ACQ Vol 13 No1 2011

additional production probes was 0%, except for Aaron who produced final consonants with 12.5% accuracy. Reliability Treatment fidelity was measured by asking an independent observer to observe videotapes of a random selection of 25% of the total intervention sessions implemented. These sessions were reviewed to ensure inclusion of important treatment elements. Treatment fidelity for all PAS and MS intervention sessions was 100%. Reliability of phonetic transcription was measured by asking a second person experienced in transcribing children’s speech to independently transcribe a random selection of 25% of the total consonant cluster probes administered. Inter-judge agreement for broad phonetic transcription of the consonant clusters was 94%. Results Accuracy on consonant cluster probe In order to investigate the effect of the two different types of intervention on the production of consonant clusters, accuracy on the cluster probe was calculated for each of the four participants at three different points in time (pre- intervention, immediately post-intervention, and 3 months post-intervention). Figure 1 shows the percent correct for the word-initial clusters that were directly targeted in the first intervention block by participants who received PAS intervention. None of the participants in either intervention group could produce this cluster type, pre-intervention. Immediately post-intervention, one participant from each intervention group had reached 30% accuracy on this cluster type, so there was no difference between the groups on this measure. Thus, Ben, who received MS intervention, improved in his production of word-initial /sp/, /st/, and /sl/ clusters over the course of the intervention, even though the intervention he received did not target this particular error pattern. Both these children show continued improvement in their production of /sp/, /st/, and /sl/ clusters when tested 3 months after the end of the intervention. It should be noted that Aaron, who received PAS intervention, did not improve in his production of word-initial consonant clusters when tested immediately post-intervention even though he had received intervention that directly targeted this cluster type.

Figure 2 shows the percent correct for word-initial clusters that were not directly targeted by participants in either intervention group. The two children who received PAS intervention made some improvement on the non-target word-initial clusters, but this was evident only when tested 3 months post-intervention. Neither child who received the MS intervention program improved in their production of the non-target word-initial clusters between the pre-intervention testing and the 3-month post-intervention follow-up. However, it should be noted that Ben, who received MS intervention, could accurately produce word-initial labial+/l/ clusters and /s/+nasal clusters pre-intervention. His pre- intervention accuracy on /s/+nasal clusters may account for his improvement in the production of /sp/, /st/ and /sl/ clusters even though he did not receive intervention that directly targeted these clusters.

Non-target word-initial clusters

100 90 80 70 60 50 40 30 20 10 0

3 months post post-intervention pre-intervention

Percent correct

Mike

Matt

Ben

Aaron

PAS intervention

MS intervention

Figure 2. Percent correct for word-initial clusters that were not directly targeted by either intervention Figure 3 reports the percent correct for word-final clusters that were not directly targeted by participants in either intervention group. Because participants in the MS group had received indirect instruction on a variety of consonant clusters at the ends of words as part of the instruction they received in word-final morphology, it was predicted that they would make greater improvement than children in the PAS group in their production accuracy of word-final clusters. Surprisingly, only the participants that received the PAS intervention (Aaron and Mike) improved in their accuracy of word-final clusters. Both children in this intervention group showed improved accuracy on final clusters when tested immediately post-intervention. However, this improvement was not maintained when Aaron was tested at the 3-month follow-up. Mike, on the other hand, continued to improve in his production of word-final clusters. When tested 3 months post-intervention, he accurately produced 6 out of 8 final clusters with the remaining two clusters being produced as 2-element clusters where previously they had been reduced to a single consonant. The only errors in Mike’s production of word-final clusters were in place of articulation ( desk produced as / dEst / and wings produced as / wImz /). Accuracy on singleton target phonological error pattern All participants received six sessions of intervention on a target phonological error pattern that did not involve consonant clusters (final consonant deletion for Aaron and

Target word-initial clusters

100 90 80 70 60 50 40 30 20 10 0

3 months post post-intervention pre-intervention

Percent correct

Mike

Matt

Ben

Aaron

PAS intervention

MS intervention

Figure 1. Percent correct for word-initial clusters that were directly targeted by participants who received PAS intervention. Scores are shown for the two children who received PAS intervention and for the two children who received MS intervention.

23

ACQ Volume 13, Number 1 2011

www.speechpathologyaustralia.org.au