Here is a list of average GPAs by department

Posted by: The Lion 1 year, 6 months ago


Courtesy of the President's Task Force on Undergraduate Education's 2009 report "An Agenda for the Future," here is a list of mean and median grades assigned by department over the course of three years. WikiCU provides a pretty thorough summary of the report, which includes such proposals as a 10-20% increase of the undergrad population and the construction of a new dorm, but let's focus on the tables on pages 80 and 81.

Here's what the report had to say, on page 72.

Fact 4: Departmental Grading Policies are Very Unequal

A final issue concerning how undergraduate education is conducted at Columbia
concerns the evaluation of students. Table 3 reports the mean and median grade by
department. The striking feature of the table is that the sciences and hard social sciences give out much lower grades than the humanities. The median student in Chemistry, Economics, and Electrical Engineering receives a B+ whereas the median student in Italian, Slavic Languages, or Music receives a grade of A. In other words, in some departments an A- is a low grade while in others it’s a high grade.


In order to separate these two hypotheses, we took data on every course taken by
every student between AY2004 and AY2006 and estimated a linear regression model in
which the dependent variable was the grade that every student received in every
department. The independent variables were a series of indicators for each student (a
student “fixed effect”), the level of the student when they took the course, whether they
were a major in that department, the log of class size, the class level, and a series of
department indicator variables. This model controls for student ability and tells us how
the same student is graded in various departments.

The results from this analysis are presented in Table 4. The results indicate that
hard sciences and hard social sciences are hard; they systematically give the same student lower grades. This strongly supports the hypothesis that the same student gets lower grades in quantitative courses than in non-quantitative courses. The table indicates that the average student in a Chemistry course will be on the margin between a B and B+, but that same student will receive an A- or A in most humanities departments. Indeed, Philosophy is the only humanities department that grades like a science department; all of the others grade much more leniently.

The analysis is sound, and the numbers don't lie. But don't drop all your econ and chem classes and register for five sections of Ear Training just yet.

There are several statistical issues that come up with this data. First of all, the latest survey of courses was taken in 2006, and judging from national reports on grade inflation, the divide between Chemistry and Music is likely to have lessened substantially.

To complicate things even further, professors are known to change grading policies between semesters depending on the aptitude of the class. A class with 40% A's in one year may have 60% the next year, and when you're dealing with lectures with as many as 530 students (in Frontiers of Science), this skews the statistics significantly.

Many would argue that allowing students to quickly judge classes and departments by GPA undermines academic pursuits and discourages intellectual discovery. This is the same mindset that opposes open course evaluations.

However, considering how much future graduate school and employment opportunities depend on GPA, students deserve to have this information available to them. Similar institutions like Dartmouth even post median grades for individual classes (Cornell discontinued the policy in 2011), and the quality of education hasn't noticably declined.

For now, this list of average GPAs is the closest thing we have to a comprehensive account of grading policies across departments. Use it to your advantage.


  • Transcripts don't lie 1 year, 6 months ago
    Points: 5

    Request an official transcript from the registrar. It lists the percent of each of your classes (of 15+ students) that received an A- or higher. It's not as useful as the raw data, or a median, but it can be VERY telling. You can get a free PDF emailed to you in SSOL.

    Link / Reply
    • Tim Weber 1 year, 6 months ago
      Points: -1

      Note that the % of A is true for CC transcripts only.

      Link / Reply
      • Ben G. 1 year, 6 months ago
        Points: 10

        That is incorrect. %A grades appear on GS transcripts as well.

        Link / Reply
        • Tim Weber 7 months, 1 week ago
          Points: 0

          Well GS doesn't really matter does it?

          Link / Reply
  • Not even a stats major 1 year, 6 months ago
    Points: 0

    Without the standard deviation this is information of limited utility--many people do extremely poorly in science and math classes, yet last year 3-4 of the first set of inductees for PBK were physics majors. In other words, there's a large spread between the best and the worst. By contrast, in the humanities and social sciences the best students receive As (A+s are rare) while the worst student might receive a B-.

    Also, what the hell does "Grades Given to Students of Identical Quality" mean?

    Link / Reply
    • Nick 1 year, 6 months ago
      Points: 9

      Also, a lot of people take classes in mixed departments. So for example a physics major who takes a lot of humanities classes might have a higher GPA than one who takes a lot of math classes or may be the other way around if s/he is more mathematically inclined. This especially a common phenomenon among premeds who might declare a certain major that's perceived to be easier to boost GPA while taking the premed requirements. So the engineers who are taking Orgo might be in for a rougher ride than the regular premeds since they have other science classes to worry about (unless, you know, again, they're really good at science, which they might well be.) There are just a lot of different variables.

      I know from personal experience, being in Applied Physics, I've personally found the curve in EE classes to actually be more forgiving which is not what these stats would have you believe. But it's related to the fact that different types of people declare EE compared to Applied Physics. Also Applied Physics gets mixed with Applied Math and Materials Science, which have such totally different requirements. Point is, in departments that curve heavily, it depends a lot more on who your classmates happen to be.

      More generally, there are a shit ton of variables which make these stats difficult to read, and GPA, while important, is kind of arbitrary, and kind of useless when comparing individuals, and above all a stupid reason to declare a certain major.

      Link / Reply
  • Anonymous 1 year, 6 months ago
    Points: 3

    The EE department doesn't curve higher than other departments; only intro to EE is particularly lenient. Maybe circuit analysis as well. The average grade in signals and systems, a second year course, was a C+ two years ago.

    Link / Reply
    • Nick 1 year, 6 months ago
      Points: -1

      I agree with that, and the statistics prove that as well. I meant that the competition is different in different departments.

      Link / Reply
  • Zach 1 year, 6 months ago
    Points: 0

    To me this suggests that the grading system itself may be flawed. Table 4 particularly seems to present some evidence that this level of grade differentiation is not a good indicator of students' long-term academic or professional achievement potential. Many notable, high-quality graduate and some undergraduate programs have taken steps to thing beyond our current grading system. Considering the broader dynamics of admissions and competition at Columbia, we could be a leader here.

    Link / Reply

New Comment

required (not published)