EDEP768
Seminar in Multilevel Modeling
University of Hawaii at Manoa
Fall, 2011
Description: General introduction to multilevel methods of data analysis for doing research in education, psychology, management, and other fields. The course develops concepts of measurement, design, and analysis. Focus is on statistical inference using various multilevel methods to investigate different types of research problems having nested, longitudinal, and cross-classified data structures.
Course Hours: Tuesday 4:30-7:00 in Wist Hall 233.
Instructor: Ron Heck ()
Office: Wist 220A, Phone: 956-4117, 9567843 (message)
Office Hours: Generally Tuesday, 1:004:00, or by appointment.
Recommended Text: Heck, Thomas, & Tabata (2010). Multilevel and Longitudinal Modeling with IBM SPSS. NY: Routledge.
Course Outline
I. Purpose
The purpose of this course is to develop an understanding of the use, application, and interpretation of multilevel modeling in the context of educational, social, and behavioral research. The course is intended to acquaint students with several related techniques used in analyzing quantitative data with nested data structures which include hierarchical, cross-classified, and repeated measures. Moreover, the course is designed to introduce students to the use of various software packages that can be used to analyze multilevel data structures, focusing primarily on IBM SPSS and Mplus. Emphasis in the course is on the mastery of concepts and principles, development of skills in the use and interpretation of software output, and development of critical analysis skills in interpreting and writing up research results using the techniques we cover.
It is assumed that you have had at least an introductory course in research and statistics as well as a multivariate statistics course.Lab computers,or your own laptops, will be used to learn how to develop analyses, display results, and provide write-ups using various techniques. SPSS should be available in the computer lab. Access to the Mplus software will be through the Mplus demo version (which can be downloaded from or one of two computers in the downstairs resource lab. The demo version has some limitations in terms of data sets and number of variables, which we will work around.
II. Course Design—The course addresses EDEA student outcome #3 (Developing educators who can use frameworks informed by theory, research, and practice to solve problems and generate policy). Course content focuses on
1. Understanding the development of theory through empirical testing and the utilization of theory in proposing multilevel models and testing them against the data.
2. Understanding how we explain events in science (e.g., description, correlation, causality).
3. Developing facility in completing a variety of cross-sectional and longitudinal analyses using the IBM SPSS and/or Mplus statistical packages.
III. Course Objectives: The objectives of the course are as follows:
1. Assist students in understanding the purposes of disciplined inquiry in education and the social and behavioral sciences.
2. Provide students with an introduction to more advanced analytic tools to investigate problems linked to more general conceptual frameworks and theories and to investigate those problems.
3. Help students gain facility with the use of the computer to analyze data.
IV. Course Requirements: Attainment of these objectives will be assessed by:
1. Students will submit 3 required short papers focusing on the analysis, interpretation, and presentation of data relevant to the analytic techniques discussed during the preceding period. The assignments will include a two-level analysis with a continuous outcome, atwo- or three-level modelwith categorical outcome, and a longitudinal (growth) model. (60%)
Each assignment will be graded using the following holistic scoring system: Excellent = 19-20 points; Good = 17-18; Fair = 14-16 points; Poor = 13 points or below. Points will be awarded according the following criteria: (i) general organization and clarity of assignment (excellent=4, fair=3, poor=2), (ii) technical accuracy (good=7 fair=5, poor=3), (iii) flow of arguments presented (good=5, fair=4, poor=3), and (iv) depth of coverage (good=4, fair=3, poor=2). Each assignment will have a due date.
2. A final paper based on a multilevel or longitudinal data set, ideally on a topic in which you have a particular interest. The paper should include an introduction to the problem, purpose of the analysis, a relevant theory or proposed conceptual model (that can be defined and tested against the data), results, and interpretation of the results. The idea is to integrate what you have been learning over the semester. For example, the paper could be a first-time analysis of some data set (perhaps your own data set), or one associated with a multivariate multilevel or cross-classified data set from the book (last part of the course) (30%)
3. Attendance, class assignments, and other course participation. If personal problems arise during the course it is helpful to keep me informed (e.g., illness or other issues that may affect your ability to complete the course, etc.). (10%)
V. Tentative Topics of Study
The Research Process
Aug. 23Introduction to the course. Logic of testing models. Conceptual frameworks, theories, and models.
Aug. 30Overview of issues in multilevel modeling. Read Heck et al. Ch1.
Sep. 06Data management issues. Categorical versus continuous outcomes.
Read Ch. 2.
Sep. 13From single-level to multilevel models. Class assignment on defining a basic multilevel model within random intercept. Read Ch. 3
Sep. 20Continue with two-level models.Investigating random slopes.
Start on Assignment #1, working with two-level models.
Sep. 27Extending the two-level model to a three-level model
Read Ch. 4
Assignment #1 due.
Oct. 04Finish up three-level models.
Oct. 11Defining a two-level model with dichotomous outcome.
Read handout on dichotomous multilevel models.
Oct. 18Work on assignment with dichotomous or other categorical outcome.
Oct. 25Begin repeated measures analysis (individual-level models).
Ch. 5
Assignment #2 due.
Nov.01Continue with individual-level growth models. Class activity.
Nov. 08Multilevel growth models.
Read Ch. 6.
Nov.15Work on multilevel growth models.
Nov. 22Defining multilevel multivariate models
Read Ch. 7
Assignment #3 due.
Nov. 29 Cross-classified data structures
Read Ch. 8
Dec.06Wrapping up.
Read Ch. 9
Dec. 14Final Assignment due (4 pm).