Between-study heterogeneity refers to the variation in true effect sizes across studies in a meta-analysis. The random-effects model accounts for this by estimating tau-squared, which quantifies the variance in true effects. High heterogeneity might indicate distinct subgroups of studies or that pooling results is meaningless. Quantifying and analyzing heterogeneity is crucial for assessing the trustworthiness of an overall effect estimate. Cochran's Q statistic, a weighted sum of squares, is traditionally used to distinguish between sampling error and true heterogeneity. It measures deviations of individual study effects from the summary effect, weighted by study precision. An approximately chi-squared distribution is assumed for Q, allowing for hypothesis testing of heterogeneity. However, Q is influenced by the number of studies and their precision, limiting its reliability as the sole indicator. The I-squared statistic, derived from Q, represents the percentage of variability not due to sampling error. It provides a more interpretable measure of heterogeneity, with common benchmarks for low, moderate, and substantial levels. The H-squared statistic is another measure based on Q, indicating the ratio of observed to expected variance due to sampling error. Tau-squared and its square root, tau, quantify the variance and standard deviation of true effect sizes, respectively. While useful, tau-squared can be difficult to interpret practically. Prediction intervals, which consider both heterogeneity variance and pooled effect standard error, offer a more informative way to represent the range of future study effects. Therefore, reporting I-squared with confidence intervals and prediction intervals is recommended for assessing heterogeneity.