Participant heterogeneity induces spurious dependence that may obscure dependence patterns displayed by a covariance matrix. A parsimonious method is proposed for the reduction of this confounding effect. The method is applied to dichotomous behavioral response data of participants diagnosed with schizophrenia, as assessed by the Pragmatic Protocol (J. Speech Hear. Disord. 1987; 52:105-119).
- Pragmatics of language