The article "Basics of Meta-analysis: I² is not an absolute measure of heterogeneity" by Borenstein, Higgins, Hedges, and Rothstein discusses the common misconception that the $I^2$ statistic provides information about the range of effect sizes in a meta-analysis. The authors clarify that $I^2$ is a proportion, representing the ratio of the variance in true effects to the variance in observed effects, and does not provide an absolute measure of heterogeneity. They emphasize that $I^2$ is useful for understanding the dispersion of observed effects relative to sampling error but not for quantifying the actual range of true effects. To address this, the authors propose reporting prediction intervals, which provide the absolute range of true effects. The article includes examples and mathematical derivations to illustrate these concepts and explains why $I^2$ is often misinterpreted. The authors conclude that while $I^2$ can be useful in certain contexts, it should not be used as a substitute for reporting prediction intervals when researchers are interested in the substantive implications of heterogeneity.The article "Basics of Meta-analysis: I² is not an absolute measure of heterogeneity" by Borenstein, Higgins, Hedges, and Rothstein discusses the common misconception that the $I^2$ statistic provides information about the range of effect sizes in a meta-analysis. The authors clarify that $I^2$ is a proportion, representing the ratio of the variance in true effects to the variance in observed effects, and does not provide an absolute measure of heterogeneity. They emphasize that $I^2$ is useful for understanding the dispersion of observed effects relative to sampling error but not for quantifying the actual range of true effects. To address this, the authors propose reporting prediction intervals, which provide the absolute range of true effects. The article includes examples and mathematical derivations to illustrate these concepts and explains why $I^2$ is often misinterpreted. The authors conclude that while $I^2$ can be useful in certain contexts, it should not be used as a substitute for reporting prediction intervals when researchers are interested in the substantive implications of heterogeneity.