wolfwings - 12:03 am - Actually, no, it's not.
The basic idea is that autistics lack some form of 'flattener' to their random choices, giving them a simple 'classical' gaussian distribution to their actions. Meaning that more extreme 'outlier' actions approach absolute 0 the more extreme they are, while normal people have a 'minimum chance' of any random choice.
It's... rather akin to taking the numbers from a more classical study, and examining the statistical curve of the entire number field instead of examining the numbers directly. Most studies worry only about the numbers, and if anything can be read from them directly, like answering (A) to one question means you're autistic.