Current Topics in Statistics for Applied Researchers
July 17 to 20, 2006
School of Nursing,
Oregon Health & Science University

Print the conference information and registration form.

This workshop is designed for researchers interested in updating and improving their statistical knowledge in order to strengthen the design and analyses of their research. Our goal is not to make you an expert in a given statistical technique but to provide you with an in-depth understanding of certain techniques so that you have greater confidence in interpreting statistical results and more meaningful discussions with your statistical consultants.

The workshop will cover four topics in-depth. For each of these topics there will be an instructional presentation, practice interpreting SPSS output, discussion of how to report the results using examples from the literature, and a presentation and discussion with a researcher who has used the approach. This year’s topics are:

Strengthening Your Research with Mediators and Moderators

Often the models on which our research is based go beyond the simple conceptualization of variables as independent and dependent. Mediator models allow us to hypothesize the process by which the independent variable effects the dependent variable and moderator models allow us to hypothesize the conditions under which the independent variable effects the dependent variable. In this session, we will work on developing hypotheses about mediators and moderators, examine regression and structural equation modeling approaches to testing these hypotheses, and practice interpreting and reporting results.

Issues in Longitudinal Design and Analysis

This session will discuss some of the important issues in designing longitudinal studies and decisions that will determine the appropriate ways to analyze such data. For instance, how many time points are needed to capture change? Do I need to worry about measurement invariance? When should a change score be computed? What is the difference between repeated measures ANOVA and multilevel modeling? What role do time-varying covariates play?

Factor Analysis

In this session we will address a variety of important issues in conducting a factor analysis with applications to instrument development; exploratory versus confirmatory approach, number of factors to extract, procedures for extraction and rotation of factors, discarding items, and creating summative scales. It will also help you become familiar with many specialized terms such as eigenvalues, factor loadings, factor pattern matrix, factor scores, principal components analysis, scree plot, communality, specificity, total variance, orthogonal versus oblique rotation, and marker items.

Cluster Analysis

Cluster analysis allows the researcher to find groups of individuals with similar patterns across an array of variables. This technique is becoming popular in the symptom research as it does not assume that all individuals experience the same set of symptoms. We will compare cluster analysis to factor analysis, discuss the different approaches to conducting and interpreting clusters, explore analyses to validate the clusters, and using the clusters in additional analyses.

In addition, each participant will have time to consult one-on-one on the design and analysis for their research with the workshop faculty.

Workshop Faculty

• Nancy A. Perrin, Ph.D. is Professor and Director of the Statistical Core at the OHSU School of Nursing and a Senior Investigator at the Kaiser Permanente Center for Health Research with a PhD in quantitative psychology. Her area of expertise is the application of multivariate statistical techniques to health outcomes. She has been co-investigator on numerous NIH funded studies utilizing a variety of research designs ranging from community-based participatory to quasi experimental to clinical trials and statistical techniques including cluster analysis, structural equation modeling, time series, regression, multilevel analysis, and multivariate analyses. She also has extensive experience in psychometrics.

• George J. Knafl, Ph.D. is Professor and Senior Scientist at the OHSU School of Nursing. He has a PhD in mathematics specializing in statistics. He is principal investigator for the NIAID-funded project Modeling HIV Subjects' Electronic Monitoring Device Data. He has extensive statistical consulting experience and his research interests include adaptive methods for statistical model selection, modeling and analysis of electronic monitoring data, modeling and analysis of longitudinal family data, statistical evaluation of survey instruments, and applications of statistics to health science research. Visit George J. Knafl, Ph.D. faculty Web page.

• Karen S. Lyons, Ph.D. is an Assistant Professor at the OHSU School of Nursing who earned her PhD in Human Development and Family Studies. Dr. Lyons’ research has focused on family experiences of chronic illness with an emphasis on psychological, relational and methodological issues. Her methodological expertise focuses on longitudinal analyses at both the individual and dyad-level, using multilevel modeling. She has published on roles of optimism and pessimism in predicting 10-year change in physical function and depression, the effects of changing physical health on relationship quality over time, frail elder-family member incongruence regarding physical function, and using multilevel modeling in family research. Visit Karen S. Lyons, Ph.D. faculty Web page.

Location

School of Nursing, Oregon Health & Science University, Portland Oregon. Visit the OHSU Parking Office Web site for maps and directions.

Portland has been proclaimed as North America's "Best Big City," according to Money magazine. Portland has unmatched natural beauty, a bustling local scene, sumptuous dining and welcoming accommodations -- all effortlessly accessed through a light-rail system. Visit the Portland Visitors Web site.

Accommodations

Doubletree Hotel – Lloyd Center
1000 NE Multnomah, Portland, Oregon 97232
503-281-6111 x4153
OHSU Special Rate: $94.00 per night (includes transportation to campus)

Print the conference information and registration form.