複變量統計分析
Multivariate Statistical
Analysis
Fall 2004 Course Syllabus
Graduate
The Course
Course
Title : Multivariate Statistical Analysis 複變量統計分析 Course
Number: 5895
Semester:
Fall 2004 Course
Units: 3.0
Time:
Friday
Course
Type: Lecture/Discussion/Lab Restriction:
Graduate students
Office:
818, SF Building(聖言樓818室) Office
hours: appointment via phone or email
Email:
hawjeng@mails.fju.edu.tw Tel:
(O) 02-29031111 ext. 2129 or (H)02-29777344
Mailbox:
台北縣新莊市242輔仁大學心理學系 Fax:
(O)02-29038438
本課程的目的在介紹各種常用於心理與教育領域的多變量統計分析方法與技術。作為一門研究所的計量課程，本課程將以基礎統計學的訓練為基礎，著重於各種多變量推論統計以及結構方程模式分析的基本概念與分析方法的介紹，例如區別分析、複變量變異數分析、因素分析、叢集分析、結構方程模式等等。課中除了原理性的討論，將配合統計軟體應用的介紹，例如SPSS、SAS、LISREL、EQS，使學生們不但能夠理解各種統計技術的原理與使用時機，並能夠具體的操作與分析，豐富學生們計量分析的知識素養與操作能力，協助學生們從事量化的研究工作。
This course introduces the multivariate
statistical techniques regularly used in educational research, such as MANOVA,
discriminant analysis, cluster analysis, canonical correlation, logistic
regression, etc. By integrating the application of computer software
such as SPSS or LISREL, this course expands beyond basic concepts of
statistical techniques to include more operational skills on doing research in
educational and relating issues.
The dual-objective of this course is to provide a sound conceptual
understanding on statistical issues, and to provide hands-on experience with
major statistical software packages.
Specific objectives this course seeks to accomplish are as follows:
1.
To
understand the basic principles of various multivariate statistical methods;
2.
To
acknowledge the advantages and limitations of statistical methods and
applications;
3.
To
apply multivariate statistical methods to practical research situations within
one’s own research area, and propose solutions to real-world research issues;
4.
To
analyze real data using software programs, SPSS or SEM software; and to propose
a complete paper on the basis of precise statistical language.
Course Format and Activities
This course combines
lecture and computer lab activities into one integrated format:
Lectures: The instructor covers each topic by
giving lectures in the class.
Lecture notes and reading material will be provided in advanced and
students are highly encouraged to be constructive and reflective preview of
assigned materials.
Computer
Labs: To enhance the concepts covered in the
class lectures, computer demos of related topics will be presented in class,
taking almost a half of the course hours.
Both real-life and simulated data will be used for these lab
activities. The computer labs take
one of two forms – demonstration of statistical software packages by instructor
or practice by students themselves.
Required: The required prerequisite for this
course is introduction to statistics or measurement course. Students without this background are
strongly encouraged to consult the instructor, although no enforcement will be
made to block enrollment.
Recommended: The recommended prerequisite is computer
knowledge and skills. Experience
with statistical packages and computer programming is highly desirable for
successful completion of this course.
Students without any background in computer use are encouraged to team
up with someone who does.
Lab homework:
This course
requires students to conduct four separate exercises in correspondent to the
class materials for lab work. The dataset
will be provided by lecturer during the class. Students are encouraged to bring with
their own data in order to get more insight from the output.
HW1: Exercise on Cluster Analysis and Log-linear analysis.
HW2: Exercise on MANOVA and DF, and compare with each other
HW3: Exercise on Logistic Regression, and compare with multiple regression
HW4: Exercise on Factor
Analysis
Evaluation of student
performance is based on the following criteria:
1.
Class participation (attendance, class interaction,
after-class discussion)………………………………… 10%
2.
Lab homework（10％ for each）……………………………………………………………………………… 40%
3.
Computer comprehension take-home examination
…………………………………………………………. 50%
Class Schedule
Revised
|
|
Week |
Topics & Activities |
homework |
|
1 |
9/17 |
課程簡介（Course introduction） |
|
|
2 |
9/24 |
複變量分析概說（Basic concepts to
multivariate statistical analysis） |
|
|
3 |
10/01 |
多變量統計的資料整備（Data screening and
preparing） |
|
|
4 |
10/08 |
集群分析（Cluster analysis） |
|
|
5 |
10/15 |
多元尺度分析（MDS） |
|
|
6 |
10/22 |
對數線性模式（Log-Linear modeling） |
HW1 |
|
7 |
10/29 |
區別分析（Discriminant Function
analysis） |
|
|
8 |
11/05 |
複變量變異數分析I（Multivariate analysis of
variance I） |
|
|
9 |
11/12 |
複變量變異數分析II（Multivariate analysis of
variance II） |
HW2 |
|
10 |
11/19 |
Midterm examination |
|
|
11 |
11/26 |
線性關係統計概說（Introduction to analysis
of linear relationship） |
|
|
12 |
12/03 |
多元迴歸（Multiple regression） |
|
|
13 |
12/10 |
邏輯迥歸（Logistic regression） |
HW3 |
|
14 |
12/17 |
典型相關（Canonical correlation
analysis） |
|
|
15 |
12/24 |
因素分析（Factor Analysis） |
HW4 |
|
16 |
12/31 |
結構方程模式I（Structural Equation
Modeling I） |
|
|
17 |
1/07 |
結構方程模式II（Structural Equation Modeling
II） |
Final exam distribute |
|
18 |
1/14 |
Final exam |
|
Textbooks
1.
Using
Multivariate statistics(4th Ed.) by Tabachnick & Fidell (2001)
2.
3.
4.
Applied Multivariate Statistical Analysis(5nd Ed.)
by Johnson & Wichern (2002)
5.
A first course on SEM by Raykov, T., &
Marcoulides, G. A. (2000)
6.
Structural Equation Modeling by Kaplan (2000)
7.
Structural Equation Modeling with EQS by Byrne
(1994)
8.
Principle and Practice of Structural Equation
Modeling by Kline (1996)
9.
Measurement, Design, and Analysis by Elazar
Pedhazur & Liora Pedhazur Schmelkin (1991)
10.
11.
(This
syllabus is subject to change upon mutual consent from the students and the
instructor.)