Meta-Analysis Reasoning Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: aa352baf-3d08-4206-9ac1-69d979aa593c

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Reason through pooling the studies in {study_pool} into a meta-analytic estimate, guarding against garbage-in. (1) Verify the studies are poolable: do they measure the same construct with comparable outcomes in comparable populations? Pooling apples and oranges produces a precise, meaningless average — the combinability judgment precedes any calculation. (2) Extract a comparable effect size and its variance from each study, standardizing metrics where needed and noting where you had to estimate a missing variance. (3) Assess heterogeneity before choosing a model: if effects vary more than sampling error explains, a fixed-effect model is inappropriate — use a random-effects model and report the between-study variance, don't hide it. (4) Compute the pooled estimate weighting each study by precision, and report the confidence interval, but present the heterogeneity alongside it so the pooled number isn't read as more settled than it is. (5) Probe for publication bias (funnel plot asymmetry, small-study effects) since missing null results inflate the pooled effect. (6) Run the influence check: does removing any single study or the smallest studies change the conclusion? State the pooled result with its caveats, and whether the evidence base is robust or fragile.

Reason through pooling the studies in {study_pool} into a meta-analytic estimate, guarding against garbage-in. (1) Verify the studies are poolable: do they measure the same construct with comparable outcomes in comparable populations? Pooling apples and oranges produces a precise, meaningless average — the combinability judgment precedes any calculation. (2) Extract a comparable effect size and its variance from each study, standardizing metrics where needed and noting where you had to estimate a missing variance. (3) Assess heterogeneity before choosing a model: if effects vary more than sampling error explains, a fixed-effect model is inappropriate — use a random-effects model and report the between-study variance, don't hide it. (4) Compute the pooled estimate weighting each study by precision, and report the confidence interval, but present the heterogeneity alongside it so the pooled number isn't read as more settled than it is. (5) Probe for publication bias (funnel plot asymmetry, small-study effects) since missing null results inflate the pooled effect. (6) Run the influence check: does removing any single study or the smallest studies change the conclusion? State the pooled result with its caveats, and whether the evidence base is robust or fragile.

Variables

Name	Type	Required	Trust level
study_pool		yes

meta-analysispooled-estimatesheterogeneitychain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18