Error consistency and the evaluation of treatment outcomes.

The consistency/variability of error substitution patterns may hold important implications for subgrouping children with speech disorders, as well as for relationships between learning and generalization patterns. There is a need to quantify and examine the range of consistency/variability within the speech disordered population as it relates to system-wide change. This investigation compared two groups of preschool children (N = 10 each) differing in the consistency/variability of errors on a variety of pre-treatment and treatment outcome measures. The Error Consistency Index (ECI), a measure of error variability across the entire phonological system, was used to identify groups at the extreme ends of the ECI distribution from a larger participant pool. Each participant was treated on three target singletons from among obstruents /s, z, f, integral, tintegral, k, g/ and liquids /l, r/ and change on these targets, as well as their generalization to untrained positions was assessed. Although there were significant differences between the variable and consistent groups on all pre-treatment measures, there were no significant group differences in target and generalization learning or in per cent consonants correct (PCC) change. These findings provide evidence to suggest that relationships observed between error variability for individual phonemes and learning of those targets may differ from those observed when consistency/variability is quantified for the entire system and change across a number of phonemes, and the system as a whole, is examined.