JCPSLP VOL 15 No 1 March 2013

at the pre-set adjusted alpha level ( p < .017) were conducted. There was a significant interaction effect for group X time for DSS per cent error rates for, personal pronoun , F (4,62) = 7.05, p < .001, N p 2 = .31, main verb , F (4,62) = 7.44, p < .001, N p 2 = .32 and sentence point , F (4,62) = 8.08, p < .001, N p 2 = .34. Simple main effect tests were completed as a follow-up to the significant interaction for each ANOVA. Findings for DSS per cent error rates in each grammatical category met the set significance level ( p < .006) for follow-up tests at post-intervention and 3 months post-intervention. Pairwise comparisons of means revealed that computer-assisted and table-top intervention facilitated significantly lower DSS per cent error rates in each category compared to no-intervention. The two intervention groups did not differ statistically ( p > .05). See Figure 3. Discussion Preschoolers with SLI who received expressive grammar intervention experienced significantly greater improvement in their grammar skills that were maintained at post- intervention and at 3 months post-intervention compared to waitlist controls. Maintenance of gains beyond the intervention period is considered an important intervention outcome indicative of development (Yoder et al., 2011). The magnitude of gain in 6 months as a result of expressive grammar intervention was above that expected for typically developing 4-year-olds at the 50th percentile, demonstrating significant accelerated growth beyond the starting point in intervention. Thus, intervention offered a therapeutic advantage over no intervention for facilitating the outcomes observed for grammatical complexity, use of morphemes, and accuracy in targeted grammatical categories. However, computer-assisted and table-top intervention resulted in similar effects on the spontaneous use of expressive grammar skills. Consequently, it is important to consider the role of expressive grammar intervention, regardless of type, for facilitating growth in these skills. The goals of language intervention for grammatical deficits are to improve children’s production and comprehension of targeted language forms (Leonard et al., 2006). By directly targeting children’s grammatical use of language forms, they become more aware of how to accurately sequence morphemes and phonemes into meaningful units. The explicit support and attention to grammatical features included in intervention may provide more time to process information and increase preschoolers’ awareness and decrease language-learning efforts during expressive grammar intervention. During computer-assisted and table-top intervention, the SLP made deliberate attempts to highlight sentence components needed to produce grammatically correct sentences. The syntactic slot-filler approach and emphatic stress were equally effective in highlighting these components for the sample of preschoolers. Implementing these techniques in a drill-play format with modelling and repetition resulted in multiple opportunities for focused practice (3000 production opportunities for the two interventions). Previous researchers highlight the necessity of explicit, repeated exposures to target forms that address grammatical productions for this population (Cleave & Rice, 1997). Computer-assisted intervention may provide motivation and increased tolerance for repetition in some preschoolers with SLI and should be considered a viable alternative to table-top intervention, where appropriate (cf. Washington et al., 2011).

The negative changes observed in DSS change scores for preschoolers in the waitlist control group on average were indicative of unsuccessful attempts at creating grammatically complex and correct productions and not a regression in these skills. An examination of the per cent error rates revealed that these children attempted to produce accurate and complex personal pronouns, main verbs and achieve the sentence points, but were unsuccessful in these attempts, resulting in the higher per cent error rates compared to cohorts in intervention. Lee (1974) and Lee and Canter (1971) suggested that as children‘s language skills develop, unsuccessful attempts at more complex productions in spontaneous language are expected. Thus, errors in syntactic productions are a normal part of grammatical development for young children with typically developing language skills. In fact for 5-year-olds, a 0.12-point decline in DSS scores at the 50th percentile is expected in 6 months. Since these errors are anticipated for typical preschoolers, errors for preschoolers who do not have typical language skills and are not receiving intervention would not be unexpected. Clinical implications The current findings expand on those reported in the Washington et al. (2011) study by providing evidence that grammatical language interventions were associated with accelerated growth to “within normal limits” for grammar development. The inclusion of computer-assisted and table-top intervention techniques, including specific step-by-step procedures, may be important in achieving this growth during SLP-led interventions for this population. Alternatively, preschoolers with SLI who do not receive intervention are at a significantly greater risk for not achieving good outcomes. Limitations and future directions Participants included in this study demonstrated specific expressive grammar deficits, thus application of these findings to other preschoolers with SLI with receptive difficulties is limited. The mixed evidence regarding the effectiveness of expressive grammar intervention for children with concomitant receptive and expressive grammatical deficits (cf. Law et al., 2012) suggests additional research is needed. Another limitation is that DSS techniques are time consuming and scores are narrower in their representation of language skills (Lee & Canter, 1971). The inclusion of MLU in this study offered another measure of linguistic gains that might not have been captured in DSS scores. Future research with other children with SLI with receptive language deficits should expand the measures used to establish acceleration in expressive grammar and perhaps also consider sampling beyond a clinical conversation with an adult. Conclusion Expressive grammar intervention offers a therapeutic advantage over no intervention for the enhanced development of spontaneous language skills in preschoolers with expressive language impairment. Acknowledgments The authors would like to thank the families, children, speech-language pathologists, graduate SLP students and research assistants who made this research possible. This study was funded by the Canadian Language and Literacy Research Network, the Ontario Graduate Scholarship, the

11

JCPSLP Volume 15, Number 1 2013

www.speechpathologyaustralia.org.au