Defining optimal language input for preschool children with hearing loss : a study of the quantity, styles of oral interaction and associated factors

Abstract

The potential impact of language input on language outcomes in typically developing children with normal hearing (CwNH) has been studied for decades. However, most research has explored structured language environment/activities, with naturalistic observations of children’s exposure to adult oral communication being rare, in-home observations often very limited, and children with hearing loss (CwHL) generally not considered. The present study compared language input in preschool children with and without hearing loss and the association of language input with language outcomes. Contextual factors, including family and child demographics–often neglected in previous studies–were considered to explore variations in language input. Language input was defined in terms of the quantity and styles of oral interaction (SOI) used by parents/caregivers. Oral interactions in natural environments were recorded using LENA (Language ENvironment Analysis) technology. LENA software automatically calculated number of adult words (AWs) and conversational turns (CTs). SOI are manually extracted from recordings. A systematic review summarizes evidence for differences in amount of language input between CwNH and CwHL and association with language outcomes, presenting current knowledge, and identifying research gaps. This dissertation comprises two papers on the bases of the findings of systematic review. First, natural language input was compared between CwHL and CwNH aged 2-5 years. Effects of contextual factors (maternal level of education [MLE], socioeconomic level of deprivation [SLD], child’s age, number of adults at home, number of siblings, child’s birth order) on language input were examined to learn more about variations in natural language input. CwHL were evaluated in terms of stability in language input from day-to-day natural oral interactions with parents/caregivers. These data were utilized statistically analysed in paper 1. Second paper, language input was analyzed in CwNH aged 2-5 years for the investigation whether children were exposed to natural language input differently based on MLE and SLD, and the association with language outcomes. Methodological limitations indicated in the review–including sample selection bias, lack of information on control of confounders, assessment protocols, limited duration and reporting of nature of observation/recordings, and lack of analysis for SOI in natural settings–were addressed. Day-long natural language input was obtained using LENA recordings. The Preschool Language Scales-5th edition was used to document language outcomes, and the New Zealand socioeconomic deprivation index and education classification systems were used as MLE and SLD measures. The review of previous studies highlighted differences in amount of language input between CwHL and CwNH. CTs and optimal SOI were found positively associated with better language outcomes. Results from the two new studies indicated quantity of language input (AWs/hour, CTs/hour) did not differ significantly between CwHL and CwNH. Results affirmed that parents/caregivers of CwHL used less optimal and more moderate and suboptimal SOI than those of CwNH. Contextual factors such as higher MLE were also significantly associated with more AWs/hour, CTs/hour, and more frequent use of optimal SOI. Recordings from four full typical days (two week and two weekend days) in CwHL confirmed the stability in language input exposure across different days. However, quantity of language input–measured by number of AWs/hour and CTs/hour– between high and low MLE and SLD groups also indicated that CwNH of high MLE and low SLD families were exposed to significantly more AWs/hour and CTs/hour and a higher proportion of optimal SOI. These language inputs were significantly associated with better language outcomes. This dissertation represents the first attempt to investigate differences in quantity of language input and SOI in CwHL and CwNH with the association of contextual factors. Families of CwHL, as well as low MLE and high SLD families, exposed their children to fewer language inputs and less optimal SOI compared with families of CwNH and high MLE and low SLD families. These family contextual factors significantly influenced exposure to language input. Less exposure to language input and optimal SOI was associated with poorer language outcomes.