I think the inverse correlation you've found is easily explained. There are folks who design languages from the perspective of academia and there are folks who design languages from the perspective of practicality, and there is not much overlap in those communities. Perl and PHP are imminently practical languages, but they never would have been put forth by academia, because they are not pure, they aren't interesting in a research sense.
It's difficult to say that with complete conviction. Racket, for example, aims to be very practical. However, most of the research in PL isn't necessarily on entire new language systems as a whole, but maybe a specific feature or category.
Disclaimer: I guess by association I'm affiliated with this group[1].
