Abstract
The problem of group performance differentials in learning to read was addressed in this exploratory study from an interdisciplinary perspective. A causal model of first grade reading achievement was developed and tested to answer questions about relative effects of biological and socialization factors on cognitive, linguistic, and affective performances and their relationships to beginning reading achievement. The stratified, random sample consisted of 234 children representing three cultural/linguistic groups (majority-white, black, and Hispanic) from 12 schools in a large urban/suburban school system.The model was estimated with LISREL and was found to have a good fit to the sample data. It accounted for 45% of the variance of reading achievement. Unexpected gender effects, independent of cultural/linguistic group and SES, were found. These were positive for girls on metacognitive understandings of reading and for boys on vocabulary. Cultural/linguistic group membership had direct and relatively large effects on cognitive and linguistic abilities. SES directly affected expressive language only. Vocabulary had no direct effects on reading achievement. Degree of match of children's language use with standard English syntax mediated all social and cognitive effects on reading and directly affected reading achievement and children's expectations for learning to read in school. Cognitive style (reflectivity-impulsivity) was directly affected by cultural/linguistic group membership with negative effects for the black group. Cognitive style directly affected children's expectations for reading only and this variable mediated the effects of cognitive style on reading achievement.Further multivariate research is needed to investigate the following as they relate to learning to read in school: (a) the possibility of early gender differences in learning; (b) the relative importance of language use by children and teachers; (c) components of early home/cultural influences; (d) the relative importance of vocabulary and syntactical factors; and (e) the relative importance of social and cognitive factors as related to cognitive style and metacognitive abilities. It was suggested that future causal learning models be estimated separately and compared for groups classified by SES, cultural background, language use, and gender.