GAUSSIAN MIXTURE MODELS FOR THE ANALYSIS OF WISC-IV DIMENSIONS: A MULTIVARIATE APPROACH TO IMPROVE THE ASSESSMENT OF INTELLECTUAL FUNCTIONING

The Wechsler Intelligence Scale for Children-IV provides four indexes that analyze the intellectual functioning in specific cognitive fields and a full-scale intelligence quotient (FSIQ) as a measure of the general cognitive ability. However, often the diagnostic process considers the FSIQ score only. This study exploits the Gaussian mixture model (GMM) as a statistical tool to analyze WISC-IV capability to support the diagnostic decision-making process in a multidimensional approach based on the joint evaluation of the four main indexes. The study was conducted on two groups of participants (10 and 12 years old with N=52 and N=47, respectively) with clinical diagnosis. In addition, N=50 observations were randomly generated from the distribution of the Italian reference populations referred to each age group. In both groups, GMM detected two components underlining different behaviors in central tendency, variability, and correlation. Comparison of GMM partitions with a supervised classification shows that group memberships are congruent.