This unit begins with an introduction to non-parametric testing and estimation, thus complementing normal-based methods like t-tests and one-way ANOVA. Much less needs to be assumed about the population distribution, and the way is opened to make comparisons of different tests for given hypotheses as the type of population distribution is altered. The unit addresses questions such as—When is a t-test the best test? Are there tests which work well for any type of population distribution? The remainder of the unit treats a selection from topics such as Bayesian statistical methods; large sample behaviour for point and interval estimation; decision theory; and Bayes estimation. -- Course Website
Instructor: Assistant Professor John Lau
Prerequisites: STAT2225 Statistical Science