Oh? Early 90s EE curricula my experience was more like 50 students came to the first "ohms law" type class and by the time we were talking about rectifier topologies (half-wave, full-wave, bridge, etc) we were down to like 10. I honestly don't know what happened to the other 40 or so.
Even first semester calculus wasn't that high of an attrition rate, maybe only 50%. Something I never really understood about the whole process of attrition is most of the dropouts happened during the really easy stuff like basic continuity concept or perhaps definition of basic derivation equation. Then everyone still present when we talked about the derivative of x squared, mostly made it thru diffeqs together although the difficulty ramped up sharply and smoothly over time. Although freshman classes may distribute on some bell curve, or graduating classes may bell curve, I don't think filter classes have a distribution anything like a bell, because then I'd expect people to randomly drop out at a decreasing rate thru the curriculum rather than all at once in the start.
(edited to add there's probably a startup opportunity to "help" attrition rates by introducing prospective EEs to ohms law or similar in other fields. All the ham radio operators (including myself) pretty much slept thru the early classes with straight As so conceptually an early introduction and filter in a non-uni setting should be possible in other fields?)