Meta-analysis

A meta-analysis is a statistical procedure that quantitatively synthesises results from multiple independent studies addressing the same research question, yielding a pooled effect estimate with narrower confidence intervals than any individual study alone. Gene V. Glass coined the term in 1976: effect sizes are extracted from each study, weighted (typically by inverse variance), and combined into a weighted average — visualised in a forest plot, where each horizontal line represents one study and the diamond at the bottom represents the pooled result. Heterogeneity — the degree to which true effects vary across studies beyond chance — is quantified by I²: below 25% indicates low heterogeneity, 50–75% moderate, and above 75% high, though I² is sensitive to the number of included studies. Publication bias, assessed via funnel plots, Egger's test, or trim-and-fill, can skew pooled estimates. In longevity research, meta-analyses aggregate observational cohorts or trials to detect modest effect sizes — such as the survival benefit of physical activity or the association between telomere length and mortality — that individual studies lack power to resolve. Their conclusions are only as valid as the studies they pool: shared systematic biases are amplified, not corrected.