Draft or revise sections of an academic manuscript in epidemiology. Use when writing methods, results, or discussion sections, or when updating manuscript text to reflect analysis changes. Applies epidemiological writing conventions and proper statistical language.
| Instead of | Write |
|---|---|
| "X causes Y" | "X was associated with Y" (observational data) |
| "No effect" | "No clear evidence of a systematic difference" |
| "X significantly affected Y" | "X was associated with a [magnitude] [direction] in Y (95% CI: ...)" |
| "Proved" |
| "Found evidence consistent with" |
| "Failed to reach significance" | "The estimate was imprecise (95% CI: ... to ...)" |
Structure:
Include enough detail for independent replication. State software versions. Reference the analysis code repository.
Structure:
Lead with the finding, then the evidence:
Reference figures and tables actively:
Structure:
For limitations:
| Term | Use precisely |
|---|---|
| Confounding | Requires a specific causal structure; use "adjustment" when structure uncertain |
| Bias | Specify type and direction; not a synonym for "error" |
| Random effects | Describe what they represent substantively, not just statistically |
| Accuracy | General performance; distinguish from calibration and sharpness |
| Calibration | Reliability of probabilistic predictions specifically |
| Significance | Avoid; report estimates and CIs instead |
When updating manuscript text: