A Second Course in Statistics: Linear Statistical Models

• Activity Number: 310
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Education
• Author(s): David Gurney*+
• Companies: Southeastern Louisiana University
• Keywords: linear statistical models ; second course ; matrix methods

Abstract

This is a proposal for a one semester undergraduate course in statistics based on linear statistical models. Linear statistical models have the form Y=XB+E where Y is an nx1 matrix of observations, X is an nxp matrix of known constants, B is a px1 matrix of unknown parameters, and E is an nx1 matrix of unknown random errors. After presenting examples that fit this form, the solution method would be developed. Then single and multivariable linear models; exponential and logarithmic models; and even polynomial and sinusoidal models would be created. The course would have a linear algebra prerequisite and would use matrix methods to develop the models and explain the underlying theory. Concepts of statistics such as correlation, unbiased estimates and degrees of freedom would be explained in detail as the course progresses.