Effects of a conversation intervention on expressive vocabulary of young children

Abstract

Entering school with a low level of vocabulary has been associated with poor academic outcomes (Biemiller, 1999). Young children with larger vocabularies typically live with parents who talk more and with greater sophistication. Introducing this kind of talk into early childhood learning environments may be a strategy for increasing children's vocabulary knowledge (Dickinson, 1991). Current observational research has consistently recorded that preschool and elementary teachers rarely talk in depth with children. This quasi-experimental study investigated the effects of a vocabulary enriched conversation intervention (500 minutes over 3 months) that took place between an adult and pairs of prekindergarten students. Three vocabulary measures were used for pre- and post-testing; a standardized expressive vocabulary test (EVT), a language sample based measure of vocabulary diversity (D) and a definition task (Schwanenflugel et al, 1996). Conversation guidelines included the intentional introduction of novel vocabulary in context, explicit definitions, joint attention tasks with varied prompts and allowance for child initiated topics. Students in the experimental group had significantly higher post-test scores on the EVT. When scores from children with lower levels of a priori vocabulary were separately analyzed the effect of the intervention was significant for both D and EVT scores. The groups did not differ significantly on the definition task. The results of the present investigation indicated that a school-based practice of enriched conversation could be an effective strategy for increasing the vocabulary levels of children starting out with less developed vocabularies. The first study examined the validity of D as a measure of lexical diversity. Messick’s (1989) framework for unitary construct validity was used to evaluate the evidence collected from the study. D performed well on internal consistency and test-retest analyses and was moderately related to the EVT. However D scores varied in relation to different prompts used in language sampling. Overall evidence suggested that D is a valid and reliable measure of lexical diversity and may be useful to educators, language therapists and researchers.