- A definitional or framing claim: how a concept is characterized in the literature.
A primary study is the natural support for the first two shapes, because it reports the original work. A review is the natural support for the third, because it synthesizes many studies into a picture of the field. Definitional claims can draw on either, depending on how the concept is treated.
This labeling step is the whole game. If you skip it, you end up citing whatever appears first in a search results list, and the mismatch shows up later when a reader asks why a broad review is standing behind a narrow numeric claim.
Why does the primary-versus-review choice go wrong so often?
Three habits push writers toward the wrong source type.
Search ranking rewards reviews. Reviews are often highly cited and easy to find, so they surface early. That makes them convenient, not automatically appropriate.
Reviews are easier to read. A review summarizes many findings in accessible prose, which tempts you to cite it for a specific number it mentions in passing. The number belongs to the underlying study, and citing the review for it blurs the trail.
Primary studies are harder to verify. You have to check the sample, the method, and whether the reported result matches your sentence. That effort is exactly what makes a primary citation strong when the claim is empirical.
A useful diagnostic: read your sentence aloud and ask, "Could a single study have produced this exact statement?" If yes, you probably need a primary study. If the sentence describes a pattern across many studies, a review is the better fit.
How do the two source types compare on the same task?
Take a concrete task: you are writing a thesis chapter on a topic and need to support two sentences — one empirical, one about the field's development. Here is how each source type performs against the same criteria.
| Criterion | Primary study | Review article |
|---|---|---|
| Best for | Specific measured results, methods, defined samples | Field-level patterns, gaps, development of a body of work |
| What you must verify | Sample, method, and that the reported result matches your sentence | Scope, inclusion criteria, and whether the synthesis covers your specific claim |
| Common failure | Overgeneralizing one study's result to a broader population | Citing the review for a specific number that belongs to an underlying study |
| Traceability | Direct: the claim sits in the study itself | Indirect: you may need to follow the review to the primary work |
| Effort to verify | Higher per claim | Lower per claim, but weaker for narrow empirical statements |
Neither column wins outright. The table shows that the two types serve different claim shapes, and the right move is to match the source to the sentence rather than to pick a favorite type.
Where does each source type fail?
Primary studies fail when you stretch them. A single study's result is bounded by its sample and design. Using it to support a sweeping statement about a whole population overreaches what the study reported. The fix is to narrow the sentence to what the study actually measured, or to add more primary studies that together support the broader claim.
Reviews fail when you use them as a shortcut to a number. A review may mention a specific finding, but that finding originates in a primary study. If your sentence needs that specific result, follow the review to the underlying work and cite it directly. Reviews also vary in scope, so a review that covers a neighboring area may not actually support your precise gap claim.
Both failures share a root cause: the citation is doing more work than the source can bear. Narrowing the claim or upgrading the source fixes it.
What does a defensible citation set look like?
A strong chapter usually needs both types, arranged by role:
- Reviews establish the landscape. Use them to frame the field, identify what has been studied, and locate the gap your work addresses.
- Primary studies carry the specific claims. Use them for measured results, methods, and defined findings.
- The trail stays visible. When a review points to a key primary study, cite the primary study for the specific claim and the review for the synthesis.
This arrangement means your reader can see both the big picture and the evidence beneath it. It also protects you when a supervisor asks, "Where does this number come from?" — the answer is a primary study, not a review that mentioned it.
A worked example: cleaning up a chapter over winter break
Suppose you are revising a thesis chapter during winter break, and your supervisor's feedback says the evidence base feels thin and loosely tied to the claims. Here is a workflow you can run.
Step 1 — Inventory your claims. Go through the chapter and tag each citation-bearing sentence as empirical, methodological, field-level, or definitional. You will quickly see which sentences are supported by the wrong source type.
Step 2 — Find the right sources for the mismatches. For empirical sentences currently backed by reviews, you need the primary studies. You can search for academic papers through WisPaper search and screen results against the specific claim each sentence makes. For field-level sentences backed by a single primary study, look for a review that maps the area.
Step 3 — Save and organize what you keep. As you confirm sources, save or upload the papers and organize them in My Library so your citation set lives in one place rather than scattered across browser tabs.
Step 4 — Check the citations themselves. Before you finalize, run your BibTeX through TrueCite, which checks references against real academic databases to flag hallucinated entries. This catches the citation that looked fine but points to a work that does not exist.
Step 5 — Ask questions of your saved papers. When you need to confirm what a saved paper actually says before citing it, use Library QA to ask questions based on papers uploaded to My Library or added from search results.
By the end of this loop, every empirical sentence has a primary study behind it, every field-level sentence has a review, and the citation list has been screened.
Starting a January project: building the evidence base from scratch
If you are beginning a new project in January rather than revising an old chapter, the order flips: you build the landscape first, then populate it with primary evidence.
Start by mapping the field with reviews so you know what has been studied and where the open questions sit. Then, for each gap you want to pursue, find the primary studies that define the current edge of knowledge. Save the whole set in My Library and set up the project in Research Projects so your sources, notes, and claims stay connected as the work develops.
When the reading load grows beyond what you can hold in your head, Scholar Agent supports longer research tasks, including literature review work, so you can move from a pile of papers to a structured view of what the evidence says. The point is not to outsource the reading but to keep the primary-versus-review distinction intact as the volume increases.
How do you know the citation set is ready?
A citation set is ready when three checks pass:
- Claim-source match: every empirical sentence is backed by a primary study that actually measured that result, and every field-level sentence is backed by a review whose scope covers the claim.
- Traceability: for any specific number or finding, you can point to the primary study that produced it, not just a review that mentioned it.
- Reference integrity: the citation list has been screened so that each entry corresponds to a real work.
If a sentence fails the first check, narrow the claim or find a better source. If it fails the second, follow the review to its underlying study. If it fails the third, correct or remove the entry.
What should you do next?
Pick one paragraph from your current draft — ideally one your supervisor flagged — and tag each citation-bearing sentence by claim type. Then fix the single worst mismatch: the empirical claim resting on a review, or the field-level claim resting on one primary study. That one repair will show you the pattern, and you can run the same loop across the rest of the chapter.




