36-402/608: Advanced Data Analysis IICourse Policies and SyllabusSpring 2010Course InformationLocation: A51 Baker HallTimes: TR 10:30-11:20Instructor: Teaching Assistants:Howard Seltman Gary Klein Jason [email protected] [email protected] [email protected] Baker Hall268-3938Office Hours:Howard: Mondays 3:00-4:00, Wednesdays 8:30-9:30 (BH232H)Gary: Wednesdays 3:00-4:00 (FMSB 320)Jason: Tuesdays 3:30-4:30 (FMSB 320)Please feel free to drop in, or e-mail to arrange another time.Web page:Blackboard (Read announcements weekly!)Backup page: http://www.stat.cmu.edu/∼hseltman/402/Physically disabled and learning disabled studentsThe Office of Equal Opportunity and Disability Services provides support services for both phys-ically disabled and learning disabled students. For individualized academic adjustment basedon a documented disability, see http://www.cmu.edu/hr/eos/disability/students/index.html.TextbookThe textbook for this course is Fred L. Ramsey and Daniel W. Schafer (2002). The StatisticalSleuth: A Course in Methods of Data Analysis Duxbury Press.Its home page is http://www.proaxis.com/∼panorama/home.htm. Data from the enclosed CDwill be used for homework assignments.PrerequisitesI assume that:• You have (more or less) mastered the material in 36-401, especially the material aboutregression models and diagnostics.• You have a working knowledge of R, and have it installed (at least version 2.9.0) on yourcomputer (update or install from http://cran.r-project.org/).• You have a strong desire to analyze data and a willingness to participate in class.1Course DescriptionAdvanced Data Analysis (ADA) I and II are key courses in the Statistics undergraduate program.In ADA I (36-401) you should have learned to apply the theory of linear regression analysis—including confidence intervals and hypothesis testing, basic distributions such as the normal, t,χ2and F distributions, basic facts and techniques for linear regression, including the materialabout regression models and diagnostics—in order to investigate model assumptions, and takeappropriate action if the assumptions don’t hold. In addition you should have learned somethingabout using R, and about writing literate data analysis reports.In ADA II (36-402), we will further develop those skills in exploring data, building and fittingmodels, investigating model assumptions, interpreting results from statistical models, and reportwriting, while learning a variety of more advanced models including PCA, logistic regression,Poisson regression, mixed models, and mediation analysis. We will also learn SAS.GradingParticipation in class: 20%Homework: 80%Total 100%Each class will include “breakout” exercises done in small groups. Full class participation creditwill be given to students who sign at least 80% of the breakout sheets and contribute periodicallyto class discussions.There will be eleven homework assignments, worth ten points each, with the lowest threedropped. Homeworks will be due by the beginning of class on Thursdays. Homework turnedin between Thursday 10:30 and Friday 10:30 will lose one point, homework turned in betweenFriday 10:30 and Monday 10:30 will lose 2 points. No credit is given for homework turned inafter that time. Cutoffs for letter grades A, B, C, and D are 90, 80, 70, and 60%.Academic IntegrityI encourage you to discuss the homework assignments with each other, but the work that youhand in must be your own. You must not copy mathematical derivations, computer output andinput, or written descriptions from anyone or anywhere else, without reporting the source withyour work. Please review the CMU policies on cheating and plagiarism.Communications• Homework, data sets, and R code for this class may be found at Blackboard or the backupweb page.• The TAs ([email protected] and [email protected]) and the instructor([email protected]) may be contacted by email. Please feel free to send us emailwith questions, comments, etc., anytime. Also, please feel free to drop by Howard’s officeor schedule special appointments by email anytime.2Other Sources of InformationA list of useful statistics texts by topic is at http://www.amstat.org/sections/cnsl/BooksJournals.cfm.One good place to search the statistics literature is the Current Index of Statisticshttp://www.statindex.org/CIS/psqlQuery.More information about R can be found at my R web page:http://www.stat.cmu.edu/∼hseltman/RTips.html.Tentative ScheduleWeek of Tuesday ThursdayJan 11 Statistics review R reviewJan 18 Simulation and Robustness Alternatives: Permutation tests andclassic non-parametricsJan 25 Outliers and Missing data Power and Model SelectionFeb 1 ANOVA I ANOVA IIFeb 8 Models with structural assumptions Nested ANOVAFeb 15 Serial correlation I Serial correlation IIFeb 22 Repeated Measures I RM IIMar 2 Multivariate Responses I Multivariate Responses IIMar 9 Spring Break Spring BreakMar 16 SAS Mixed Models in SASMar 23 Binary Outcomes Tables of CountsMar 30 Logistic Regression I Logistic Regression IIApr 6 Poisson Regression I Poisson Regression IIApr 13 Some Experimental Design Issues CarnivalApr 20 Coding a new test I Coding a new test IIApr 27 Mediation I Mediation