In meta-analysis, what does the I^2 statistic measure?

In meta-analysis, what does the I^2 statistic measure?

• A. The percentage of total variation across studies due to heterogeneity rather than chance.
• B. The overall effect size across studies.
• C. The probability that the observed results are due to chance alone.
• D. The number of studies included in the meta-analysis.

Answer: (A)

In meta-analysis, the I^2 statistic measures the proportion of variation in effect estimates across studies that is due to real differences between studies (heterogeneity) rather than random sampling error. Put simply, it tells you how much of the observed differences in study results reflect true diversity in effects rather than just chance. It’s expressed as a percentage, where 0% indicates little or no heterogeneity and higher values indicate more inconsistency among study results. This helps you decide how much to trust a single pooled effect and whether a random-effects model might be more appropriate than a fixed-effect model. It’s not about the overall pooled effect size, nor is it the probability that results are due to chance (that would be a p-value from a heterogeneity test), and it doesn’t count the number of studies.

In meta-analysis, the I^2 statistic measures the proportion of variation in effect estimates across studies that is due to real differences between studies (heterogeneity) rather than random sampling error. Put simply, it tells you how much of the observed differences in study results reflect true diversity in effects rather than just chance. It’s expressed as a percentage, where 0% indicates little or no heterogeneity and higher values indicate more inconsistency among study results. This helps you decide how much to trust a single pooled effect and whether a random-effects model might be more appropriate than a fixed-effect model. It’s not about the overall pooled effect size, nor is it the probability that results are due to chance (that would be a p-value from a heterogeneity test), and it doesn’t count the number of studies.