Statistics 6220 Spring 2006 Syllabus

Text: Introductory Statistics for Researchers, under contract with Duxbury Press. Binder copies on sale in Statistics Department office (room 204) for $40, cash only.

Optional: Student version of SAS-JMP for Windows. Available in bookstore. If you prefer a different statistics package such as SAS, SPSS, Minitab, etc., that's fine. If you don't have any previous experience using statistics packages, I recommend SAS-JMP as inexpensive and easy to use.

Six months of SAS-JMP for $29.95:Student License

Grading:

Homework problems (30%)

Exam I (20%)

Exam II (20%)

Final Exam (30%)

Exams and Final are open book, open notes. I aim for testing fundamental understanding of important concepts, rather than memorization and computational detail. Explaining and interpreting results will be emphasized as well as calculations. The final exam is comprehensive.

Some homework problems will require a statistical package. The statistics computer lab in room 307 has SAS and SAS-JMP; it is open every evening and staffed with a graduate student. Demonstrations in class will use SAS-JMP.

Academic Honesty: UGA requires an explicit statement of policy concerning cheating and plagarizing.

Philosophy: Everyone's goal should be the same: to foster understanding of good statistics practice and procedures. Discussion of homework problems and group studying of class material can be very helpful. Explaining concepts to others is a very good way of checking and solidifying your own grasp of the concepts. There is no competition between students. I don't grade on a curve, and I don't compare students' abilities in any way that might be contradictory to the class goal.

Policy: Working together on homework problems is allowed and encouraged, but the write-up of the assignments must be done individually. Your exams must be all your own work.

Course Description:

The emphasis is on applications. Examples are taken from newspapers, magazines, and professional journals, and many of the problems are based on real data or real situations. Interpretation, study design, biases, possible confounding, sources of variation, and other aspects of statistical thinking are the focus throughout the semester. The idea is that statistics is never done in a vacuum, but with the purpose of making decisions or conclusions, or modeling or describing relationships. Every time a new formula or derivation is presented, we connect the concept to a variety of situations in the real world. The focus is on interpretation of results in the context of the problem as well as methods.

Topics:

Statistical Studies

Aggregation Bias

Experiments vs. Observational Studies

Confounding Factors

Variables and Distributions

Hypothesis Testing and Statistical Significance

Quick Review of 6210

Chi-square test of independence

T-tests

Sources of Variation

Analysis of Variance

One-way ANOVA

Data Snooping

Multiple Comparisons

Randomized Block Design

Factorial Experiments

Simple Linear Regression

The Principle of Least Squares

The Regression Model

Regression Effect Bias

Residual Analyses

Multiple Regression

Two-line Regression Models

ANCOVA Models

Two-way ANOVA

Two Continuous Predictors

Confounding Factors in Multiple Regression

Residual Analyses

Quadratic Regression and Extensions

Weighted Regression

Logistic Regression

Modeling the Probability of an Event

Odds and Odds Ratios

Multiple Logistic Regression