Sure. Sometimes. And sometimes (as in my own case on this subject), the perception came first. This is a complex phenomenon, maybe several phenomena under a single category, possibly with both genetic predispositions and environmental triggers at work. Plus, sample groups that keep moving geographically and are hard to define rigorously.
But dismissing personal observation is not scientific. It's illogical to say that because publicity can make a phenomenon appear more frequent, whenever a phenomenon appears frequent, it must be as a result of publicity. Gut impressions are actually the source of most hypotheses. The scientific method comes in when you test them. A lot hinges on who does the testing, and how well-designed and honest the tests are.
Clusters of cases is one reason. The clustering may be random, or it may be related to an external factor. There could, for example, be a school catering to a specific population in that area.
Publicity plays a huge role in the perception of incidence. The (now withdrawn) report on a link between autism and vaccines was published and widely publicized. People who never had a reason to notice autistic people began to notice them, and automatically accepted the premise that autism has increased.
It is actually not scientific to elevate personal observation to the status of measured data. As a scientist, if I notice that there seems to be more or less of something than usual, my first question is, is there a real increase or decrease in what I have observed? In many cases, it turns out that it is *not* real. If it does turn out to be real, then I have to proceed to the next question: why? In the case of autism, we have had such a huge broadening of the definition of autism that the question of whether or not the increase in incidence is real has not been scientifically answered.
The true incidence of autism spectrum disorders is likely to be within the range of 30–60 cases per 10 000, a huge increase over the original estimate 40 years ago of 4 per 10 000. The increase is largely a consequence of improved ascertainment and a considerable broadening of the diagnostic concept. And, as this recent article discusses, there has been a lot of progress made towards understanding the genetic basis of autism.