Comparing Grammatical Items in Spontaneous Speech Vs. Parent Report in ASD

Background:  The Vineland Adaptive Behavior Scales (Sparrow et al., 1984) uses parental report to examine communication and social skills in children with ASDs, and exhibits high reliability when compared with children’s actual pragmatic usage (Reichow et al., 2008).   However, specific grammatical items (e.g., wh-questions, tense/aspect, negation) may be less reliable, given how widely varying these are in the language development of children with ASDs (Goodwin et al., 2012; Tager-Flusberg, 1994).  Moreover, parents may be more sensitive to some grammatical items than others.   The current study takes a longitudinal approach, comparing parental report on the Vineland of wh-questions, verb tense, and negation with the children’s concurrent spontaneous production.  Comparisons are made between high-functioning and ‘middle-functioning’ groups of children with ASD.

Objectives:  We investigate the degree to which parental report of children’s grammatical usage on the Vineland ‘matches’ those children’s production in spontaneous speech.

Methods:  Eleven ASD children (MA=31.23 months) were assessed every four months for two years.  HFA (n = 5) children had Mullen scores in the normal range at visit 1; MFA (n = 6) children produced the target grammatical items but had Mullen scores at least 1.5 SD below the normal range at visit 1.  At each visit, the children participated in a 30-minute semi-structured play session with their parent.  All speech was transcribed and coded for types and tokens of Wh-questions, progressive and regular past verb inflections, and negation.  At each visit, the Vineland was also administered, and coded for parental report of the child producing what/where, who/why, and when questions,  verbs ending in “-ing” and “-ed”,  and negatives in sentences.  Data were analyzed into three categories:  Vineland and Speech at same levels, Vineland more advanced than Speech, and Vineland less advanced than Speech.

Results:  Chi-squares were performed across visits for each group separately, and combining visits to compare groups.  For negation, 88-90% of Vineland-Speech comparisons were at the same level, and this did not differ across visits nor across subgroups.  In contrast, for wh-questions, only 46% of Vineland-Speech comparisons were at the same level; 34% showed Vineland more advanced than Speech and 20% showed Speech more advanced than Vineland.  Moreover, this distribution varied significantly by subgroup (X² (2) = 14.13, p < .001), with more Speech-more-advanced-than-Vineland children in the MFA group.  For both verb inflections, 71% of Vineland-Speech comparisons were at the same level, 11% showed Vineland more advanced than Speech, and 18% showed Speech more advanced than Vineland.  This distribution varied significantly by subgroup for the progressive inflection, (X² (2) = 10.5, p< .05), but not for the regular past. Again, more children in the MFA group produced progressive inflections in Speech than were reported to do so by their parents.

Conclusions:  Parental report in the Vineland varies in how consistently it captures children with ASDs’ production of grammatical items.  Production of negation use was highly accurate; however, production of progressive inflections, and especially wh-questions, was under-rated, particularly with children in the middle range of language use.