How to Critically Evaluate a Source (Not Just Summarise It)

A practical guide for students · Last updated 11 October 2026

"Critically evaluate the evidence" is one of the most common instructions on a marking rubric, and one of the most commonly missed. Most students report what a study found. Markers are asking something different: how much weight that finding can actually bear.

Summary vs evaluation — the actual difference

Summary

A 2013 meta-analysis found that CBT produced moderate-to-large effect sizes for depression, comparable to medication in mild-to-moderate cases.

Evaluation

That meta-analysis pooled over 100 trials of widely varying quality and delivery format, and its authors themselves flagged substantial heterogeneity — so the headline effect size likely masks real differences between, say, therapist-led and app-based delivery, rather than describing one consistent effect.

Both sentences are accurate. Only the second one is doing the thing the rubric is actually asking for: interrogating whether the finding deserves the confidence it's being cited with, rather than just reporting that it exists.

Five questions that actually do the evaluating

What was the study design, and what can that design actually prove? A randomised controlled trial can support a causal claim; a correlational or cross-sectional study generally can't, however large it is.
How big and representative was the sample? A finding from 40 undergraduates and a finding from a multi-site trial of 4,000 adults are not equally strong evidence for the same claim, even when they point the same direction.
Who funded it, and does that create a conflict of interest? Industry-funded drug trials, for instance, are consistently more likely to report favourable results than independently funded ones — worth naming when it's relevant, not as an accusation, but as context.
Does the stated conclusion actually match the data reported? It's common for a paper's abstract to claim more certainty than its own results section supports. Read past the abstract before you cite the claim.
How does this sit against the rest of the literature? Is this the consensus finding, a notable outlier, or part of a genuinely contested debate? A single study is weak evidence for anything until you know where it sits.

This is the exact distinction LitMark's Hit the Mark checks for: it marks a criterion like "critically evaluate the evidence" as only partly met when an essay reports findings without assessing their quality — not just whether the right papers were mentioned.

The honest takeaway: naming a study isn't evaluating it. If your sentence would still be true with "and that's definitely right" deleted from the end, it probably isn't evaluation yet.

← Back to LitMark