Generalized linear mixed models are frequently applied to data with clustered categorical outcomes. The effect of clustering on the response is often difficult to practically assess partly because it is reported on a scale on which comparisons with regression parameters are difficult to make. This article proposes confidence intervals for estimating the heterogeneity due to clustering on a scale that is easy to interpret. The performance of the proposed asymptotic intervals and percentile bootstrap intervals are compared by simulations and in an application.