Measuring change in speech over time: words or segments?

Background: Measuring change is an integral part of the therapy process; what is measured and how it is measured validates intervention. Commonly in the literature, phonological analysis has been carried out using measures that focus on segments, e.g. Percent Consonants Correct (PCC) rather than whole words. However, more recent studies suggest that analysis of children’s phonological development should include measures that track change towards correct word forms, rather than just correctness of individual consonants. Objective: To determine whether a measure of percentage of whole word proximity is a more sensitive measure of change in children’s developing phonological systems than PCC. Method: Twelve children aged between 3;02 and 4;09 with SSDs participated in this study. At initial assessment speech accuracy in single words was assessed using a standardised test of phonology, the Diagnostic Evaluation of Articulation and Phonology (DEAP) (Dodd et al 2002). Speech accuracy in connected speech was assessed using a constrained speech sample elicited using pictures from the Renfrew Action Picture Test (RAPT) (Renfrew 2003). Both assessments were re-administered 8 weeks later. PCC and PWP were calculated for each assessment at initial and follow up assessment, and results compared to establish whether one measure was more sensitive to change in the samples than the other. Results: No one measure was more sensitive than the other for all cases. Individual analyses exemplify how change is better captured by one method or the other in different cases. Conclusions: The effectiveness of this study is limited by smaller than expected amounts of change over the two time points. However, analysis of individual cases where some change occurred showed that PCC was more accurate for some and PWP was more accurate for others. Clinical implications are discussed.