<p>Below is the list of the 10 National Universities with the highest percentage of classes with fewer than 20 students:</p>
<p>National University (state) Undergraduate enrollment Percentage of classes with fewer than 20 students U.S. News rank
New School (NY) 6,970 92.1% 128
Immaculata University (PA) 3,223 81.6% 194
Edgewood College (WI) 1,941 81.3% 170
Cardinal Stritch University (WI) 2,984 80.7% RNP*
Columbia University (NY) 5,888 79.7% 4
Harvard University (MA) 6,641 79.5% 1
University of Chicago 5,238 77.8% 5
Nova Southeastern University (FL) 6,174 76.1% RNP
Yale University (CT) 5,310 75.8% 3
Northwestern University (IL) 8,443 75.0% 12</p>
<p>While large classes may often be the norm at National Universities, many of these schools still offer a significant number of smaller classes, according to data reported by colleges to U.S. News in a 2011 survey of undergraduate programs.</p>
<p>The New School in New York offers the highest percentage of classes under 20 students, with 92.1 percent. Among the top 10 schools on this list, four are ranked among the top 10 Best Colleges: top-ranked Harvard University, third-ranked Yale University, fourth-ranked Columbia University, and fifth-ranked University of Chicago.</p>
<p>^Ahahaha that’s pretty funny! I knew there was something fishy about the 19 person limits for sosc and hum and the 18 person limits for language classes</p>
<p>I guarantee Harvard and Columbia and Yale are keeping an eye on class size, too, for the same reason. But in the end, the students at all the schools may be the real winner because if this, so this might be one example of how the rankings have had a good affect on schools overall.</p>
<p>That’s better than suddenly trimming down all classes to half their usual sizes every year at the time when inspectors arrive, as some colleges actually do (won’t name which).</p>
<p>Although the US News rankings may incentive colleges to reduce class size, a reduction from 20 to 19 seems unlikely to significantly increase learning. This seems more like a case of colleges gaming the system by taking advantage of arbitrary cutoffs in the US News ranking. A better statistic than percent of classes under 20 that would avoid this kind of gaming might be to take the geometric mean of class size.</p>
<p>“A better statistic than percent of classes under 20 that would avoid this kind of gaming might be to take the geometric mean of class size.”</p>
<p>Exactly. Of course, the same thing could be said for SAT scores, high school class ranks, etc.</p>
<p>However, US News is aiming to provide measures that can be easily interpreted by parents, teachers, etc. If US News used geometric mean, most people would have no idea what they were talking about, lowering the usefulness of the rankings, exerting a negative effect on magazine sales, etc.</p>
<p>I’m not sure that the geometric mean of SAT scores or class ranks are necessarily very meaningful. In particular, the geometric mean of class rank would be largely determined by the number of students with low class ranks. At least for the SAT metric there’s no arbitrary cutoff like 20 students that can be easily gamed.</p>
<p>Percent of class under 20 is a component of the main US News ranking.</p>
<p>“I’m not sure that the geometric mean of SAT scores or class ranks are necessarily very meaningful. In particular, the geometric mean of class rank would be largely determined by the number of students with low class ranks. At least for the SAT metric there’s no arbitrary cutoff like 20 students that can be easily gamed.”</p>
<p>Of course there is. For SATs, only the 25th and 75th percentiles are calculated. That means that any school can admit a very unqualified bottom 1/4 of its class and have this not noticed by anyone. In a similar way, universities could game the “classes with over 50 students” metric by condensing any classes with over 50 students into a single huge class with over 500 students.</p>
<p>“Of course there is. For SATs, only the 25th and 75th percentiles are calculated. That means that any school can admit a very unqualified bottom 1/4 of its class and have this not noticed by anyone. In a similar way, universities could game the “classes with over 50 students” metric by condensing any classes with over 50 students into a single huge class with over 500 students.”</p>
<p>I guess the above possible although colleges don’t have the same amount of control over applicant’s SAT scores as they over class sizes. If the above hypothesis is correct there would likely be large discontinuities in acceptance rate by SAT scores around the 25th and 75th percentiles. Unfortunately, I don’t think data on that is publicly available. In any case this isn’t nearly as easy as switching all size 20 class to a size 19 class which doesn’t significantly affect the college yet produces a substantial increase in ranking.</p>
<p>You know traditionally UChicago really didn’t have an issue with big class size since it was so much smaller than the other top schools (undergraduate). But these days the college is the same size as Princeton and Yale though smaller than Harvard, Stanford, Duke, Penn, etc., as you know.</p>
<p>Some posters above are cynical about UChicago “manipulating” class size for ranking points in USNWR. While that may well be a factor (and what strategies do other top ranked schools employ?), it is also reasonable to view it, and accept it happily, as an ancillary benefit to keeping CORE sections at a not-to-exceed size conducive to vibrant discussions and the Socratic method so helpful in creating the value the CORE classes deliver at a student’s intro point to a UChicago education.</p>
<p>The number needs to be set arbitrarily for this purpose, or any individual CORE class would be at risk of ballooning to a size that would defeat the purpose. It could have been 20, maybe as high as 22, but why not 19, smaller being better. But smaller is more expensive to deliver, so might as well optimize around the USNWR ranking point to achieve the balance between cost to the university and pedagogic benefit.</p>
<p>I don’t fault for Chicago for gaming the system here. Limiting classes to 19 instead of 20 or 22 costs the school very little and helps the US News ranking. However, I think US News criteria (% under 20 and % over 50) for measuring class size is stupid and the geometric mean would be a better metric. Although most readers might not know the technical definition of geometric mean, most people probably have a rough idea of what average means.</p>
<p>EDIT: As to phuriku’s earlier point about taking the geometric mean of SAT scores, I don’t think that is a good metric as subtracting 200 points from each section of the SAT should make no substantive difference in metrics involving SAT scores but would severely mess up the geometric mean [ex. if the SAT sections were scored on a 0-600 scale instead, a single 0 would set the geometric mean for that section to 0].</p>
<p>As someone who just failed to add in 4 different classes with a capped size of 50, I can say the prestige of having small class sizes doesn’t do anything if one can’t get into those classes because of it. :[</p>
Keep 'em coming!! :)</p>