Many selecting among elite private colleges and universities value small class sizes.
Reported percentage of small classes (fewer than 20 students excluding labs & other break-out sections) among the Top 15 LACs & Top 15 National Universities (all are private schools) as ranked by US News 2020 edition. All figures are based on the 2018-2019 academic year and are, therefore, pre-Covid.
Percentage of classes with fewer than 20 students ranked:
Columbia University–82%
Claremont McKenna College–81%
University of Chicago–79%
Northwestern University–78%
Haverford College–78%
Williams College–77%
Hamilton College–76%
Wash & Lee–75%
Princeton University–74%
Johns Hopkins University–74%
Swarthmore College–74%
Colby College–74%
Yale University–73%
Harvard University–72%
MIT–71%
University of Pennsylvania–71%
Duke University–71%
Amherst College–71%
Brown University–70%
Pomona College–70%
Bowdoin College–70%
Carleton College–70%
Smith College–70%
Stanford University–69%
CalTech–68%
Wellesley College–67%
Vassar College–67%
Vanderbilt University–66%
Middlebury College–66%
Grinnell College–65%
Dartmouth College–62%
Notre Dame–62%
P.S. CC’s website does not allow lists to have more than one school ranked at the same number even though several schools are tied at 74% & at 70%.
Again, the above percentages do not include any labs, recitation, discussion sections or any other supplementary meetings. Those percentages are listed in the “sub-sections” category on the schools’ CDS. No section may be counted more than once.
This is an odd list. It provides some interesting data points, but the binary categorization of class size (a class of 21 students is not small?), and the analysis only of the top 15 LAC’s and uni’s (when many schools throughout the top 50 and beyond have very similar qualities) limit the usefulness of the information. I’m pretty sure that the 26% of “large” classes at Hopkins are qualitatively quite different from the 26% of “large” classes at Colby. And meanwhile no comparison is made with Bates or Cornell because they’re not top 15. Meh.
FWIW, I suspect the reason Claremont McKenna ranks so high is that all of its intro science classes are taught through Keck Science (shared by CMC, Pitzer, and Scripps) and thus aren’t counted in an analysis of classes that are coded as CMC-taught, even though many are taught by CMC profs. The largest 5C’s class my daughter took was an introductory Keck Science class (although it still was not a giant, anonymous lecture course by any stretch).
Limiting this to the Top 15 (actually top 16 schools in each category due to ties) LACs & National Universities was to be able to compare only the most elite private schools. So, yes, this just a list comparing 32 elite private National Universities & LACs. The Top 15 ranked National Universities are all private schools.
(If this list was limited to the top 25 or top 50 or top 100 in each category there would be dissatisfied readers.)
Choosing to designate classes with fewer than 20 students was not a choice made by me. It is a standard part of each school’s CDS (Common Data Set).
Please note that the CollegeBoard’s statement above references “publishers”, not “Publisher’s”.
Great idea & astute observation as this limited list was intended only to compare elite private schools.
I agree that public National Universities merit separate consideration. However, public schools are not as accessible to all on as equal a basis as are private schools, therefore, listing of public schools has more limited usefulness for readers seeking small classes at the most highly selective schools.
Understood. I’m just not sure that the fine gradations compared here are meaningful, especially when comparing across school types. A person could look at this and conclude that they would have fewer “large class” experiences at MIT than at Carleton, which would be a comically inaccurate impression.
I think most private elite schools will have plenty of small classes, not sure minor gradations are important. But I have a kid interested in looking at state schools but we are both put off by thought of giant classes (at least beyond intro level), so would love to see comparisons as to whether any state schools do have small classes …
Public National Universities ranked in the Top 50 National Universities reported percentage of classes with fewer than 20 students:
UCLA = 51%
UC-Berkeley = 53%
Michigan = 57%
Virginia = 55%
UNC = 38%
UC-Santa Barbara = 50%
University of Florida = 50%
UC-Irvine = 53%
UC-San Diego = 43%
UC-Davis = 36%
College of Wm. & Mary = 47%
Illinois = 38%
Texas = 38%
Georgia = 45%
FSU (Florida State University) = 57%.
Many large public universities have honors colleges or programs which offer small classes for many subjects including introductory courses. But, public schools are not the focus of this thread due to variations in tuition, limitations of amount of non-residents, and difficulty of sorting out honors colleges & honors programs, in addition to domination by University of California schools.
While the list is interesting, it is not a good representation of how likely a particular student is to see small class sizes. The small class sizes are usually dominated by less popular majors and certain types of upper level classes, such as seminars. This is particularly true with larger private colleges. Smaller LACs are generally more likely to have small classes in more popular majors and lower level classes.
For example, the list above says 72% of Harvard classes have <20 students. The Harvard Q guide list the following mean enrollment by subject. Student’s view is the mean enrollment seen by students, so popular lectures are weighted more heavily than unpopular small seminars. If there was a survey of students of all students taking econ classes that asked how many students are enrolled in your econ class, it would be closer to this >200 value since a large portion of students are taking huge intro to econ type classes, which may have 600+ students. The Dean’s view instead looks at the mean class size across all offered classes, regardless of how likely students are to take the classes. The numbers below are a few years old. I expect CS classes would have considerably larger mean enrollment today.
Harvard Mean Class Enrollment by Subject – Student’s View
Economics – 230 students
Computer Science – 210 students
General – 170 students
Chemistry – 140 students
…
Philosophy – 30 students
Expository Writing – 20 students
Visual & Environmental Studies – 20 students
Freshman Seminar – 10 students
Slavic Languages – 10 students
Harvard Mean Class Enrollment by Subject – Dean’s View
General – 73 students
Computer Science – 50 students
Chemistry – 46 students
Economics – 38 students
…
Religious Studies – 17 students
Expository Writing – 15 students
Visual & Environmental Studies – 12 students
Freshman Seminar – 11 students
Slavic Languages – 8 students
I usually add a space between “1” and “.”, so the software does not recognize it as a list. It would be nice if the CC software had away to turn off automated formatting.
@Data10: This may be why the CDS (Common Data Sets) also list the percentage of classes with 50 or more students.
Northwestern University & the University of Chicago have the most impressive class sizes when considering both the percentage of classes under 20 students as well as the percentage of classes with 50 or more students if one values small classes.
More useful for a student concerned about class sizes would be to check each college for the class sizes of the classes that the student is interested in.
Also, note that some colleges try to game class size metrics by capping many class sizes at 19, 29, 39, 49.
Agree that students should examine specific majors at particular colleges if small class size is a priority.
Do you know any colleges that game the system ? (Legitimate question; not trying to challenge your statement.) Also, not sure if “gaming the system” is harmful or misleading in this situation as all schools are free to limit class sizes.
You can also check Northeastern’s class schedule at Banner . For example, most English writing 1110 and 1111 courses are capped at exactly 19 students, rather than 20 students. The ranking bump for having 20 sections of 19 students versus 19 sections of 20 students is likely the reason why 19 is a much more common enrollment cap than 20. Also, if there are 48 students for a given class, having one section of 19 and one section of 29 looks better for class size metrics than two sections of 24.
Northeastern University’s tactics have been heavily discussed in the media due to an unusually swift rise in the rankings. Boston’s Northeastern University may be an anomaly.
Some may find this non-conclusive study to be of interest:
Gaming may interfere with an optimal allocation of resources. For example, two sections of the same course may be optimally taught under available resources at 20 students each. With a gaming scenario, however, an artificial aspect appears, under which these classes might be adjusted to 19 and 21 students so that at least one of the classes falls below the externally influenced threshold.
Also, at a larger university, I bet major matters when looking at class size. Classics classes might be consistently smaller than a bigger major like political science or econ.
For class size comparison to be meaningful, it needs to be by major or department. A CS class is going to be relatively large anywhere (if a lower division CS class at a college has only a dozen students, it’s probably not very good). A college with a significant portion of its students in CS, or interested in CS, is likely to impact its average class size.
︎1.