Tuesday, September 05, 2006

Opportunistic Definitional Change

On the Freakonomics blog, Steven Levitt discusses a recent report that shows a correlation between autism and older fathers, and he wonders whether economics has much to say about the rise in autism cases. In the comments, “Sarah” nails it:
I have a friend whose son was clearly “not right”. After lots of agonizing, and lots of testing, he was diagnosed with Asperger’s-type autism.

Now, it’s well-known that diagnosis of anything can often bring tremendous relief to someone who has been suffering with a nameless problem, but in this case, the relief was something different: it was relief that he QUALIFIED for state-funded programs to help autistic kids. That meant physiotherapy, speech therapy, a special preschool, swimming lessons, and more.

I don’t deny that this boy will really benefit from these programs, and I’m glad he’s getting them. And, obviously, his parents don’t want him to be autistic. But… there must be many cases (autism, and other syndromes or diseases), where an upswing in prevalence reflects the pushing of borderline cases into diagnosis, in order to qualify people for benefits. If that’s not economics, I don’t know what is!
At least one other commenter took umbrage at the implication (by commenters less careful than Sarah) that autism and Asperger’s syndrome are just made-up. But you don’t have to deny the reality of the condition to recognize the economic incentives at work.

As experts in the field recognize, autism and Asperger’s are spectrum disorders, which means there’s no sharp dividing line between those who have them and those who don’t, just as there’s no specific height that qualifies a man as tall, and no specific number of grains of sand that defines a heap. Yet there really are tall people, there really are heaps, and there really are people with autism or Asperger’s. (As my co-blogger Tom likes to say, the existence of dawn does not abolish the difference between night and day!)

In our paper on slippery slopes, Mario Rizzo and I argue that slippery slope effects are especially likely in the presence of gradients like these. When law or policy grants privileges to, or imposes duties on, people who fall into a vaguely defined class, new cases will continually arise that push the class-defining boundary one way or the other. For instance, if retarded people are exempt from the death penalty, the dividing line between retarded and not retarded becomes very important – and some criminal defendants will have a rather strong interest in being classified as retarded.

In the context of spectrum properties, the arbitrariness of any strict definition will always produce hard-to-justify distinctions – such as executing someone with a 70 IQ while sparing someone with a 69 IQ, or granting funds to a middle-class family with a child who is borderline-autistic by current standards while refusing funds to a poorer family with a just barely more functional child. The natural tendency will be to shift the boundary to produce consistency of treatment.

In principle, the class-defining boundary can slide in either direction, becoming more or less inclusive. But when public policy rewards membership in a class, the incentives naturally point toward more inclusive definitions, as seems to be the case here. The direction of definitional change is most predictable when the benefits of expansion are concentrated on a small group (such as parents of children who are borderline autistic by present standards and the professionals who treat them), while the costs are spread thinly over a large group (such as taxpayers).

I won’t be surprised if future econometric studies reveal that the rate of autism and Asperger’s diagnoses has grown fastest in those states that provide public funding for children so diagnosed, and possibly grew earliest in those states that adopted the funding earliest. Levitt is the acknowledged expert at this kind of analysis.

The potential for a slippery slope does not necessarily argue against having programs to fund treatment of autistic/Asperger’s children, of course. But it should at least make us wary of their growth. We should not expect any definition to produce a clean separation between the cases-that-should-be-funded and cases-that-should not, and that vagueness creates a built-in mechanism to encourage growth of the program.


Anonymous said...

I think ADD is a made-up ailment, fostered mainly by drug manufacturers and caretakers (teachers, parents, etc.) who don't want the hassle of dealing with the fact that kids are kids, energetic and curious.

Anonymous said...

When I was in jr. high school, I was "dragged out" of history class to be tested by a speech therapist. She claimed I had a speech impediment that required her intervention. I went along with the charade because it meant missing a dull class, and spending time with a pretty lady who gave me her undivided attention. I don't think she changed the way I talk any. I'm still not very knowlegeable in history; and just what's so terribly wrong with talking like Daffy Duck or Porky the pig?!

Jadagul said...

I wonder (although I might know if I read the paper, which I'm too lazy to do) if 'older-father' proxies for someone more able to game the system and get a diagnosis that helps his kid. And anonymous the first, as far as I can tell, there are people with a real problem, and an inability to focus for any length of time. The ritalin or whatever enables them to perform dramatically better than they would otherwise. That said, I suspect that this set of people is a small fraction of the total set of people with diagnosed ADD.

Glen Whitman said...

Anon -- I agree with Jadagul. ADD is a real thing, but its boundaries have been stretched in exactly the manner described in the post. ADD is susceptible to that kind of expansion because ADD, like autism, is a spectrum disorder. One end of the spectrum is total dysfunction; the other end is what most of us would consider normal childhood behavior.