PH 717: Applied Longitudinal Data Analysis Spring 2016 Location: All lectures will be held in the Joseph J. Zilber School of Public Health building, Room 110 Schedule: Fridays 9:30 AM-12:10 PM Instructor: Cheng Zheng, PhD Assistant Professor Joseph J. Zilber School of Public Health Office Hours: By appointment via email Email: zhengc@uwm.edu Office phone: (414) 227-3015 Course Description: Short description: This course will cover data analysis techniques for longitudinal data with focus on application in public health with related examples using SAS. Full description: This course will cover data analysis techniques for longitudinal data that are time-dependent. Methods and techniques will include mixed models, marginal models and transition models for longitudinal data analysis. Techniques to handle missing and drop out events for longitudinal data analysis will also be covered. The course will be taught in an applied perspective with emphasis on public health applications. The students may use any software they are familiar with (e.g. R), but the examples will be taught in SAS. Prerequisites: PH 711 (P) and PH712 (P), or consent of instructor Course Learning Outcome: By the end of the course, students will be able to 1. Select the appropriate analysis method for longitudinal data. (MPH #6, Biost #2, #3, #4, #5) 2. Analyze longitudinal data and make correct inferences to answer scientific questions using statistical software. (MPH #8, Biost #6) 3. Communicate with public health collaborators for their model selection and interpretation of results. (MPH #11, Biost #1, #8, #10, #11) 4. Critique longitudinal data analysis methods used in published public health literature. 1 (MPH #8, Biost #8) Class Format: Lecture Credit Hours: This course counts for 3 graduate credits. Investment of time outside of class will vary student-by-student. The requirements for this class will require no less than 48 hours per credit hour or no less than 144 hours of time in accordance with UWM’s Credit Hour Policy. The workload is an estimate and that students are assessed on their performance, not on the time put into the course. Required and recommended readings: Required: Fitzmaurice, G.M., Laird, N.M. and Ware, J.H. (2011) Applied longitudinal analysis, 2nd edition. New York: Wiley. Optional: Diggle, P.J., Heagerty, P., Liang, K.Y. and Zeger, S.L. (2002) Analysis of longitudinal data, 2nd edition. Oxford University Press Optional: SAS/STAT(R) 9.3 User’s Guide. http://support.sas.com/documentation/cdl/en/statug/63962/HTML/default/viewer.htm#titlep age.htm Course Requirements: To meet course objective, students are expected to: 1. Complete reading assignments and attend class with good preparation; 2. Complete homework assignments in a timely manner; 3. Complete midterm and final exams; 4. Complete the course project. Attendance: Attendance is required. From the 3rd absence, 1% will be deducted from total course score for each absence. Classroom conduct: All students will be expected to conduct themselves in a professional and collegial manner at all times. Assignments: Homework: A total of 12 homework assignments will be distributed. Homework assignments will be distributed on Fridays and due in one week. Each task will be graded on a scale of 0-10. The lowest two will be dropped. The purposes of these assignments are to help students understand when to select each model, how to implement and interpret each model and how to critique an incorrectly used method. The assignment will include short-answer questions 2 based on supplement reading materials and small data analysis projects. The homework aims to prepare students for the exams. Course project: A data analysis project with several specific scientific questions will be assigned to the student early in the semester. The students are expected to select the appropriate analysis method and implement it using statistical software. Then the students shall write a data analysis report to answer the scientific questions using output from statistical software. The report shall be written in a form that a non-statistician collaborator can understand. In the report, the student shall write clearly the method they use and the interpretation of their result. Exams: Both mid-term and final exams will consist of a series of multiple-choice and short answer questions. The exam aims to evaluate students’ knowledge of different models, and ability to select and interpret the model correctly. The exams’ form will be in-class closed book with one double sided A-4 cheat sheet allowable. These exams aim to test whether the student grasps the materials covered in class. Late homework/project: late assignment submission will be accepted with 10% point deducted per each late day (i.e. you will not obtain score if you submit 10 days after due date). Grading: The weights assigned to each of assignments above will be MPH Program % of Grade Competencies Assignment Due Date Homework Midterm Exam Fridays (See outline) March 18th 30 #6 15 #6 Course Project Final Exam April 22nd May 12th 30 #6, #8, #11 25 #6 For this course, grades will be based on the following scale: Percent 94 – 100% 90 – 93% 87 – 89% 84 – 86% 80 – 83% 77 – 79% 74 – 76% 70 – 73% 67 – 69% Letter Grade A AB+ B BC+ C CD+ 3 Track Competencies #2, #4, #8, #11 #2, #4, #5, #8 #1, #2, #3, #4, #5, #6, #10, #11 #2, #4, #5, #8 64 – 66% 60 – 63% < OR = 59% D DF Course Evaluations: The Joseph J. Zilber School of Public Health administers end of semester course evaluations. Students enrolled in this class will receive an evaluation via PantherMail during the last full week of the semester, and they must complete it before the last day of final exams. If you do not use your PantherMail, then please forward all messages to your primary email account, so you do not miss this correspondence. General Information: In the event of disruption of normal classroom activities due to an outbreak, or any other public health emergency, the format for this course may be modified to enable completion of the course. In that event, you will be provided an addendum to this syllabus that will supersede this version. Incomplete Grade: An "Incomplete" grade will be given only for a major reason that occurs at the end of the semester and only if the bulk of the course work is complete. The student must make arrangements with me to complete the course work by a designated time. Contesting a grade: Students are expected to contact the instructor within 2 weeks of receiving a grade on any assignment if the student feels she/he was graded unfairly. Comprehensive information on UWM policy: Specific points are mentioned below. The policy can be found at http://www.uwm.edu/Dept/SecU/SyllabusLinks.pdf. Accommodation for Religious Observance: Students will be allowed to complete examinations and other requirements in advance of religious observance given that the student informs the instructor at the beginning of the semester or no later than 3 weeks prior to absences related to religious observance. Drop /Withdrawal/Repeat Policies: A student may drop a full-term course(s) through the end of the eighth week of classes. Special Needs: Students in need of special accommodations in order to meet course requirements are expected to contact the instructor as soon as possible to make arrangements. Cancellation of Class: If the canceled class is an exam date, the exam will be held on the next class day. If weather conditions warrant the cancellation of class, the UWM home page, the radio or TV will announce the closing. Also check UWM email. Academic Misconduct Policy: Academic misconduct is an act in which a student seeks to claim credit for the work or efforts of another without authorization or citation, uses unauthorized 4 materials or fabricated data in any academic exercise, forges or falsifies academic documents or records, intentionally impedes or damages the academic work of others, engages in conduct aimed at making false representation of a student's academic performance, or assists other students in any of these. Prohibited conduct includes cheating on an examination; collaborating with others in work to be presented, contrary to stated rules of the course; submitting a paper or assignment as one's own work when a part or all of the paper or assignment is the work of another; submitting a paper or assignment that contains ideas or research of others without appropriately identifying the sources of those ideas; stealing examinations or course materials; submitting, if contrary to the rules of a course, work previously presented in another course; tampering with the laboratory experiment or computer program of another student; knowingly and intentionally assisting another student in any of the above, including assistance in an arrangement whereby any work, classroom performance, examination or other activity is submitted or performed by a person other than the student under whose name the work is submitted or performed. In fairness to all students and to promote academic integrity, the Instructor of this course accepts responsibility to deal effectively with any instance of academic dishonesty should it occur. Students who violate academic standards as set forth in UWS Chapter 14 and UWM Faculty Document 1686 (http://www4.uwm.edu/acad_aff/policy/academicmisconduct.cfm) will be confronted and must accept the consequences and sanctions levied against them for their actions. Plagiarism and Cheating: (Student Handbook – pgs. 154-155) Dishonesty, including but not limited to cheating, plagiarism, or knowingly supplying false information or deceiving the school and its officials is a violation of the student conduct policy. Any student who is found to have violated this policy is subject to disciplinary sanctions up to and including suspension or permanent dismissal. Please be aware that plagiarism is presenting another’s ideas as one’s own and includes paraphrasing as well as copying without proper citations or quotation marks. What is copyright? Copyright is a form of protection provided by the laws of the United States (title 17, U.S. Code) to the authors of “original works of authorship,” including literary, dramatic, musical, artistic, and certain other intellectual works. This protection is available to both published and unpublished works. Section 106 of the 1976 Copyright Act generally gives the owner of copyright the exclusive right to do and to authorize others to use their materials. You must get permission to use copyrighted original works of authorship if you plan to make your project available to the public in any way. For more on gaining permission, see: http://www4.uwm.edu/ltc/copyright/getting-permission.cfm 5 Course Outline FLW: Fitzmaurice, G.M., Laird, N.M. and Ware, J.H. (2011) Applied longitudinal analysis, 2nd edition. New York: Wiley. DHLZ: Diggle, P.J., Heagerty, P., Liang, K.Y. and Zeger, S.L. (2002) Analysis of longitudinal data, 2nd edition. Oxford University Press Weeks Week 1 1/29 Description Introduction; FLW: Chap 1.1-1.4, Longitudinal and clustered data FLW: Chap 2.1-2.5, Longitudinal data: basic concepts Week 2 2/5 Overview of linear models; HW1 due FLW: Chap 3.1-3.6, Overview of linear models for longitudinal data Week 3 2/12 Maximum likelihood and restricted maximum likelihood estimator; HW2 due FLW: Chap 4.1-4.5, Estimation and statistical inference Week 4 2/19 Week 5 2/26 Modeling the mean; HW3 due FLW: Chap 5.1-5.9, Modeling the mean: analyzing response profiles FLW: Chap 6.1-6.6, Modeling the mean: parametric curves Modeling the covariance; HW4 due FLW: Chap 7.1-7.8, Modeling the covariance Week 6 3/4 Linear mixed effect models; HW5 due FLW: Chap 8.1-8.9, Linear mixed effect models Week 7 3/11 Comparison between fixed effect and random effect models; Model checking; HW6 due FLW: Chap 9.1-9.8, Fixed effects versus random effects models FLW: Chap 10.1-10.7, Residual analysis and diagnostics Review of generalized linear models; Mid-term Exam FLW: Chap 11.1-11.8, Review of generalized linear models Week 8 3/18 Week 9 3/25 Marginal model; Generalized estimating equations; HW7 due FLW: Chap 12.1-12.4, Marginal models: introduction and review FLW: Chap 13.1-13.6, Marginal models: generalized estimating equations (GEE) 6 DHLZ: Chap 8.1-8.4, Marginal Models Week 10 4/1 Week 11 4/8 Generalized linear mixed effects models; HW8 due FLW: Chap 14.1-14.8, Generalized linear mixed effects models FLW: Chap 15.1-15.7, Generalized linear mixed effects models: approximate methods of estimation DHLZ: Chap 9.1-9.4, Random effects models Comparison between marginal and mixed effect model; Overview of missing data; HW9 due FLW: Chap 16.1-16.6, Contrasting marginal and mixed effects models FLW: Chap 17.1-17.6, Missing data and dropout: overview of concepts and methods DHLZ: Chap 13.1-13.4, Missing values in longitudinal data Week 12 4/15 Multiple imputation and weighting method; modeling drop-out process; HW10 due FLW: Chap 18.1-18.7, Missing data and dropout: multiple imputation and weighting methods DHLZ: Chap 13.5-13.8, Missing values in longitudinal data Week 13 4/22 Transition models; Course project due DHLZ: Chap 10.1-10.4, Transition models Week 14 4/29 Sample size; Power; Design; HW11 due FLW: Chap 20.1-20.6, Sample size and power FLW: 21.1-21.6, Repeated measures and related designs Week 15 5/6 Multilevel models; Time dependent covariate; HW12 due FLW: Chap 22.1-22.5, Multi-level models DHLZ: Chap 12.1-12.4, Time-dependent covariates Appendix A: Competency Sets Addressed in this Course MPH Core Competency: #6 Utilize appropriate quantitative and/or qualitative methods in public health practice and research. #8 Collect, synthesize and critically analyze information and data to identify and address, and inform public health issues and interventions. 7 #11 Communicate effectively about public health issues with diverse audiences using a variety of strategies and modalities. Biostatistics Track Competency: #1 Function as a collaborator with community partners on public health projects and in developing recommendations for appropriate study designs that advance social justice and population health. #2 Translate research objectives into testable hypotheses. #3 Differentiate between quantitative problems that can be addressed with routine methods and those requiring input from a doctoral-level biostatistician. #4 Demonstrate a broad knowledge and understanding of statistical techniques used in public health studies and health-related scientific investigations. #5 Identify and apply a variety of appropriate statistical methods for developing inferences about public-health-related questions. #6 Demonstrate basic programming skills in multiple statistical software packages and data management and integration techniques for public health and big data projects. #8 Interpret and critique statistical analyses in publications for public health professionals. #10 Demonstrate effective written and oral communication skills when reporting statistical results to different audiences of public health professionals, policy makers and community partners. #11 Formulate and produce graphical displays of quantitative information (e.g., scatter plots, box plots and line graphs) that effectively communicate analytic findings. 8