There was a recent conversation on the ASAN discussion list about research priorities and the ASAN's policies about supporting (or not supporting) certain lines of research. While I cannot speak for the ASAN as a whole, I can outline my own opinions on the matter.
It is common practice to divide research into so-called "basic research" and "applied research". Simply put, basic research involves finding out what the heck's going on; applied research figures out how to change and/or improve it. Basic research produces knowledge which informs and guides applied research; applied research produces what we commonly refer to as "technology".
For an example, a study that investigates the way autistics learn would be basic research. A study that investigates how to best teach autistic students would be applied research. For obvious reasons, the latter sort of study benefits considerably from the former -- it's far easier to design a teaching method when you understand how your student learns.
Judging the value of applied research is comparatively simple and straightforward. Its value is a function of that which it creates. Research into ways to make autistic children act like neurotypical children, for instance, is highly valued by those who view this as an important treatment goal and actively opposed by many (if not most) of those who do not.
Attempting to judge the value of basic research, however, is notoriously more tricky. Most people, however, judge the value of a concerted effort into investigating a basic research topic as a function of the value (or percieved/expected value) of the applied research that it is expected to inform.
Because of this, you can tell a lot about a person (or an organization) by the areas of research (basic or applied) that they prioritize. Autism Speaks, for instance, funds a lot of research into the genetics of autism. What does this area promise?
Well, given that gene therapy is (at least for now) a pipe dream and the extent to which we (don't) understand the human genome, the main benefit of knowing which genes cause or contribute to autism would be in the area of testing. Simply put, that knowledge would allow us to develop a genetic test.
The primary "advantage" of a genetic test is that it can be done prenatally. Going by historical data on how such information has been used, this means that it will be used to selectively abort fetuses that will develop (or are judged "at risk of developing") into autistic children.
This says quite a bit about Autism Speaks.
As for my own research priorities? I just want more basic research into the ways autistics think and learn. There's a reason why I almost always enjoy Michelle Dawson's papers.