|Year of offer||2019|
|Subject level||Graduate coursework|
|Fees||Subject EFTSL, Level, Discipline & Census Date|
Extensions of the multiple regression model are examined. Topics include non-linear least squares, maximum likelihood estimation and related testing procedures, generalised least squares, heteroskedasticity, autocorrelation and models with stochastic regressors. Limited dependent variable and panel data models and issues involving time-series data are introduced. Theoretical concepts are illustrated by applied examples.
Intended learning outcomes
On successful completion of this subject students should be able to:
- Explain various problems of interpretation of causal estimates in regression and how the problems may be addressed using instrumental variables.
- Apply least squares methods to estimation and inference for linear regression models with panel data.
- Apply least squares to estimation, inference and interpretation for single and multiple equation models for stationary and non-stationary time series data.
- Derive, simulate and interpret statistical properties of least squares estimators in these settings.
High level of development: written communication; statistical reasoning; application of theory to practice; interpretation and analysis; critical thinking; synthesis of data and other information; evaluation of data and other information.
Moderate level of development: problem solving; use of computer software; receptiveness to alternative ideas.
Some level of development: oral communication; accessing data and other information from a range of sources.