Browsing by Author "Anttila, Elina"

Children’s language skills can be evaluated by formal testing, parent report or assessing spontaneous speech in recorded language samples. This study focused on measures of lexical diversity determined from children’s spontaneous speech. These measures are one way to assess children’s lexical skills. However, there is only little research on the relation between the most recent lexical diversity measures and validated measures of lexical skills. The first aim of this study was to find out if there is a relation between the lexical diversity assessed from spontaneous speech of preschool children and lexical skills assessed by validated methods. The second aim was to examine if the measures of lexical diversity predict the lexical skills of a child assessed by validated measures and if so, which of the measures is the best predictor of a child’s lexical skills. The data of the study was collected in LEINIKKI research project. The participants (N = 68) were 2;6–4;0-year-old, healthy, Finnish-speaking children. The spontaneous speech of the participants was recorded in free play settings, the speech of the children was transcribed and token, type, type-token ratio (TTR) and moving-average type-token ratio (MATTR) values were determined from the first ten minutes of each language sample. These values were used as measures of the children’s lexical diversity. The following validated methods were used to assess the children’s lexical skills: the lexical skills assessing part of the parent report method LEINIKKI and Wechsler Preschool and Primary Scale of Intelligence 3rd edition subtests Picture Naming and Receptive Vocabulary. The relation between the lexical diversity measures and the validated measures of lexical skills was examined by Spearman’s correlation. It was also researched if the measures of lexical diversity had the ability to differentiate the children who scored the least in the validated lexical measures. Linear regression models were used in examining whether the measures of lexical diversity predicted the lexical skills assessed by validated methods when the age of the children was considered as a background variable. The token values neither correlated with the children’s lexical skills nor differentiated the least scoring quarter in the validated lexical measures. The type and TTR values correlated significantly (p < 0,05) with two out of three validated measures of lexical skills but were not statistically significant predictors of the children’s lexical skills. The type value did not differentiate the least scoring quarter in the validated lexical measures whereas the TTR value did in two out of three validated lexical measures. The relation of the MATTR value to a child’s lexical skills was the most significant: the MATTR value correlated highly significantly (p < 0,001) with all validated measures of lexical skills, differentiated the least scoring quarter in all three validated lexical measures and was a highly significant (p < 0,001) predictor of lexical skills in all regression models. This study supports the use of the MATTR value instead of the more traditional measures of lexical diversity when assessing the lexical diversity of preschool children. The MATTR value can be used especially in supplementing the view of a child’s lexical skills acquired by other measures. However, more research on the topic is needed.