About our Workshops
Click on the name of a workshop for more information.
Introduction to Factor Analysis: Factor Analysis is a widely used multivariate technique. It is a data reduction technique that examines the underlying relations that exist among a set of variables. In doing so it assumes that a small number of unobserved variables, called factors, are responsible for the correlation among a large number of observed variables. This one hour workshop is intended as an introduction to factor analysis.
Introduction to Multilevel Models: Multilevel models (also referred to as mixed models or hierarchical models) are used when observations are not independent. Data clustering occurs with many experimental designs (e.g., split plot designs), with social science data collected simultaneously on different units of analysis (e.g., households and their individual members), with measurements taken on the same units over several time periods (longitudinal studies) and with spatial data. The purpose of this workshop is to introduce the concepts that form the basis of these models, the underlying statistical model and the estimation techniques that are used.
Refresher on Interpreting Linear Regression Parameters:: This workshop is intended for researchers who have taken at least one statistics class covering linear regression. Participants will review in a hands-on fashion how to make sense of their linear regression output, with an emphasis on the meaning of the regression parameters. Amongst others, topics covered will be interpretation of parameters of continuous variables , categorical variables, interaction and polynomial terms.
Basic Data Analysis: This workshop is designed to teach researchers how to organize and begin analyzing data. Some of the topics include:
- Organizing and entering data
- Cleaning and manipulating data
- Visualizing data
- Basic statistical analysis
Statistical Analysis with Missing Data: This is a two-hour workshop to introduce new methods for dealing with missing data and the software available to implement them. We will discuss the mechanisms that cause missing data, advantages and disadvantages of commonly-used approaches to dealing with missing data, some new and better approaches, and how to implement them with available statitstical software.
Introduction to Logistic Regression Analysis: This workshop is designed to teach researchers how to use logistic regression. After attendance at the workshop, you should be able to:
- Decide whether logistic regression is appropriate for your data
- Interpret results from a logistic regression analysis
- Test which logistic regression model best fits the data
- Check whether the assumptions are being met
- Understand the terminology used in software manuals so that you can choose the appropriate options for a particular analysis.
Analysis of Longitudinal Data: This is a week-long workshop. Its purpose is to provide participants with an overview of issues related to the design and analysis of longitudinal data, that is, data collected on individuals over time. These data occur frequently in research in many fields, and often present statistical challenges that cannot be addressed with the knowledge gained from typical introductory statistics courses.
Introduction to Sampling Design and Analysis of Complex Surveys: Large-scale surveys are increasingly made available to and used by researchers in many fields for secondary data analysis. These surveys have often a complex design which might include features such as stratification, multi-stage sampling and unequal sampling probabilities. The statistical analysis of such a survey will yield incorrect results if the design is not properly taken into consideration.
The purpose of this workshop is to offer participants an overview of complex survey design and the aspects that need to be taken into consideration when performing a statistical analysis. Examples will be presented and practical aspects such as available software will be discussed.
The workshop is intended for participants who have the equivalent of at least two semesters of statistics.
Designing Experiments: This workshop is offered by Emeritus Professor W. Federer from the Department of Biological and Computational Statistics. Dr. Federer is an expert in design of experiments and author of several books on the subject including co-author of "Variations of Split Plot and Split Block Experiment" recently released.
The principles of designing experiments will be discussed. Reasons for adhering to these principles will be given. The various units related to conducting experiments will be presented. Methods for controlling extraneous variation in experimental material will be given. With this background, it is a simple procedure to obtain a randomized form for a variety of experiment designs using the software toolkit Gendex. An experimenter should not use valuable time to obtain a randomized for of a plan for an experiment in the 21st century. Such a toolkit as Gendex allows an experimenter to obtain a variety of plans for consideration.
Getting Started with Data Analysis
Although this workshop is open to all researchers, it is especially intended to help honors students getting started with the analysis of quantitative data. The focus of the workshop is on practical data analysis and topics covered will include the organization, summarization, basic analysis and presentation of data. Participants are encouraged to bring their own data if possible because a statistical consultant will be available after the workshop to answer their specific questions.
Introduction to Classification and Regression Trees
The classification and regression tree methods are nonparametric analysis techniques for partitioning populations into meaningful subgroups. When the response variable is categorical, the technique is referred to as Classification Trees. When the response variable is continuous, the method is referred to as Regression Trees. The purpose of this workshop is to introduce researchers to the concepts of tree-based modeling techniques, including the interpretation of results and software implementation.
Introduction to Functional Data Analysis
This workshop is offered by Dr. Giles Hooker, a faculty member in the departments of Biological Statistics and Computational Biology and Statistical Sciences.Functional data analysis is the study of high frequency data that may be thought of as describing a collection of smooth curves. Classical examples of these data include data obtained from medical or environmental monitoring devices and optical or mechanical tracking devices. Functional data analysis can also be appropriate whenever an underlying smooth process is thought to be the quantity of interest in noisy and less frequent data.
Using several examples we will illustrate ways to describe the variation among a group of curves, to describe differences between groups of curves and to understand the effect of one set of curves on another. We also discuss some techniques that are unique to functional data: curve alignment and the analysis of rates of changes or derivatives. This workshop will serve as an introduction to a class that will be taught in the fall. The workshop will consist of a lecture and a computer lab using Matlab software. No previous knowledge of Matlab will be assumed but the equivalent of two semesters of statistics and some previous experience with linear regression would be needed to benefit from this workshop.
Introduction Logistic Regression for Responses with More than Two Categories
The binary logistic regression is a model that is used when the dependent variable is binary. When the dependent variable is a nominal variable with more than two categories, a multinomial logistic regression can be used. When the response categories of the dependent variable are ordered, ordinal logit models can be used. The workshop is intended for people who would like to learn about multicategory logistic regression analysis in their research. No previous experience with or knowledge is necessary, although knowledge of binary logistic regression is helpful.
The emphasis will be on deciding when multicategory regression is appropriate, the terminology used, and the interpretation of the results. Output from statistical software will be used in explaining how to interpret results.