Look at alumni-earned PhDs in STEM fields on a per capita basis:
http://www.nsf.gov/statistics/infbrief/nsf13323/
http://www.reed.edu/ir/phd.html
About half the top 10 colleges by this measure, and a little more than half the top 50, are LACs.
Sometimes you’ll see push-back against these numbers on CC (suggestions, for example, that LAC alumni may be gravitating to weak programs, or that these numbers only reflect a greater inclination to pursue academic/research careers among the kinds of students who choose to attend LACs). There may or may not be a solid basis for these suggestions; I haven’t seen good evidence either way. To me, it stands to reason that LACs will tend to do a better job than large research universities at motivating and preparing undergrads for graduate-level research. At a LAC, classes are smaller, they are almost always taught by professors (not TAs), and so there can be a greater volume of professor-graded papers and essay/long-answer tests.
The down-side to LACs is that they generally can’t offer as many courses and majors as a big research university can. Also, a LAC may not as easily be able to attract superstar scholars (relatively few Nobel laureates, for example, are affiliated with LACs). In my opinion, the academic advantages of LACs tend to outweigh these disadvantages for most good students in the arts and sciences (unless, perhaps, you are comparing them to a few tip-top research universities.)