The Elicitation Method for Past Tense Verb production in Children with Specific Language Impairment and Typical Language

Document Type


Publication Date



Regular (e.g., jumped) and irregular (e.g., fell) past tense verb acquisition in children with typical language development (TL) occurs between ages 3-5. In children with specific language impairment (SLI), acquisition of these forms is extended and errors in spontaneous conversation may even continue into adulthood. However, there is a lack of consensus as to whether probed or spontaneous language samples give a more accurate representation of a child’s linguistic skills. The first aim of this study was to determine if there were differences in regular and irregular past tense verb production accuracy between two Elicitation Methods: probed vs. spontaneous language sampling. The second aim was to determine if accuracy and error patterns differed between children with SLI and children with TL. The participants included 11 children with SLI (mean age: 5 years) and 20 children with TL (mean age: 3 years 6 months). Each participant received a battery of tests to determine language status. This battery included two elicitation methods for regular and irregular past tense: a probe and a spontaneous language sample. The Rice/Wexler Test of Early Grammatical Impairment probed past tense verb production using picture prompts and a standardized verbal routine. Additionally, a language sample was recorded in which participants told three thematically related stories provided spontaneous productions. The first two stories were read by the examiner first and the child was asked to retell it. The first story was presented in the present tense. The second story was presented in the past tense. The third story was made up by the child based on the pictures and the tense was free to vary. These stories provided the language sample that was then transcribed and coded for a statistical analysis of verb production. Within and between groups ANOVAs revealed statistically significant differences between the probe and spontaneous language samples, with the probe yielding higher accuracy for regular and irregular past tense verb production in both groups. There was no significant Group effect or Group by Elicitation Method interaction. Analysis of the types of errors produced revealed a statistically significant Group by Elicitation Method interaction. Post hoc analysis found for regular past tense verbs, children with SLI produced more stem-form errors than children with TL. For irregular past tense forms, children with SLI produced more stem form errors, while children with TL produce more overregularization errors. The observed pattern of errors is consistent with inclusionary criteria for SLI, the literature, and theoretical foundations. The results add to the literature about the accuracy of probe and spontaneous language sample elicitation methods.


Johnson City, TN

This document is currently not available here.