For the past few years I have been responsible for 1.3 undergrad courses in the spring semester. The class I teach myself has about 25 students, whereas the course I only responsible for part of has closer to 65. I have taught them entirely as lectures in the past and focused on honing the material to what I think are the central concepts. Overall, it has been decently effective, but I remain frustrated that a decent proportion of the class still can't wrap their heads around some key information by the time the final rolls around.
Although it would be convenient to blame this on the students not studying enough... back in my day... uphill both ways... everything cost a nickel... etc., the situation is not all that different from the "that reviewer didn't understand!" refrain of the grant game. The solution is the same: change your approach to clarify so that they can't miss the point. Last year I restructured the lectures and tried to build in some reinforcement of the topics I wanted to sink in, with only a minor improvement in retention.
So, I am musing about ways to break up the lecture and get the students involved. For the bigger class I've talked about incorporating clickers before and gotten some good advice in that comment thread about dealing with the smaller class. In particular, Alyssa was a big proponent of the "think, pair, share" approach and several others supported that idea.
In playing with this concept, one of the major issues I run into is that what I teach does not really have "problem sets" that are ready made for this strategy. The same applies for the online homework that Arlenna suggested. I think my solution will have to involve using exam-like questions that take the concepts just discussed and makes students tie those back to previous material.
Of course, the other key piece of this is avoiding a full overhaul of my lectures. Whereas I'm happy to tweak and update, anyone who has built a course from the ground up knows how time consuming that process is. This spring will be the last semester before my tenure packet goes in and I'll be dedicating my time to getting papers out and proposals written, not making sure I push my teaching evals from good to great. In a cost benefit analysis, it unfortunately doesn't make sense.
And so it goes, doing the best you can with the time you can spend.