Leader: Associate Professor P R Rayment and J Hille
Not offered in 2006.
Synopsis: The unit content is organised in two sections: the first covers the design and analysis of sample surveys, thereby balancing the focus on experimental design and analysis in unit STA2216; the second section focuses on robust and nonparametric methods, which are valuable when data do not conform to classical parametric procedure assumptions.
Objectives: On completion of this unit students will be able to: assess the requirements of a proposed survey and develop the appropriate questionnaire and/or data collection techniques to meet these requirements; design an efficient sample survey using the relevant simple random, stratified random, cluster or wildlife sampling techniques; obtain point estimates for all required parameters and establish appropriate error bounds for these estimates; apply appropriate robust and nonparametric procedures to a range of problems including location comparisons and regression fits where traditional parametric methods are inappropriate; utilise robust estimators of location and dispersion including trimmed and Winsorised estimates of location and estimates of linear regression coefficients; utilise a statistical package to apply the statistical techniques covered in the unit.
Assessment: Assignments (two): 30% + Examination (three hours): 70%
Contact Hours: Three hours of lectures and one 1-hour tutorial/computing workshop per week
Prerequisites: MTH1210 or STA1010
Prohibitions: MAT2770, MAT2790, MAT3242, MAS3212, MAS3322, MAT3770, MAT3790, MTH2257