JCPSLP VOL 15 No 1 March 2013

Table 1. Preschoolers’ pre-test characteristics Total sample

Computer-assisted

Table-top (n = 11) 4;2 to 4;10 Female = 3

No intervention

(n = 34)

(n = 11)

(n = 12)

Age-range

3;6 to 4;11 Female = 7 Male = 27

3;11 to 4;6 Female = 3

3;6 to 4;11 Female = 1 Male = 11 99.42 (4.30) 99.75 (2.90) *8.58 (4.99) *4.51 (.54) *5.01 (.70) 106.92 (8.24)

Gender-distribution

Male = 8

Male = 8

Receptive-word (PPVT-IIIB) Receptive-sentence (CELF-P)

101.85 (4.88) 101.21 (6.95) *10.26 (3.97)

103.64 (5.71) 103.36 (8.65) *10.09 (2.30) *4.81 (1.08)

102.73 (3.77) 100.36 (8.10) *12.27 (3.38)

Expressive -SPELT-P

Expressive-DSS Expressive-MLU

*4.84 (.82) *5.26 (.76)

*5.21 (.66) *5.62 (.75)

*5.17 (.73)

Nonverbal IQ (KBIT-2 matrices subtest)

109.15 (10.10)

112.27 (12.35)

108.45 (9.61)

*Raw score. Note. Means reported with standard deviations in parentheses. Standard scores are reported for all measures, except for expressive-language. Receptive-word (PPVT-IIIB; Dunn & Dunn, 1997), receptive-sentence (CELF-P; Wiig, Secord, & Semel, 1992), expressive-SPELT-P (Werner & Krescheck, 1983), expressive-DSS (Lee, 1974), expressive-MLU (Brown, 1973; Miller, 1981), nonverbal IQ (KBIT-2; Kaufman & Kaufman, 2004).

The current study The hypothesised link observed between grammatical errors and processing constraints suggests that SLI has a complex nature that necessitates grammatical language interventions (Leonard et al., 2007). A secondary analysis of the Washington et al. (2011) data was completed for the DSS scored language samples to determine if expressive grammar intervention facilitated accelerated growth (i.e., to within normal limits), representing performance outside the pre-test range for the spontaneous use of grammar skills better than no intervention. Additionally, the authors tracked decreases in per cent error rates for targeted grammatical categories (e.g., personal pronoun , main verb , sentence point ) for intervention and no intervention groups. The following research questions were addressed: 1. Do computer-assisted and table-top intervention result in accelerated growth in grammatical development compared to no intervention? 2. Do computer-assisted and table-top intervention result in significantly lower per cent error rates for targeted grammatical categories compared to no intervention? Method Participants Following ethical approval, 34, 3- to 4-year-olds ( M = 4;4 months, SD = 5 months) who were randomly selected from an intervention waitlist at a government-funded preschool speech-and-language initiative in Ontario, Canada met the Washington et al. (2011) study criteria (see below). Their parents identified them as Caucasian (n = 32), Asian (n = 1), or other (n = 1) and monolingual English speakers. The sample included 27 boys and 7 girls, residing in urban and rural regions. All participants met the diagnostic criteria for SLI of an expressive nature as outlined in the Washington et al. (2011) study. These included normal hearing range, normal receptive language and nonverbal cognition, i.e., one standard deviation from the mean on the Peabody Picture Vocabulary Test-IIIB (PPVT-IIIB; Dunn & Dunn, 1997); the receptive portion of the Clinical Evaluation of Language Fundamentals-Preschool (CELF-P; Wiig, Secord, & Semel, 1992); and the Kaufman Brief Intelligence Test – 2: Matrices Subtest (KBIT-2; Kaufman & Kaufman, 2004). For expressive grammar, children demonstrated skills at or below the 10th percentile, on the Structured Photographic Expressive Language Test-Preschool (SPELT-P; Werner

& Kresheck, 1983) and a spontaneous language sample scored for grammatical complexity (see Procedures). Language assessment results revealed that all participants experienced grammatical deficits affecting the accurate production of 3rd person singular present progressive sentences containing a subject-verb-object (e.g., The boy + is eating + a hot-dog or He + is eating + a hot-dog). Following consent to participate, 22 of the 34 preschoolers were randomly selected to receive intervention, leaving 12 participants to remain on the waitlist for intervention. This selection allowed for equal numbers in each intervention group and in the control group (see Table 1). Half of the participants in intervention received computer-assisted intervention (n = 11) and the other half received table-top intervention (n = 11). Results of Univariate ANOVAs revealed non significant between-group differences for age ( p = .126), gender ( p = .902), nonverbal IQ ( p = .443), receptive language word-level ( p = .087), and receptive language sentence-level ( p = .374), and expressive grammar on the SPELT-P ( p = .080) and DSS ( p = .127). There was no study attrition. Table 1 describes participants’ demographic information. Speech-language pathologists (SLPs) or graduate SLP students evaluated participants’ language and cognitive skills during a 90-minute individual assessment in a clinical setting to determine participant suitability (pre-intervention). This session included the collection of a 45-minute spontaneous language sample during play (using a standard procedure and including a dollhouse, toy household objects and people and the Spot Bakes a Cake [Hill, 2003] and Where’s Spot [Hill, 2000] books). At least 100 intelligible utterances were collected and digitally recorded from the participants at post-intervention and 3 months post-intervention, representing a break in intervention (cf. Washington et al., 2011), to establish spontaneous language outcomes associated with intervention. Language samples were transcribed and coded by assessors who were blind to group assignment and assessment time points. DSS procedures were used to analyse the language samples (Lee, 1974). To obtain a DSS score, 50 consecutive utterances containing a subject and verb were selected. Each utterance was scored for grammatical accuracy (i.e., the DSS sentence point) and the eight DSS Procedures Assessment

8

JCPSLP Volume 15, Number 1 2013

Journal of Clinical Practice in Speech-Language Pathology