Arch Neurol. 2008;65(1):155. No New Neurobiology Yet for Autism

First, they are present in individuals in the nonautistic group of the studies they cite. The autistic-related traits of these studies derive from group average comparisons that hide—as can be seen in inspection on the nonautistic control group data—individuals who have the same trait as that identified at the group level with autism. None of the reviewed physical or functional traits indeed identified with autism in the cited studies is, in fact, specific to autism. Moreover, such normal "outliers" argue that from a population perspective, the majority of individuals with a trait identified by group analysis with autism, in fact, paradoxically do not have autism. This is because a small percentage of a large number (the population of nonautistic individuals) can easily be larger in absolute terms than a high percentage of a tiny one (individuals with autism make up only a small faction of the total population).