What is up with the statistics department?

Maybe this is just because I knew I wanted to be a statistics major, but I heard things about the statistics department at Columbia my freshman year that left me somewhat terrified to do so. Three years later and I can say that you should not be afraid—if you do want to major in statistics, that is. And if you do, here’s what you’ll be getting into, and I promise, it is manageable and, honestly, fun!


  • Prerequisites:
    • Mathematics: Calculus I, II, and III; Linear Algebra
      • Your pure mathematics background! Standard exemptions for Calculus I and II (having a sufficient AP score from AP Calculus BC, etc.) apply. 
      • What will you be using most in statistics? Integration and differentiation stay with you (especially partial derivatives and multiple integrals). Linear algebra (matrix multiplication, spectral decomposition, eigenvalues, and so on) is a big deal, especially for elective courses. Luckily, matrices are sexy!
    • Computer science:
      • There are several courses that can fulfill this. I took the one offered by the department, STAT 2102, because I didn’t trust my ability to survive a class actually taught by the Computer Science department. Coding is coding: great when it works, awful when it doesn’t.
  • Core courses in probability and statistics:
    • After taking Calculus-based Introduction to Statistics and Probability Theory, you don’t technically have to take these courses in order, but having a background in probability theory and statistical inference is only going to help you in the more specific courses.
    • Calculus-based Introduction to Statistics and Probability Theory
      • The weeder classes—genuinely. The introductory course, in particular, is way harder than you think it should be, and if you finish it not just thinking that it was an emotional and mental trial, but a class on material you actually don’t like, stop here. Probability theory is, well, theory, and contains the parts of statistics that I think are the most conceptually difficult to understand at first (how many different ways people can sit at a table and whatnot). 
    • Statistical Inference
      • Fun! The things you talked about in your baby statistics courses—hypothesis testing, confidence intervals, maximum likelihood estimation—in more depth.
    • Linear Regression Models
      • Exactly what it says. A mix of application in R and theory, and a pretty straightforward class.
    • Statistical Computing and Introduction to Data Science
      • All the applied statistics (and programming) you can handle. Sometimes taught in the truly cruel format of a two and a half hour Friday course, but you’re not just sitting in lecture the whole time, so it’s more or less bearable.
    • Elementary Stochastic Processes
      • How do you spell stochastic? What does it mean? Why on earth are these things considered elementary? Honestly the hardest class I’ve ever taken, but also, because there’s something wrong with me, one of my favorites, too. Incredibly theory-based, and a lot of talk about drunk people trying to walk home (or into the river).
  • Three approved electives in statistics or, with permission, a cognate field.
    • Lots of spunky options here on various topics: Bayesian statistics; multivariate statistical inference; survival analysis; times series; nonparametric statistics; the list goes on. Don’t just go by the topic—pay attention to who’s teaching it, too, since that will definitely impact how much you enjoy the course.
    • Introduction to Modern Analysis I and II deserve a special shoutout if you’re planning to apply to doctoral programs or other postgraduate study. It’s usually a requirement for those programs, and even if it’s not, you’re going to want to have taken it. A year-long course (hence I and II) in the mathematics department, you technically need to have taken Calculus IV to take it, but just ask your instructor nicely to get around that.

Overview of the classes:

  • One surprising thing about the statistics department is that beyond the most introductory courses, all statistics classes are open to both undergraduates and the department’s MA students, meaning you have a mix of all different ages and schools (especially since Barnard’s statistics department is the same as Columbia’s, or vice versa).
  • This means that since you are indeed technically getting a master’s level education—MA students have a different curriculum and requirements, but still, you’re in classes with them—the classes are hard. Every statistics class has made me fight for my life at some point or another, though some more than others.
  • To that point, though, even though the department definitely has its fair share of why-are-you-a-professor professors, most of the professors I have had give good, cohesive lectures, have clear expectations, and prove to be extremely helpful in office hours. I have also yet to meet a statistics TA I didn’t love. The courses have high standards, but most professors are going to help you get there instead of leaving you high and dry. And no one curves down.
  • Most statistics professors are good professors. All statistics professors at Columbia, good or bad, are also the weirdest people you will ever meet. I can’t even explain it. It goes beyond quirky. None of them are normal. I’ve had other professors who have been Statement Professors, who have left a strong impression, but this is a department-wide thing in statistics. None of them are normal. I mean it.
    • Example: One professor changes outfits between classes so that his clothes are coordinated with his lesson plans. His directory photo is also him and Kim Kardashian. I’ll leave you to find out who he is based on that.
  • Professor recommendations: Ronald Neath; Anne Van Delft; Carsten Chong; Johannes Wiesel; Wayne T Lee; Gabriel Young. You won’t go wrong with them, though you’ll want to write down the things they say for future generations.

Department communications:

I just wrote Professor Neath’s name as a professor recommendation, and I stand by that, but it should be noted that, since he’s the director of undergraduate studies for statistics, it can also be a journey to get him to respond to your email. Once you get a hold of him, he’s straightforward and helpful, but your best bet is to show up to his office hours directly.

Applying for the major: You just have your little SSOL moment!

Last-minute tips:

  • If you don’t like statistics, don’t major in statistics. That’s true for any major, but especially this one. The “weeder courses” (Calculus-based Introduction, Probability Theory), are grueling, both because of the wildly varying quality of the professors and the fact that it is difficult material. 
    • If you do like statistics, however, I promise you, you’re not doomed to hate your department. I love statistics! Maximum likelihood estimation makes me happy! You will love it too. Just do your research on the professors, reach out to them and your TAs, study hard, and enjoy learning from the most delightfully weird professors known to man.
  • If you don’t know how to get an answer to a question, it probably means you need to use the law of total expectation and/or law of total variance.

School of Social Work but also Statistics via Columbia