Contribution of Selected Variables to the Prediction of Speech Proficiency for Adolescents with Cleft Lip and Palate

Abstract

The purpose of this study was to examine the relative contribution of clinical data obtained during speech examinations at age 4 years and at each subsequent age through 13 years to the prediction of judged speech proficiency at age 14 years for children with unilateral cleft of the lip and palate. Psychological scale values of speech proficiency were obtained for 50 adolescents and used as the dependent variable in each analysis. The 16 independent variables used in the analysis at age 4 years included four nonspeech variables (gender, type and age of primary surgery, and pharyngeal flap surgery) and 12 speech measures obtained from speech examination. In subsequent analyses at ages 5 through 13 years, the rate of change between adjacent age levels for each of the 12 speech measures was also included, for a total of 28 independent variables. The data were analyzed using the MAXR stepwise regression procedure , first for the total group of subjects and then separately for males and females. The results of this investigation indicated that a large percentage of the variance in judged speech proficiency at age 14 years can be accounted for, using clinical data obtained for speech examinations at ages 4 through 13 years. When the regression analyses were performed for the total group of subjects at ages 4 through 13 years, the most efficient set of predictors accounted for 50 to 75 percent of the variance in the dependent variable. The variable of gender was identified as the single most important predictor in nine of the 10 regression analyses, alone accounting for at least 40 percent of the variance in judged speech proficiency. When the regression analyses were performed separately for males and females, approximately 50 to 80 percent of the variance was accounted for in judged speech proficiency for males and 50 to 90 percent for females using one, two, or three of the independent variables. The most efficient set of predictors varied across age levels for both groups. In addition, these predictors differed between males and females.