Measuring variability in early child language: don't shoot the messenger.

Feldman et al. criticize the MacArthur Communicative Development Inventories (CDIs) as having too much variability, too little stability, and insufficient ability to predict early language delay. We present data showing that these characteristics of the CDI are authentic reflections of individual differences in early language development rather than measurement deficiencies. We also respond to their critical assertions concerning sociodemographic influences on the CDI scores.