STATISTICS WEEK OF SHORT COURSES 2006
The University of Florida

The Department of Statistics at the University of Florida has 30 faculty members, many of whom are internationally recognized, and more than 50 graduate students. The department prides itself on the balance of theory and applications taught in its curriculum. The strengths of the department include Categorical Data Analysis, Nonparametric Statistics, Longitudinal Data Analysis, Statistical Modeling, Design of Experiments, Monte Carlo and Computer Intensive Methods. The instructors teaching these courses are experienced instructors who have extensive experience both as consultants to government and industry and as instructors.

INSTRUCTORS

James Booth (Professor, Cornell University, Department of Biological Statistics & Computational Biology, Affiliate Professor at the University of Florida), has published over 30 articles on statistical methodology, and has taught a wide variety of courses on the theory and application of statistics including topics such as Linear Models, Multivariate Analysis, Categorical Data Analysis, and Generalized Linear Models.

George Casella (Distinguished Professor, Chair) is an active researcher in decision theory, statistical confidence, environmental statistics, statistical genomics and the theory and application of Monte Carlo and other computationally-intensive methods. He is a Fellow of the American Statistical Association and the Institute of Mathematical Statistics, has served as Theory and Methods Editor of “Journal of the American Statistical Association” and has authored five textbooks, including: Statistical Inference, 2nd ed., 2001, with Roger Berger; Variance Components, 1992, with S.R. Searle and C.E. McCullough; Theory of Point Estimation, Second Edition 1998, with Erich Lehmann, and Monte Carlo Statistical Methods, 1999, with Christian Robert. More about his academic record can be found at http://www.stat.ufl.edu/~casella

Mary Christman (Associate Professor) is active in statistical modeling for environmental studies. Her interests include geostatistical modeling for zoogeography and biodiversity, statistical methods for accounting for uncertainty and for sampling error in models of population dynamics, statistical methods of adjusting for censoring, methods for statistical analyses when data have zero-inflation, and developing sampling designs and estimators for rare and elusive species. She has received the Distinguished Achievement Award from the American Statistical Association's Section on Statistics and the Environment and the Graduate Education Award from the Marine-Estuarine-and Environmental Sciences Program at the University of Maryland.

Malay Ghosh (Distinguished Professor) has made major contributions to several areas of statistical inference, both frequentist and Bayesian. He is been engaged in research on parametric and semi-parametric Bayesian inference related to case-control data. Dr. Ghosh has taught short courses in USA, Finland and Iran, on Bayesian Data Analysis in the Deming Conference and Merck Research Laboratories. He has published two books and more than 200 research articles. He is a Fellow of the American Statistical Association, Institute of Mathematical Statistics, and is also an elected member of the International Statistical Institute. More about his academic record can be found at http://www.stat.ufl.edu/~ghoshm
Jeff Gill (Associate Professor, U.C.-Davis, Department of Political Science, Affiliate Professor at the University of Florida). Major areas of research are: Markov chain theory, MCMC, Bayesian hierarchical models, elicited prior specifications, and social science applications. Books authored include: Numerical Issues in Statistical Computing for the Social Scientist (Wiley), Generalized Linear Models: A Unified Approach (Sage), and Bayesian Methods (Chapman & Hall). He has also published articles in “Statistical Science”, “Sociological Methods and Research”, “Political Analysis”, “Electoral Studies”, and other scholarly journals.
Bhramar Mukherjee (Assistant Professor) is an active researcher in statistical analysis of case-control studies, Bayesian methods and experimental design. She currently supervises several doctoral students working on diverse design and inference questions in epidemiology and has authored several articles in this area. Dr. Mukherjee has considerable collaborative experience working with medical practitioners, biologists, epidemiologists and geneticists. She has offered several statistical methods courses to students from non-quantitative disciplines, emphasizing how to use standard statistical software like SAS and SPSS to solve real life problems. More about her academic record can be found at http://www.stat.ufl.edu/~mukherjee
P.V. Rao (Professor Emeritus), has consulted extensively within the University of Florida Health Science Center and as a Statistician in the Pediatric Oncology Group. He is the author of the text Statistical Research Methods in the Life Sciences (Brooks/Cole) and has considerable experience in the application of survival analysis methods to studies on the treatment of childhood cancers. An elected member of the International Statistical Institute, he has published extensively in the theory, methodology and application of statistics. More about his academic record can be found at http://www.stat.ufl.edu/~pvrao
Alex Trindade (Assistant Professor) is an active researcher in time series and statistical finance. He regularly teaches a time-series core course, structured into an applied computer intensive format, in UF's Quantitative Finance co-Ph.D. program. He has been involved in reliability research projects with The Boeing Company and the Institute for Defense Analysis and has published in the “Annals of Statistics”, the “Journal of Multivariate Analysis”, and “Statistica Sinica”. More about his academic record can be found at http://www.stat.ufl.edu/~trindade
Mark Yang (Professor Emeritus) has worked extensively in theoretical and applied statistics and recently in statistical genetics. He is the author of Introduction to Statistical Methods in Modern Genetics, (2000). He has worked with Bell Labs, the Department of Energy, the Naval Research Lab and NASA. He received the 1998 University of Florida's Excellence in Teaching Award. Dr. Yang has published numerous papers in statistical, IEEE and genetics journals and is an Associate Editor "Statistica Sinica". He is a Fellow of the American Statistical Association and currently serves as president of the Florida Chapter of ASA. More about his academic record can be found in http://www.stat.ufl.edu/~yang/publications/index.html
Yasar Yesilcay (Senior Lecturer) received a Ph. D. in biostatistics from the University of North Carolina, as well as a Certificate of Demography from Office of Population Studies of Princeton University and Certificate of Sampling from the Survey Research Center at University of Michigan. He has led a number of local and national surveys in Turkey and taught in Turkey, Jordan, Oman and the USA for 30 years. He has contributed to designing and implementing undergraduate curricula for a minor in statistics for three different universities in three countries (Middle East Technical University, Ankara, Turkey, Sultan Qaboos University, Muscat, Oman and James Madison University in Virginia).
Linda Young (Professor) has worked extensively in statistical modeling of ecological and environmental studies. She is lead author on the text: Statistical Ecology: A Population Perspective, Springer, 1988. Her interests include statistical methods for combining data sets with differing spatial support, geostatistical modeling for public health and ecological processes, statistical methods of adjusting for censoring, and sampling of biological populations. Dr. Young is past Vice President of ASA, past President of the International Biometric Society Eastern North American Region and current Chair of the Committee of Presidents of Statistical Societies.

COURSE DESCRIPTIONS

Introduction to Categorical Data Analysis
March 13-14: James Booth

An introductory course on statistical methods for analyzing categorical data. Topics include variable classification and sampling, two and three-way tables, logistic regression, multicategory logit models, and methods for dealing with overdispersion and dependent responses. A basic knowledge of statistical concepts and methodology, such as confidence intervals, p-values, t-tests and linear regression, will be assumed. Application of the methods will be illustrated using SAS and R.

Recommended text: An Introduction to Categorical Data Analysis by Alan Agresti. (Wiley, 1990; 2nd Edition, 2002, ISBN: 0471226181).

Data Analysis using Monte Carlo
Methods & Bayesian Stochastic Simulation CANCELED
March 13-14: George Casella & Jeff Gill

Monte Carlo statistical methods, particularly those based on Markov Chains, have matured to the point that they are now part of the standard set of techniques used by statisticians. This short course serves as an introduction to the application and underlying workings of Monte Carlo methods.

A variety of interesting statistical applications are used to illustrate how these methods enhance statistical practice. The course begins with basics of random number generation and illustrations of how simulation approaches often supply easy methods for solving difficult problems. We explore techniques for Monte Carlo integration and optimization, and the more recent Markov chain Monte Carlo techniques such as the Gibbs sampler and the Metropolis-Hastings Algorithm. We will evaluate in detail the process of designing and coding Bayesian statistical models that do not have closed-form analytical solutions. These methods are particularly useful in analyzing data that are modeled using a hierarchical structure.

We will use examples from economics, engineering, and political science/public policy, biostatistics and other areas to provide details of the construction and analysis of such models, and illustrate algorithms such as EM and Data Augmentation. The emphasis will be on applying MCMC theory in a practical manner and producing useful results. Each attendee is strongly urged to bring a laptop computer that has a copy of R and WinBugs installed on it. (Both are available free of charge, and can be downloaded from http://www.r-project.org and http://www.mrc-bsu.cam.ac.uk/bugs). A number of examples will be worked out and, time permitting, each attendee will have the opportunity to analyze and interpret the data and models used. Attendees are also encouraged to bring their own data sets for analysis.

We do not assume that attendees have any familiarity with Monte Carlo techniques, or with any Markov chain theory. We assume familiarity with basic theoretical statistical concepts such as densities, distributions, probability and expectations, the Law of Large Numbers, the Central Limit Theorem, and maximum likelihood estimation. Hierarchical models are often analyzed using Bayesian methods, and familiarity with these methods is desirable but not essential, as the basics will be covered. Some necessary background can be gained from the text Statistical Inference by George Casella & Roger Berger (Duxbury 2001, ISBN: 0534243126).

Spatial Statistical Analysis in the GIS Environment
March 13-14: Mary Christman & Linda Young

Spatial statistical methods, now incorporated into Geographic Information System software, permit rapid analysis and subsequent mapping of statistical quantities. A variety of interesting applications are used to illustrate how the integration of spatial statistics and the display capabilities of GIS enhance understanding of data and interpretation of the maps.

After a basic introduction to spatial autocorrelation and exploratory spatial data analysis, we explore techniques of variogram estimation and ordinary, universal and indicator kriging. Mapping of predictions, standard errors, probabilities and quantiles are also covered.

Each attendee is strongly urged to bring a laptop computer with a copy of ArcGIS Versions 8 or 9 with the Geostatistical Analyst extension installed. A number of examples will be worked out and, time permitting, each attendee will have the opportunity to analyze and interpret the data and models used.

We do not assume any familiarity with spatial statistics. We do assume familiarity with basic statistical concepts such as probability distributions, parameter estimates, standard errors, confidence intervals, and hypothesis testing at the level of Bain, Engelhardt’s Introduction to Probability & Mathematical Analysis. Familiarity with regression methods is desirable but not essential, as the basics will be covered. We do assume participants have basic GIS skills, such as use of ArcMap.

Copies of all course slides and example output discussed will be provided.

Statistical Analysis of Case-Control & Cohort Data
March 13-14: Malay Ghosh & Bhramar Mukherjee CANCELED

This short course is designed to train statisticians, epidemiologists and other scientists on how to analyze data collected through case-control and cohort studies.

The first part of the course will present analysis tools for both unmatched and matched case-control data. SAS codes with several real examples will be provided for each topic, using PROC FREQ, PROC LOGISTIC and PROC PHREG. SAS Macros for implementing a matched design will be discussed. Special topics will include discussion of some recent applications of case-control design in genetic epidemiology and the Bayesian alternative. The second part of the course will focus on classical methods of analyzing cohort data starting with a discussion on simple measures of incidence rates, rate ratios and risk ratios, followed by Cox regression, Stratified Cox regression and Poisson regression. Implementation of these methods using SAS will be demonstrated.

The short course emphasizes concepts, interpretation and implementation rather than technical details. However, the attendee is expected to have familiarity with basic principles of statistical inference. Copies of course slides and SAS programs will be provided to the attendees. We will mainly use material from the following texts:

  1. Statistical Methods in Cancer Research: Volume I: The Analysis of Case-control Studies, N.E. Breslow and N.E. Day, International Agency for Research on Cancer, 1980, Oxford University Press. ISBN 92 832 0132 9
  2. Statistical Methods in Cancer Research: Volume II: The Design and Analysis of Cohort Studies, N.E. Breslow and N.E. Day, International Agency for Research on Cancer, 1987, Oxford University Press, ISBN 92 832 01 82 5
  3. Applied Logistic Regression, Hosmer D. W. and Lemeshow, S., Second Edition, Wiley. 2000. ISBN 0 471 35632 8
  4. Categorical Data Analysis, Agresti, A. Second Edition, Wiley, 2002. ISBN 0 471 36093 7.

Core Topics:

  1. Disease Rate, Odds Ratio, Relative Risk
  2. Logistic regression analysis for unmatched case-control data.
  3. Cochran-Mantel-Haenszel test, McNemar’s test for stratified contingency tables and matched pairs data.
  4. Conditional logistic regression for analyzing matched case-control data.
  5. Design and sample size issues in case-control studies.
  6. Cox proportional hazard regression.
  7. Stratified Cox regression.
  8. Poisson Regression

Special Topics:

  1. Bayesian methods for analyzing case-control data.
  2. Use of case-control design in genetic epidemiology studies.

Survival Analysis CANCELED
March 15-16: P.V. Rao

This course will cover statistical methods for analyzing and interpreting clinical trials data. However, the methods introduced in this course can be used in a wide variety of applications such as engineering, sociology and insurance. Starting with the definitions of survival and hazard functions, the course will cover current methods of modeling survival data and making inferences about their associated parameters. The use of SAS procedures, LIFETEST, PHREG and LIFEREG will be illustrated with examples.


Familiarity with interpreting results of analysis of variance and multiple regression analysis will be assumed.

Required text: Modelling Survival Data in Medical Research, Second Edition, by David Collett (Chapman & Hall/CRC, 2003, ISBN 1584883251). If you do not have a copy, please order one from your local bookstore, an internet provider, or on the registration form.

Core Topics:

1. Definitions and examples
2. Preliminary concepts
3. Inferences about survival functions using Kaplan-Meier estimators
4. Comparing survival functions using weighted logrank tests
5. Modeling survival data using proportional hazards (PH) models
6. Checking violations of PH model assumptions
7. Modeling survival data using accelerated failure time models
8. Analyzing recurrence times (time permitting)

Statistical Financial Modeling CANCELED
March 3-4: Alex Trindade

Data collected over time usually exhibit a high degree of correlation. Its analysis therefore requires a different approach from the traditional linear statistical model framework with i.i.d. errors. This short course will provide a "hands-on" introduction to modern applied time series methods, with particular emphasis on techniques suitable for the modeling and forecasting of financial data. The classical ARMA model with its nonstationary and seasonal variants will form the bulk of the course. Specialized methods for finance will be seen to be extensions of this class. Applications will be illustrated with software.

No previous knowledge of time series is required but attendees should be familiar with elementary distributions (normal, t, chi-square) and inferential concepts (confidence intervals, hypothesis tests, p-values). Knowledge of trigonometric functions, matrix notation and exposure to multiple linear regression is also necessary. Ideally , attendees will have taken a two semester undergraduate mathematical statistics sequence, such as might be covered in for example Mathematical Statistics with Applications (Duxbury 2002, by Wackerly, Mendenhall, and Scheaffer).

Required text: Introduction to Time Series and Forecasting, 2nd Ed., by Brockwell and Davis (Springer 2002, ISBN 0387953515) and accompanying software ITSM2000. If you do not have a copy, please order one from your local bookstore, an internet provider, or on the registration form. Although not a requirement, maximum benefit from the interactive exercises will be derived if attendees bring along a laptop computer outfitted with ITSM2000. Copies of all course slides and example output discussed will be provided.

Core Topics:

1. Trend and seasonality; testing an estimated noise sequence
2. ARMA(p,q) processes, the ACF and PACF, modeling and forecasting with ARMA processes
3. Non-stationary and seasonal time series models, ARIMA and SARIMA models
4. Regression with time series errors
5. Transfer function modeling
6. Forecasting techniques, ARAR and Holt-Winters algorithms
7. Modeling changing volatility: ARCH and GARCH processes
8. Multivariate Methods

Statistical Genetics and Microarrays
March 15-16: Mark Yang

Statistical genetics plays a crucial role in biology and medicine consequently in biostatistics. To really understand how statistics works in this area, one needs to start with some basic facts in biology and biochemistry and the statistical models based on them. This class will start from the basic molecular genetics to the different forms of genetics data that require substantial statistical help, including gene hunting, heritability estimation, microarray and genomics. The emphasis is at the conceptual level but with mathematics details whenever necessary.

Core Topics:

1. Biological and chemical background for Modern Statistical Genetics
2. Linkage Analysis with Qualitative Trait
3. Genetics of Quantitative Trait (QTL)
4. Small Area Gene Mapping by Linkage Disequilibrium
5. Functional mapping and growth curves
6. Statistical Methods in Microarray Analysis
7. Software in Microarray Data Analysis

Survey Design: Planning for Data Collection CANCELED
March 15-16: Yasar Yesilcay

Data collection activities are carried out by many private and public agencies. Often, those charged with collecting data have some knowledge about sample selection and estimation (called sample design), but very limited, if any, knowledge about survey design. Few textbooks cover survey design in detail.

Survey design, the process of planning every phase of the study, is a multistage process requiring repeated review and revision so that the resulting data are obtained in a cost efficient manner and lead to reliable inferences. This process includes all the activities that must be carried out from the beginning of data collection, including sample design, to the final stage of the distribution of the final report.

Core Topics:

1. The careful and detailed specification of the objectives of the survey
2. specification of the population of interest (differentiating between the concepts of ideal population, desired
    population, target population and the sampled population)
3. Specification and definition of variables of interest, methods of observation, desired accuracy and precision
4. Specification of the available and needed resources
5. Survey Design (Specification of the details selection of the sample and the estimation methods)
6. Organization of the field work for data collection
7. Planning the data summary and analyses
8. Planning and implementing a pilot study (or pretest)
9. Revising all of the above in the light of the pretest
10. Rechecking the consistency of all the above steps
11. The actual implementation of sample selection, data collecting, analyses and inferences
12. Dissemination of the results

The course will also cover various techniques of sample selection, data collection and estimation.

Some knowledge of elementary statistics (at the level of Moore and McCabe, Introduction to the Practice of Statistics) is expected of the participants. Some knowledge of sampling, estimation and inference is desirable

Required text: Designing Surveys: A Guide to Decisions and Procedures by Czara & Blair. (Pine Forge Press, 2nd Edition, ISBN: 0761927468). If you do not have a copy, please order one from your local bookstore, an internet provider, or on the registration form.

GENERAL INFORMATION

Fee Structure

Statistics Short Courses include expert instruction with printed detailed course notes, continental breakfast, lunch, morning and afternoon refreshments breaks and evening receptions. One 2-day course is $1,100 and the cost for 4 days of courses is $2,000. Required textbooks may be included at an additional cost. http://www.register123.com/event/profile/form/index.cfm?PKformID=0x144553a77a
Lodging

A block of rooms has been reserved at The Florida Mall Hotel, at the Florida Mall, Orlando Florida. The special room rate is $109 single or double, plus taxes. The Florida Mall Hotel guests, who stay in the 11-story high-rise hotel, enjoy free in-room high-speed internet service, LCD TVs, an outdoor pool, a whirlpool and an exercise facility. The Hotel is connected to Florida Mall which has over 200 specialty stores, department stores and 30 eateries.

The Florida Mall Hotel is an ideal meeting headquarters within easy reach of all the excitement and attractions that Central Florida has to offer. Situated in south Orlando at Sand Lake Road and U.S. 441, The Florida Mall Hotel is six miles west of Orlando International Airport and seven miles from downtown. The hotel offers free parking, on-site car rental, attractions tickets and transportation is available to: Walt Disney World, Universal Studios Florida, Sea World, Church Street Station and the Kennedy Space Center.

Make reservations by calling 1-800-588-4656 [FAX: (407) 855-1585] or mail to:
The Florida Mall Hotel
1500 Sand Lake Road
Orlando, Florida 32809

In order to receive the special room rate ($109, single or double) you (or your agent) must clearly indicate that you are attending the University Of Florida Statistics Week Of Short Courses. The Hotel requires a one night deposit to confirm a reservation. Deadline for hotel reservations is February 6, 2006. After this date, rooms will be on a space and rate availability basis.

NOTE
Please make your hotel and travel reservations early. The Orlando area is a major conference and tourist area, and March is a very popular month to visit Florida. Airlines, auto rentals and hotels fill up quickly at this time of the year.
Refunds

Full refunds, less a $75 processing fee, will be made if a written request is postmarked by February 6, 2006. The University of Florida reserves the right cancel individual courses or this event. In the event the University of Florida cancels a course or this event, fees paid for the courses and texts will be fully refunded. However, the University of Florida will not be responsible for any other expenses paid by participants, including, but not limited to, travel and hotel costs, regardless of the date of cancellation.

In compliance with the ADA act, participants with special needs can be reasonably accommodated by contacting Carol Rozear in the Department of Statistics at the University of Florida before January 28, 2005. She can be reached by phone at (352) 392-1941 ext. 207, by fax at (352) 392-9734, or by calling 1-800-955-8771 (TDD).