Limitations, honestly

Every study has limitations. Reviewers will find them anyway. Stating them yourself, clearly, makes the paper more credible, not less.

What counts as a limitation

• Sample: small, from one place, one population, one dataset.
• Design: observational, not randomised; no blinding; short follow-up.
• Measures: proxies for what you really care about.
• Generalisation: settings the results may not transfer to.
• Analysis: assumptions that may not hold, choices made after seeing data.

A good limitation paragraph

For each important limitation: what it is, how it might affect the results (in which direction, if you can tell), and what would fix it in future work.

ExampleWeak: "This study has some limitations." Strong: "All images came from one region in a single season, so accuracy may be lower under different lighting and leaf colours. Testing on images from other regions is the most important next step."

Don't bury them, don't drown in them

Give limitations their own paragraph or subsection. Cover the ones that matter most for your main claim; don't list fifteen trivial ones to look humble.

Limitations shape the claim

If a limitation is serious, your conclusion should be smaller to match. That's the link to the next lesson.

RememberName each important limitation, say how it might change the result, and how future work could fix it.