MATH3851 Experimental Design and Categorical Data

MATH3851 is a Mathematics Level III course. See the course overview below.

Units of credit: 6

Prerequisites: MATH2801 or MATH2901 and MATH2831 or MATH2931

Exclusions: MATH2810, MATH2910, MATH3830, MATH3930

Cycle of offering: Yearly in Semester 2

Graduate attributes: The course will enhance your research, inquiry and analytical thinking abilities.

More information: This recent course handout (pdf) contains information about course objectives, assessment, course materials and the syllabus.

The Online Handbook entry contains up-to-date timetabling information.

If you are currently enrolled in MATH3851, you can log into UNSW Moodle for this course.

Course Overview

MATH3851 is divided into two parts:

Experimental Design and Categorical Data Analysis. In Experimental Design, you will learn why is it important do design your experiment before starting the data collection. You will learn about principles that will allow you to extract maximum amount of information for a given sample size from available sources. The statistical experiment that needs to be designed, is usually about collecting data to see how certain set of input variables (factors) influence another set of output variables. You will study how to set optimally your factorial designs and how to analyse the information from them using solid statistical reasoning. You will also learn about the importance of randomisation in design and how to analyse randomised designs. These techniques have wide applications in sciences and engineering.

In Categorical Data Analysis, you will learn about specific statistical tools and techniques that are specifically tailored towards analysing discrete valued data. At the beginning, you will learn how to answer simple questions about the presence or absence of association between categorical variables using cross-tabulated data. Later in the course, you will also learn how to model the association between the categorical variables by using techniques such as Logistic and Poisson regression and Log-linear models for contingency tables. A variety of applications in medicine, public health, economics and engineering will be used to illustrate the usefulness of these methods.

Modern Statistical packages will be used to illustrate both the computational and data visualisation parts of the course.