Instructor Dr. Jeffrey Harring 1230E Benjamin Building Phone: (301) 405-3630 Email: harring@umd.edu Office Hours Tuesday 2:00-3:00pm, or by appointment Course Objectives, Description and Prerequisites Longitudinal data often arise from repeated measurements collected over time or other conditions. The primary objective of this course is to equip you with the tools necessary to analyze longitudinal data. The course culminates with a final project (more will be said in regards to specifics of the final project later in the semester). The assignments will provide SAS and R code needed for performing the myriad of steps required for the final project. This course will cover modern approaches to analyzing longitudinal with emphasis on linear mixed effects models for normal data, generalized linear models for non-normally distributed data, such as response variables following binomial or Poisson distributions and nonlinear mixed effects models for nonlinear profiles. Traditional methods such as repeated measures ANOVA/MANOVA will be discussed in reference to linear mixed effects models. We will discuss graphical data exploration, correlation structures, parameter estimation / testing / inference, model selection, diagnostics and model limitations. Modeling will be done using R (see notes below). Prerequisites This course requires prior experience with inferential statistics up through multiple regression (e.g., EDMS 651); some familiarity with matrix notation; or permission of the instructor. Guidelines Students should be familiar with the basic notation of random variables, statistical inference (confidence intervals, hypothesis testing), multiple linear regression and logistic regression. Familiarity with matrix notation is helpful; we will review this very briefly at the beginning of the course. During the semester, the underlying statistical theory will be outlined using matrix notation, but deep understanding of the theory is not necessary to complete the out-ofclass assignments or exams; the main focus will be on applications. The course is accessible to graduate students in all fields. 1
Required Textbook Fitzmaurice, G. M., Laird, N. M., & Ware, J. H. (2004). Applied Longitudinal Analysis. New York: Wiley. Additional Reference Textbooks Students may find the following texts useful for more in-depth mathematical treatment of course topics or alternative perspectives. Note, none of these texts is required. Hand, D. J., & Crowder, M. (1996). Practical Longitudinal Data Analysis. New York: Chapman & Hall. Hedeker, D., & Gibbons, R. D. (2006). Longitudinal Data Analysis. New Jersey: Wiley. Pinheiro, J. C., & Bates, D. M. (2000). Mixed-Effects Models in S and S-PLUS. New York: Springer-Verlag. Singer, J., & Willett, J. (2003). Applied Longitudinal Data Analysis. New York: Oxford. Verbeke, G., & Mollenberghs, G. (2000). Linear Mixed Models for Longitudinal Data. New York: Springer-Verlag. Weiss, R. E. (2005). Modeling Longitudinal Data. New York: Springer-Verlag. Computing All statistical modeling for in-class lectures will be presented in SAS and R. R, is a free implementation of the S statistical computing language and environment. S has become more or less the standard language for statistical computing, at least among statisticians. You can download R at the home page of the R project: http://www.r-project.org. R is made available for several operating systems including OS X (Macintosh), Linux/Unix as well as Windows. Can I use another, more familiar, statistical software package? Yes. There are many different options at you disposal: SPSS, LISREL, Stata, etc. While these programs can be used in completing the assignments and final project, they will not be officially supported. Saying that, if you are quite comfortable in SAS, then there may be little reason to jump on the R bandwagon. Homework There will be several homework assignments given throughout the semester. I encourage you to work together in computing and discussing the problems. However, each student is expected to independently write up the submitted assignment using her or his own computing 2
and giving explanations in her or his own words. All assignments will involve computing; please attach only relevant computer output to what you turn in. Some assignments may also include reading and writing about a related journal article. You will get about two weeks to work on each homework assignment. Exam There will be one in-class exam towards the middle of the semester. The content of the midterm exam will cover topics presented in class up to that point. The exam will be openbook & self-created crib sheet. My notes will be off-limits for the exam. Students are allowed 3 sheets of notes (front and back) to use on the exam. Students will want to bring a calculator to the exam. Project There will be a final project towards the end of the semester (no final exam). Project details will be handed out in the beginning of March. At that time each student must submit a project proposal of no more than one page and get approval from the instructor by early April. Unlike the homework, the final project will be completed by you alone, without the aid of discussing the computing, approaches, etc. with other classmates, students outside the class or faculty (not even your mother). Students are on their honor to do the final project completely independently; any student found doing otherwise will be subject to the maximum University penalties. Grading This course is not graded on a curve. Your homework, exam and project will be combined using a weighted average grading scheme with the corresponding weights given below. Final letter grades will then be assigned based on the given scale (there will be no rounding). Assessment Weight Overall Course Percent Grade Total homework points 40% 100 93% A Total midterm exam points 30% 92 -- 90% A- Total final project points 30% 87 -- 89% B+ 83 -- 86% B 80 -- 82% B- No Extra Credit 77 -- 79% C+ 73 -- 76% C Your course grade will be based 70 -- 72% C- only on the above assessments. 67 -- 69% D+ There will be no extra credit 63 66% D opportunities. Please do not ask for 60 62% D- exceptions. 60% F 3
Incompletes Incompletes for this course will be given on a case-by-case basis. The most valid reason for an incomplete is an unforeseen event that gravely interferes with a student s ability to perform at an adequate level. Incompletes will not be given for unqualified poor performance. Course Topics and Readings Below is a list of topics and readings we will cover throughout the semester. I do not provide a detailed calendar, as I am not sure how long topics will take to cover in class. However, for the most part, we will follow the order of listed topics. Students should keep up with the readings whether or not the material is explicitly covered in class or not. Topics Fitzmaurice et al.. Chapter Introduction 1, 2 Longitudinal Data Structures 2, 3 Profile Analysis 5 Linear Mixed Effects Model A, B, 6, 8 Estimation and Testing 4 Modeling Covariance Structures 7, 8, 9 Non-normal Response Variables 10, 11 Nonlinear Profiles 12, 13 Missing Data 14 Dynamic Covariates 15 Other Topics if Time Permits Accommodations for Emergencies In the event that the University closes on the day of class (for instance, a huge hurricane rips through the campus), we will obviously have no class. However, if the University does not shut down and there is a threat of inclement weather, etc., we will still have class unless you hear from me otherwise. With that said, please check your email and/or the course website, sometime during the day of class for any last minute postings of announcements regarding the course. If you need to be gone from class, out of common courtesy I would appreciate if you would send me an email to let me know. All students are expected to take the exams and/or submit assignments on the specified dates and no make-up exams are planned (see section on make-up exams below). You must contact me before an exam if you are going to be absent or you will receive a zero for that assessment. 4
Academic Accommodations In compliance with and in the spirit of the Americans with Disabilities Act (ADA), I want to work with you if you have a documented disability that is relevant to successfully completing your work in this course. If you need academic accommodation by virtue of a documented disability, please contact me as soon as possible to discuss your needs. Students with documented needs for such accommodations must meet the same achievement standards required of all other students, although the exact way in which achievement is demonstrated may be altered. All requests for academic accommodations should be made as early as possible in the semester. For further information concerning disability accommodations, please contact Dr. William Scales at the Disability Support Service (301) 314-7682. Academic Integrity The University of Maryland, College Park, has a nationally recognized Code of Academic Integrity, administered by the Student Honor Council. This Code sets standards for academic integrity at Maryland for all undergraduate and graduate students. As a student you are responsible to uphold these standards for this course. It is imperative that you are aware of the consequences of cheating, fabrication, facilitation, and plagiarism. For more information on the code of Academic Integrity or the Student Honor Council, please visit http://www.studenthonorcouncil.umd.edu/code.html for details. On plagiarism -- It is important that the student synthesize pertinent information from the readings and class lectures when writing up homework assignments. Synthesis does not occur when large blocks of text are copied from the textbook or my notes and used to answer questions. It is understood that the student will have to use some verbatim phrases and definitions from the textbook or notes. This is not considered a case of scholastic misconduct. What must be avoided is extensive verbatim copying of information from the textbook or my notes when answering the longer questions on the assignments. Make-Up Examinations The University policy states: An instructor is not under obligation to offer a substitute assignment or to give a student a make-up assessment unless the failure to perform was due to an excused absence, that is, due to illness (of the student or a dependent), religious observance (where the nature of the observance prevents the student from being present during the class period), participation in university activities at the request of university authorities, or compelling circumstances beyond the student s control. Students claiming excused absence must apply in writing and furnish documentary support for their assertion that absence resulted from one of these causes. 5