This study examines the decision-making process used for differential diagnosis of a sample of Hispanic children referred for autism spectrum disorders (ASDs). Of the sample of 28 children, 18 were diagnosed with ASDs. Of the 10 children who were not diagnosed with ASDs, 80% were found to have multiple diagnostic labels or comorbidities. Mann-Whitney U analyses determined the differences between the children with the most severe social impairment, children with less severe social impairment and the non-autistic children on several domains commonly used to assess ASDs. These analyses indicated significant differences in some characteristics of the children in the sample. Based on these results, a decision-tree for the diagnosis of children with and without ASDs with comorbid disorders was developed.