Non-parametric Statistical Tools for the Social Sciences
Sometimes researchers do not a priori have a specific model or parametric assumption in mind when modeling social science data. Nonparametric approaches are useful in these circumstances for visualization as well as for suggesting underlying assumptions for subsequent parametric models. This talk reviews the basics of nonparametric data analysis from bivariate smoothing, kernels, and splines, to full regression-style generalized additive models. Example data will be drawn from the social and public health sciences.
Workshop is limited to 25 participants
[[{"fid":"75","view_mode":"teaser","fields":{"format":"teaser","field_file_image_alt_text[und][0][value]":"Dr Jeff Gill","field_file_image_title_text[und][0][value]":"Dr. Jeff Gill"},"type":"media","link_text":null,"attributes":{"alt":"Dr Jeff Gill","title":"Dr. Jeff Gill","style":"line-height: 1.44; background-color: transparent;","class":"panopoly-image-quarter media-element file-teaser"}}]]
Presenter: Dr. Jefferson Gill, Professor, Department of Political Science, Professor, Division of Biostatistics, Professor, Department of Surgery
Washington University
Jeff Gill. Professor, Washington University. (BA UCLA, MBA Georgetown, Ph.D. American University, Post-Doc Harvard). Major areas of research and interest are [Methodology and Statistics] Bayesian approaches, Markov chain Monte Carlo, queueing theory, nonparametrics, missing data, generalized linear model theory, model selection, circular data, and general problems in statistical computing; [Epidemiology] mental health outcomes for children exposed to war, foot-and-mouth disease, containment policy,and measurement/data issues; [Medicine] pediatric traumatic brain injury, linkages between obesity and cancer (including human energetics and mouse models), models of Warfarin dosage, psychiatric trauma, physiological effects of stress; [Political Science] voting, terrorism, Scottish politics, expert elicitation, bureaucracy. He is the author of "Essential Mathematics for Political and Social Research," with Cambridge University Press. and is the author of five other books including the forthcoming second edition of "Bayesian Methods for the Social and Behavioral Sciences"(Chapman & Hall/CRC), which is the leading Bayesian text for these disciplines. His journal work has appeared in the Lancet-Neurology, Journal of Politics, Electoral Studies, Statistical Science, Political Research Quarterly, Quarterly Journal of Political Science, Sociological Methods and Research, Public Administration Review, Journal of Public Administration Research and Theory, Annals of Statistics, JASA, Journal of Statistical Software, Political Analysis, and elsewhere.
Date: Friday, January 27, 2017
Time: 9:00am - 4:00pm - Lunch will be included
Location: Cowden Building, Ostrom Lab, Room 124
Workshop Fee: $0 (lunch included)
Register HERE