The development of the Minnesota Visual Autism Symptom Scale (MN-VASS)

Abstract

The development and psychometric characteristics of the Minnesota Visual Autism Symptom Scale (MN-VASS) was described. The relationship between the literature surrounding autism symptoms, the diagnostic criteria for autism, and the resulting content of subscales was presented. Item analyses were conducted using item total correlations. All of the item total correlations were above .30, suggesting adequate item functioning. Two internal consistency estimates of reliability were computed for the MN-VASS: a split-half coefficient expressed as a Spearman-Brown corrected correlation and coefficient alpha. For the split-half coefficient, the scale was split between whole subscales so that the traits of autism measured on the scale would be equally divided between the halves. The value was .89 for split half and .90 for coefficient alpha, suggesting satisfactory reliability. Test Retest reliability was reported as a correlation coefficient of .90 for the entire scale (n=22), with subscale correlations ranging from .63 to .93. Inter rater agreement was between 60% and 87%. Convergent validity was investigated between the Childhood Autism Rating Scale and the MN-VASS. A correlation of .89 was reported between the two scales. Teacher/therapist evaluations of the scale suggest that the MN-VASS is a useful and valid measure of the important and teachable behaviors which can be addressed through a program of instruction for children on the autism spectrum.