The courses you elect to take for your first semester will depend on your background. Three of the classes offered in Fall are covered on the First Year Exam (STA 601 , STA 721 , and STA 711) so you should begin by considering these classes. Students generally choose THREE of the following classes to take their first semester:
- STA 611 Introduction to Statistical Methods - This class is taught from the first half of Statistical Inference by Casella and Berger. This class covers topics such as basic probability theory, transformations, expectations, common families of distributions, multivariate random variables, and sampling distributions. If you have little or no mathematical statistics background then be sure to take this class. Most first-years will probably find this class unnecessary, but if you feel like you need some review, you may want to consider auditing.
- STA 601 Modern Statistical Data Analysis - This class introduces students to classical and Bayesian statistical thought. This class also introduces students to basic statistical analysis (such as exploratory and graphical analysis) using R. This class will also touch on LaTex. Some first-years will not need this class.
- STA 721 Linear Models - A thorough introduction to linear regression, with a focus on Bayesian modeling and computation. Model selection, Bayesian model averaging, and other topics. Most (not all) first-years decide to take this class.
- STA 711 Probability and Measure Theory - STA 205 is an introductory class on probability from a measure theoretic viewpoint. Topics include rigorous treatment of Lebesgue measure and integration, convergence of random variables, Lebesgue's dominated convergence theorem, etc. Almost all first-years take this class.
- MATH 241 Real Analysis - The topics covered in Math 241 are also covered in STA 711 (205), so many students take one or the other. If you really like probability (mathematical theory) and want to work mostly in statistical theory, then taking MATH 241 at some point is suggested.
- MATH 216 Applied Stochastic Processes - Discusses Markov chains, martingales, Brownian motion, etc.
- STA 841 Generalized Linear Models - Covers topics such as log-linear models, Poisson regression, probit models, logit models, models for longitudinal, clustered, and multivariate data, latent factor models, stochastic search algorithms, etc. Generally for students who are familiar with Bayesian methods such as Gibbs sampling, Metropolis-Hastings, and have had an advanced course in linear models.
A simple suggestion would be to begin with STA 601, 721 and 711. STA 841 is a permission number course so you'll need to talk to your advisor or the instructor to determine if you are qualified. A "typical" course of study can be found on the department website stat.duke.edu/phd under the links on the left side.