Jeanette O'Hara Hines|Statistics and Actuarial Science

Director

Statistical Consulting Service

Associate Professor

Department of Statistics and Actuarial Science
University of Waterloo
Waterloo, Ontario N2L 3G1
Canada

(519) 888-4567 Ext. 5535
(519) 746-1875

E-mail:
Office: M3 4203
Webpage for Statistical Consulting: www.stats.uwaterloo.ca/stats_navigation/Consulting/StatConsulting.shtml

Research and Scholarly Interests

Jeanette Professor O'Hara Hines' research focuses on the practical needs of researchers in the biological sciences, who frequently pose challenges with new ways of gathering data, or with new objectives. As an applied statistician concerned with meeting those needs, she routinely examines the reliability of existing methods and of possible alternatives, which allows her to develop new methods, improve the reliability of existing ones, and improve the effectiveness of research in the biological sciences-as some of her recent work suggests.

One of Professor O'Hara Hines' ongoing research projects is the analysis of clustered or longitudinal data with categorical responses, a type of data frequently gathered in the biological and medical areas. She has extended this work to include the added complication for longitudinal data of drop-outs, or cases which disappear from the study. Since such a disappearance is usually related to the purpose of the study rather than simply being a random phenomenon, any valid analysis needs to take these drop-outs into account. The primary intent of this work is to provide researchers in the life sciences and applied statisticians with sound and practical guidance about analyses of clustered or longitudinal data with categorical responses.

Another research project involves the examination of a single sequence of correlated biotelemetric location data for the purpose of estimating the home range/utilization distribution of a radio-tagged animal using a ``kernel density estimator,'' which estimates likely positions between observations. Ecologists have expressed concern that the estimators of home range currently in use do not take into account the sometimes irregular spacings between observations or the differing environments in the study area, with some environments being far more attractive to the animal than others-such as berry-rich meadows rather than stone out-croppings for bears, or rivers rather than fields for fish. Moreover, current methods ignore the (very important) effects of redundancy (sleeping bears tend not to move in their sleep, for example) and changes over time (foraging patterns of bears alter as winter approaches), concerns that the proposed methods do address. Future work, again prompted by the ecologist's need for sound methodology, will further improve on the proposed methods.

Another research area arises from reciprocal transplant experiments, involving pairs of environments. These experiments measure the successes in one environment of offspring with parents native to the other environment; and also the successes in either environment of offspring with one parent native to each environment. In principle, such experiments should provide valuable information about the nature of the adaptations and the genetic effects, but sound and robust statistical methods are needed. Professor O'Hara Hines' contribution involved developing statistical methodology from basic principles for testing various hypotheses about fitness, as opposed to the use of the only technique previously available, an ad hoc adaptation of a regression method of analysis.

Recent Publications

O’Hara Hines, R.J. and W.G.S. Hines. 2010. Indices for covariance mis-specification in longitudinal data analysis with no missing responses and with MAR dropouts. Computational Statistics & Data Analysis, 54, 806-815.
O’Hara Hines, R.J. and W.G.S. Hines. 2007. Covariance miss-specification and the local influence approach in sensitivity analyses of longitudinal data with drop-outs. Computational Statistics & Data Analysis, 51, 5537-5546.
O’Hara Hines, R.J. and W.G.S. Hines. 2006. Letter to the editor concerning ‘An index of local sensitivity to nonignorable drop-out in longitudinal modelling by G. Ma, A.B. Troxel, and D.F. Heitjan in Statistics in Medicine, 2005, 24:2129-2150'. Statistics in Medicine, 25, 3217-3218. (more)

Biography

Jeanette O'HaraHines Professor O'Hara Hines is currently the director of the Statistical Consulting Unit. In this role, she directs the departmental consultant and graduate RAs assigned to the unit and also provides statistical advice to researchers on campus. She has also organized well-received invited sessions on statistical consulting at the Statistical Society of Canada (SSC) conferences.

Professor O'Hara Hines is currently a regional representative on the board of both the SSC and the International Environmetrics Society. In addition to being an accredited professional statistician (PSTAT), she is currently on the SSC's accreditation committee accrediting other statisticians. In the recent past, Professor O'Hara Hines was chair of the SSC's committee for Women in Statistics, and Canadian chair of the Women's Caucus. In these roles she organized invited sessions for female researchers at SSC and ASA conferences.

Professor O'Hara Hines has been an invited speaker at SSC and Joint Statistical Meeting (JSM) conferences. (more)