The role of noninitial clusters in the Children's Test of Nonword Repetition: Evidence from children with language impairment and typically developing children.

The Children's Test of Nonword Repetition (CNRep) is one of the most popular tests of nonword repetition. The test is composed of nonwords of different length, and normative data suggest that children experience more difficulties in repeating long nonwords. An analysis of the distribution of phonological clusters in the test shows that noninitial clusters are unequally distributed in the test: They only appear in long nonwords (four and five syllables). For this reason, we hypothesized that the difficulties children encounter with long nonwords may be influenced by the phonological complexity of the clusters and not just by the challenge for working memory associated with length. To test the hypothesis, we compared repetition performance in long nonwords with and without a noninitial cluster in 18 children with language impairment and 18 typically developing children. The analysis shows that long nonwords with noninitial clusters are repeated less accurately by both groups. In addition, there was an interaction between cluster and age: The effect of cluster is absent in younger children and gradually increases with age. These findings suggest that phonological complexity may be having an impact on the length effect normally observed in the CNRep, and this impact may be particularly evident in older children.