A comparison of three methods for obtaining articulatory responses.

This study compared speech-sound productions of 18 children with articulation errors obtained with three different speech-sampling methods. The same 20 stimulus words were elicited by each of three sampling methods: (1) Continuous Speech Task, (2) Modeled Continuous Speech Task, and (3) Spontaneous Picture-Naming Task. Subjects performed differently on each of the speaking tasks. Statistical treatment of the mean score data revealed that significantly more errors were elicited with the Continuous Speech Task than with the Modeled Continuous Speech Task or the Spontaneous Picture-Naming Task. In addition, significantly more errors were elicited with the Modeled Continuous Speech Task than the Spontaneous Picture-Naming Task. The Continuous Speech Task required the greatest amount of time to administer, followed by the Modeled Continuous Speech Task, and finally the Spontaneous Picture-Naming Task.