John R. Skalski
Q SCI 483
Analysis of linear regression models and introduction to nonlinear models. Model selection using generalized F-tests; residual analysis. Application to categorical, count, binomial, transformed variables. Introduction to matrix formation of regression models and applications. Prerequisite: Q SCI 381; Q SCI 482. Offered: Sp.
Basic topics of ordinary least-squares including model fitting, parameter estimation, confidence interval estimates, residual analyses, hypothesis testing, sample size calculations, and power analyses. Generalized linear models will be presented for non-normally distributed data. Weighted regression, nonlinear regression, and loess smoothing will also be covered using the free statistical software R.
Student learning goals
Perform single and multiple variable regressions
Analyze continuous, count, categorical, and binomial data
Construct and evaluate model fit
Graphically and tabular summarize data
Use free statistical software R
General method of instruction
Two 2-hour lectures per week, along with a 2-hour computer lab using R.
Working knowledge of basic statistical inference, as taught in QSCI 482, essential.
Class assignments and grading
Nine weekly homework assignments will analyze a variety of ecological data sets using the methods introduced in the lectures and presented in the computer labs. One midterm and one in-lab final examination.
Grade based on total numerical score from nine weekly homework assignments (450 points), midterm exam (100 points), and final exam (150 points).