**Instructor:** Gary C. Ramseyer

**Office:** 451 Degarmo Hall

**Phone:** 309-438-7939

**Email:** gcramsey@ilstu.edu

**Web Site:** http://www.ilstu.edu/~gcramsey

**TEXT:**

Tatsuoka, M.M. and Lohnes, P.R. (1988). Multivariate Analyses: Techniques for
Educational and Psychological Research (2nd ed). New York: Macmillan (Reprinted
by PIP PRINTING).

**OBJECTIVES:**

This course is designed to broaden and enrich the student's knowledge and
understanding of statistical methodology as it pertains to the study of multivariate
techniques used in the behavioral sciences. Particular emphasis is placed on
procedures involved with multiple dependent variables studied simultaneously in a
comprehensive design. Since this is the fourth and most advanced course in the
statistical sequence, at the end of the semester a student should be able to read and
intelligently assess almost all of the research literature in a particular field.
Moreover, a student should be able to apply many of the multivariate techniques
presented in this course to his or her own research studies. Also, as a by-product of
this course, the student should develop a proficiency in the use of matrix algebra.

**REQUIREMENT:**

A ten-digit scientific calculator.

**EVALUATION:**

The course grade will be based on two equally-weighted components, weekly hand-in
assignments and a final examination. The final examination involves the solution of
a single data-set problem requiring the application of almost all the concepts
presented in the course. Classroom attendance is mandatory except for emergency
situations.

**TOPICS:**

- Some matrix algebra.
- Multiple regression and the simplification of its principles through matrices.
- The multivariate normal distribution, Hotelling's T
^{2}and Wilk's Lambda criterion for significance tests of group differences (one-way MANOVA). - More matrix algebra, linear transformations, and axis rotation.
- Eigenvalues and principal component analysis.
- Discriminant analysis using the generalized eigenvalue approach.
- Canonical correlation using the generalized eigenvalue approach.
- Multiple independent variables with multiple dependent variables (two-way MANOVA).

RETURN to Courses Taught on Home Page of Gary Ramseyer.