Description:

This Master Prompt would focus on determining whether meta-analytic conclusions remain credible under publication bias, selective reporting, influential studies and alternative analytical decisions. It would guide researchers through funnel plots, Egger's regression, Deeks' test where appropriate, fail-safe approaches and other suitable assessments of small-study effects.

It would then extend into leave-one-out analysis, influence diagnostics, outlier assessment, alternative model specifications, exclusion of high-risk-of-bias studies and robustness checks. Importantly, researchers would be instructed not to interpret funnel-plot asymmetry automatically as publication bias and not to equate a nonsignificant bias test with proof that bias is absent. Sensitivity analysis is particularly important because the uploaded literature contains examples where excluding a single dataset materially changes statistical conclusions.