StockholmUniversity Fall 2011

Department of Statistics

Linda Wänström

Course Description for Multivariate Methods, basic level

Contents of the course

Often we are interested in the relationship between two sets of variables or the relationships among a large set of variables. Statistical methods for analyzing more than one independent and/or more than one dependent variable can be classified as multivariate methods. This course introduces some of the most important multivariate methods such as Principal Components Analysis, Factor Analysis, Cluster Analysis, Discriminant Analysis, MANOVA, Canonical Correlation and Structural Equations. Some knowledge in Matrix algebra is needed to understand the methods, and an overview of Matrix algebra is presented. Many multivariate methods rely on Multivariate normal assumptions, and this distribution is therefore covered. Practical applications are performed using the statistical software programme SAS.

LITTERATURE

  1. Sharma, S. (1996). Applied Multivariate Techniques. Wiley, New York.
  2. Extra material.

LEARNING GOALS

After taking the course the studentshould be able to

  • Describe the most common multivariate methods
  • Use statistical software to analyze data using some multivariate methods and interpret the results

EXAMINATION AND GRADING CRITERIA

The examination of the learning goals consists of an exam in two parts: A) a written part that covers the theory of multivariate methods (50 points maximum) and B) a computer part consisting of data analysis in SAS and documentation in MS Office Word (50 points maximum). In order to pass the course, the student needs at least 25 points on both of the exam parts.

Grading Criteria

A(Excellent): The student can apply multivariate methods with statistical inference, which is not directly covered in the course material, in a correct and well structured way. The student can present and interpret results, describe and show understanding of concepts, methods, and theory used in performing multivariate techniques in a clear and correct way. At least 90 points on the exam.

B(Very good): The student can apply multivariate methods with statistical inference, covered in the course material, in a correct and well structured way. The student can present and interpret results, describe and show understanding of concepts, methods, and theory used in performing multivariate techniques in a clear and correct way. 80-89 points on the exam.

C(Good): The student can apply multivariate methods with statistical inference, covered in the course material, in a correct and well structured way. The student can present and interpret results, describe and show understanding of concepts, methods, and theory used in performing multivariate techniques in a good way. 70-79 points on the exam.

D(Satisfactory): The student can apply multivariate methods with statistical inference, covered in the course material, in a correct way. The student can satisfactory present and interpret results, describe and show understanding of concepts, methods, and theory used in performing multivariate techniques. 60-69 points on the exam.

E(Sufficient): The student can apply multivariate methods with statistical inference, covered in the course material, in a mainly correct way. The student can satisfactory present and interpret results, describe and show understanding of concepts, methods, and theory used in performing multivariate techniques. 50-59 points on the exam.

F(Insufficient): 0-24 points on part A and/or part B.

For the grade F, the student may retake the exam at a later occasion (se schedule). No complementation is possible.

TEACHING

There are 12 lectures (F1-F12) and 6 computer labs (D1-D6), see separate schedule. The computer labs are mandatory. During the labs, students will practice using the statistical software SAS to analyze data using multivariate methods. Instructions for the labs will be handed out during each lab.

READING

Content / Preparation
F1 / Introduction. Matrix Algebra. The Multivariate Normal distribution / Sharma ch. 1 - 3
F2 / Principal Components Analysis / Sharma ch. 4
F3 / Exploratory Factor Analysis / Sharma ch. 5
F4 / Confirmatory Factor Analysis / Sharma ch. 6
D1* / Principal Components Analysis / Sharma ch. 4
F5 / Cluster Analysis / Sharma ch. 7
F6 / Discriminant Analysis / Sharma ch. 8
D2* / Exploratory Factor Analysis / Sharma ch. 5
F7 / Discr. Analysis / MANOVA / Sharma ch. 11
F8 / MANOVA / Sharma ch. 11
F9 / Test of Assumptions / Sharma ch. 12
D3* / Confirmatory Factor Analysis / Sharma ch. 6
F10 / Canonical Correlation / Sharma ch. 13
D4* / Discriminant Analysis / Sharma ch. 8
F11 / Structural Equations / Sharma ch. 14
D5* / MANOVA / Sharma ch. 11
F12 / Structural Equations / Sharma ch. 14
D6* / Structural Equations / Sharma ch. 14

*Mandatory attendance

EXAM:

Part A: 27 October 9-14. Laduvikssalen.

Part B: 31 October 9-12. B319/B397.

RE-EXAM:

Part A: 23November14-19. Brunnsvikssalen.

Part B: 28 November 13-16. B319.