PZ Meyers points to a study showing that most scientists do not perform activities for outreach to non-scientists, and that only a very small fraction (about 3%) do so very actively --- where "very actively" means "six or more times per year", which doesn't even seem that active to me
Well, in fact, this is a subject near and dear to my heart. The reason behind the lack of outreach is obvious: there are almost no institutional incentives for professional scientists to work on outreach activities. From grad school, through postdoc work, through faculty hiring, through tenure review, you win funding and professional advancement based almost solely on research output within your discipline. If you want to do outreach activities, that's fine, but you'll see few career benefits from doing so. Therefore, in practice, the people who end up doing non-negligible amounts of outreach tend to be one of the following:
- people who don't land tenured research positions, and choose an alternative career path;
- people in the tiny minority of superprolific superstars who can publish lots of research while also doing lots of outreach on the side; or
- people who have invested the ten to fifteen years (grad school + possible postdocs + six year tenure clock) of tunnel-vision research needed to land a tenured position, and now have the freedom and security to do whatever they want.
Regarding the third category, one would think that there are many tenured professors out there, and there are. So why don't more of them do outreach? Well, after you've spent a decade being driven to do research and little else, habits form and harden. Also, the long pre-tenure career track selects people who care passionately about research, and therefore prefer to spend their time doing research instead of other activities. In practice, therefore, few tenured professors dedicate much time to outreach. By that time, the die has been cast --- although some tenured professors do get fired up and fighting. It helps if somebody pokes them with a sharp stick (think Ed Felten vs. the RIAA, or biologists who take down creationists).
Now, in the grand scheme of things, this system makes some sense: the American scientific/economic/military juggernaut gets incredible value from the specialization of labor wherein its research universities' primary role is to generate basic scientific research. Contra the popular perception of academia as an inconsequential playground for the effete, American universities are highly optimized for this mission, and accomplish it more effectively, and on a larger scale, than any other institution in human history.
By contrast, the benefits of having a bunch of scientists running around proselytizing laypeople are less obvious (although if you're reading this, then you probably agree that those benefits are substantial). From this point of view, having scientists spend their labor on outreach is actually a form of inefficiency that should be weeded out. And, if we view the arena of competition at the level not of individual scientists but of university departments, all incentives point towards improving efficiency in this respect. A department gets funding, develops a reputation, and attracts top grad students and faculty candidates primarily by having a record of strong research.
So, whatever the long-run costs of isolating science from public life, the short-run features of the competitive landscape drive scientists, and the institutions that employ them, away from outreach. Fortunately, we are not bound by the gradient descent algorithm of natural evolution. We have brains, and we can think further ahead than one step, or even change the landscape. I think that there's a growing sense, at least within computer science, that if we do not school the public sphere, the public sphere will school us. So I'm optimistic that my discipline, at least, will change course, though the change will come more slowly than I'd prefer.