Griswold's Conjecture

For the time being, I am going to name the following statement Griswold's Conjecture:

Employment of early intervention for children diagnosed with autism leads on average to poorer outcomes than would have been the case without such intervention.

For the purpose of this conjecture, “early intervention” is defined as any technique administered to autistic children with the intent of ameliorating or removing autistic behaviors and attributes. Examples would include applied behavioral analysis, pharmacology and chelation.

I am uncertain if anyone else has attempted to state a similar hypothesis (and if they have, I will gladly rename the conjecture), but I believe it is important to place this statement as formally as possible within the autism debate arena; for at the present moment, the autism research community is unabashedly assuming its negation. In nearly every research article and in nearly every accompanying press release, autism scientists can be heard repeating the mantra that early intervention greatly improves outcomes for autistic children. The correct reply to such statements should be, “Really? So you have disproven Griswold's Conjecture? Can you show me your results?”

With increased awareness and diagnosis of autism, it has become more and more apparent that in the past most autistic individuals went entirely unrecognized. Therefore, many of these individuals would have been raised in circumstances similar to non-autistic individuals, and would not have been exposed to intervention techniques designed to lessen or remove their autistic characteristics. Although specific outcomes would have varied greatly, statistics make it clear most of these individuals did not end up in institutions or in other similarly poor circumstances. In all likelihood, many of these individuals—perhaps a sizeable majority—somehow made their way into general adult society, more or less indistinguishable from the other members of the population.

Today, circumstances have changed greatly. More and more autistic children are being exposed to early intervention techniques targeted specifically to the amelioration or removal of autistic behaviors and attributes, and there is a growing body of anecdotal evidence suggesting these techniques lead to very poor outcomes (see for instance here, here and here). Without further investigation, it remains unclear whether early intervention does not in fact produce exactly the opposite result from what the autism research community intends, does not produce a very large increase in negative outcomes for autistic individuals.

Therefore, I am placing this challenge before the autism research community: disprove Griswold's Conjecture, or else quit assuming its negation.