The gist of this article is that it might, particularly in certain types of courses and circumstances, but we're making a very big case about changes in education (that smaller classes are automatically better) on the basis of very little real evidence. Are there not ways to teach large sections just as effectively for certain outcomes (content knowledge, for instance) as for small sections? Most likely, but not much real study has been put into the topic to know for sure.
http://www.insidehighered.com/views/2007/12/06/barwick