sci-results-writer

name

sci-results-writer

description

Draft, revise, or audit English SCI Results sections from study findings, statistical outputs, tables, figures, forest plots, RCS plots, regression models, subgroup analyses, sensitivity analyses, clinical outcomes, or reviewer comments. Use when the user asks to convert raw findings into manuscript-ready Results text, keep Results separate from Discussion, report null findings, or write figure/table callouts.

SCI Results Writer

Core Rule

Report what was found, in a logical sequence, without interpreting mechanisms, implications, or recommendations. Use past tense and keep claims statistically traceable.

Read references/source-guide.md when source-derived Results rules are needed.

Workflow

1. Identify the research question, outcome hierarchy, statistical models, and table/figure order.
2. Select findings that answer prespecified questions or explain key analyses.
3. Order results to mirror Methods unless clinical logic requires a clearer sequence.
4. Draft concise paragraphs with sample size, effect estimates, 95% confidence intervals, p values, and figure/table callouts.
5. Include negative, null, unexpected, and robustness findings when relevant.
6. Remove explanation, speculation, literature comparison, and recommendations.

Common Order

1. Study flow, sample, or included studies.
2. Baseline characteristics.
3. Primary outcome.
4. Secondary outcomes.
5. Subgroup, sensitivity, nonlinear, or exploratory analyses.
6. Adverse events, bias, heterogeneity, or robustness checks.

Rules

• Use "was associated with" unless causal inference is justified.
• Put exact values near the claim.
• Refer to figures and tables in numbered order.
• Use text to highlight key findings, not to repeat entire tables.
• For RCS plots, report overall and nonlinear p values, reference point, curve direction, and uncertainty at extremes.

Output Format

Provide:

1. Results
2. Figure/table callouts
3. Notes / Items to verify