Reframe a non-significant result using effect size and power

Role

quantitative analyst

Task

The study of {{research_question}} produced {{reported_result}} with n = {{sample_size}}. Rewrite the conclusion so it reports the estimated effect with its interval, states which practically important effect sizes remain compatible with the data, and distinguishes an underpowered null from evidence of absence. Add the sample size that would be needed to detect an effect of {{minimum_effect}}.

Context

Stakeholders are about to cancel {{initiative_name}} on the basis that the analysis 'found nothing'.

Inputs

• Reported estimate, interval, and p-value: {{reported_result}}
• Sample size {{sample_size}} and design
• Smallest effect that would matter: {{minimum_effect}}
• Variability estimate used for power calculation

Constraints

• Never state that no effect exists because the threshold was not crossed
• Report the interval and describe its practically relevant endpoints
• Show the power or required-sample reasoning, including inputs used
• Recommend a decision that reflects uncertainty rather than a binary verdict

Output Format

A rewritten conclusion paragraph, a compatible-effects note, and a power/sample-size box with inputs and result.

Quality Criteria

• Interval-based language replaces threshold language
• Practically important effects assessed for compatibility
• Power inputs stated so the calculation is checkable
• Decision recommendation acknowledges uncertainty