τ² (variation between studies)

τ² (tau squared) is the estimated variance of the true effects between studies, on the scale of the effect. τ is its square root: the standard deviation of the true effects. Statable estimates τ² with REML and gives an interval. Unlike I², τ² does not depend on the precision of the studies.

Example: τ = 0.56 for log risk ratios: the true effects have a standard deviation of about 0.56 on the log scale.

Caution: With fewer than about five studies τ² is very uncertain and the interval very wide.