Common Research Mistakes (and How to Avoid Them)

Most committee revisions and reviewer rejections trace back to a small set of recurring, avoidable mistakes — not weak ideas, but weak execution of the methodology around them.

Design & planning stage

• Vague research questions that can't be operationalized into measurable variables or clear inclusion criteria.
• No sample size justification — see our sample size guide.
• Mismatched design and question — e.g., using a cross-sectional design to make causal claims. See our study design comparison.

Data collection & analysis stage

• Choosing a statistical test after seeing the data, rather than pre-specifying analysis in the protocol — a practice that invites accusations of p-hacking.
• Ignoring test assumptions (normality, independence, equal variances) without checking or addressing violations.
• Confusing statistical and practical significance — see our guide to confidence intervals and p-values.

Writing & reporting stage

• Overclaiming conclusions beyond what the data and design actually support.
• Incomplete methods reporting that a reader couldn't use to reproduce the study.
• Inconsistent numbers between the abstract, results, tables, and figures — a fast way to lose a reviewer's confidence in everything else.

The pattern behind most of these

Nearly every item above traces back to decisions made too late — choosing a test after the data is in, defining criteria after screening has started, writing methods from memory instead of from a protocol. Locking in decisions upfront, in a written protocol, prevents almost all of them.