Approaches to the Prediction of Language Abilities in a Sample of Children Who Have Developmental Delays

Abstract

Prediction of the quality of language was explored using planned comparisons of three approaches, one cognitive, one neurodevelopmental, and one a combination of the two. Subjects were 37 children, ages 5-9 years, whose significant developmental delays included language and speech skills. The cognitive predictors were mental age (MA) and IQ from the Stanford-Binet Intelligence Scale. Neurodevelopmental predictors consisted of fine motor skill quotients (MQs) and dichotic speech processing scores. Chronological age (CA) was also evaluated as a predictor. A composite language ability score constituted the dependent variable. Results of regression analyses showed that CA and MQ, and MA and MQ, were nearly equal in their predictive strengths and were substantial predictors of composite language scores. Larger multiple correlations (low .8 range) were found when combinations of MA, IQ, and MQ or CA, IQ, and MQ were used as predictors. Statistical control over the 4-year age range revealed that approximately equal amounts of prediction of language scores were attributable to CA and a combination of MA, IQ, and MQ. Each of the latter variables contributed important amounts of unique variance to the language score prediction. Dichotic ear scores did not relate to cognitive or language scores and were ineffectual as predictors in regression analysis. Results indicated that children of the type studied have language and speech delays that show substantial relationships to their verbal cognitive abilities and MQs, in addition to their CAs.