What courses do you need before applying for Stats Ph.D.?

[svent]

Just curious, how many actual stats classes do you need in general before applying for Stats programs? I know pretty much any program's website is wishy-washy on these details, but are basic classes on probability and stats sufficient, or do you need more advanced classes like stochastic processes, regression analysis, etc.? Obviously actual math classes (calculus, linear algebra, real analysis, potentially measure theory or functional analysis) are important.

 

It seems like most applicants don't take the subject GRE, have relatively weak math backgrounds compared to the pure math folks. Are any advanced stats classes even necessary, or do these programs like to recruit people from math/CS/etc.?

[cyberwulf]

A lot of people coming into stats programs (particularly the top ones) are pretty "math-y" and may not have a ton of stats background. For admissions purposes, grad schools need to distinguish between a group of fairly talented students, and performance in "hard" math classes like real analysis, measure theory, etc. provides one potential separator. Stat classes are less helpful for us because: a] curricula can vary widely across institutions, and b] they tend to be "easier", so that a larger proportion of the applicant pool gets high grades.

Edited June 30, 2015 by cyberwulf

[Ellies]

Basic stats is sufficient. Good grades in advanced math classes are a lot more important than taking many stats classes.

[bayessays]

While I do think math classes are more important, having at least had some exposure to statistics, especially a calculus-based probability/stat course is very helpful to show that you have some idea what you're getting into (and is required by some programs for admission).