Paul Rosenbaum

Education

PhD, Harvard University, 1980
AM, Harvard University, 1978
BA, Hampshire College, 1977

Career and Recent Professional Awards

Fellow, American Statistical Association, 1992
Fellow, Center for Advanced Study in the Behavorial Sciences, Stanford, 20002001
George W. Snedecor Award from the Committee of Presidents of Statistical Societies, 2003

Academic Positions Held

Wharton : 1986present. (named Robert G. Putzel Professor, 2001; Robert B. Egelston Term Professor of Statistics, 199192; Joseph Wharton Term Associate Professor and Professor of Statistics, 198691)

Previous appointment : University of Wisconsin, Madison

Other Positions

Senior Research Scientist, Research Statistics Group, Educational Testing Service, 1986
Research Scientist, Research Statistics Group, Educational Testing Service, 198386
Statistician, Division of Statistics and Applied Mathematics, Office of Radiation Programs, U.S. Environmental Protection Agency, 198081

Professional Leadership

Member, Committee on National Statistics, National Research Council, 199699
Member, Committee on Data and Research for Policy on Illegal Drugs, National Research Council, 19982000
Member, Advisory Board of the Measurement, Methodology and Statistics Program of the U.S. National Science Foundation, 19992001

For more information, go to My Personal Page

Paul R. Rosenbaum (2016), The crosscut statistic and its sensitivity to bias in observational studies with ordered doses of treatment, Biometrics, to appear.

J.H. Silber, Paul R. Rosenbaum, Matthew McHugh, JM Ludwig, Herbert L. Smith, BA Niknam, Orit EvenShoshan, LA Fleisher, RR Kelz, Linda K. Aiken (2016), Comparison of the value of better nursing work environments across different levels of patient risk, JAMA Surgery, 151. 10.1001/jamasurg.2015.4908

Dylan Small and Paul R. Rosenbaum (2016), Constructed second control groups and attenuation of unmeasured biases, Journal of the American Statistical Association, 111, pp. 11571167. 10.1080/01621459.2015.1076342

Dylan Small and Paul R. Rosenbaum (2016), An exact test of fit for the Gaussian linear model using optimal nonbipartite matching, Technometrics, in press.

Jeffrey H. Silber, Ville A. Satopaa, Nabanita Mukherjee, Veronika Rockova, Wei Wang, Alexander S. Hill, Orit EvenShoshan, Paul R. Rosenbaum, Edward I. George (2016), Improving Medicare's Hospital Compare Mortality Model, Health Services Research, 51 (52), pp. 12291247.

J.H. Silber, Paul R. Rosenbaum, R.N. Ross, JM Ludwig, W Wang, BA Niknam, Alexander Hill, Orit EvenShoshan, RR Kelz, Lee Fleisher (2016), Indirect Standardization Matching: Assessing Specific Advantage and Risk Synergy, Health Services Research , to appear.

Paul R. Rosenbaum (2015), Some counterclaims undermine themselves in observational studies, Journal of the American Statistical Association, 110.

Jesse Hsu, JR Zubizarreta, Paul R. Rosenbaum, Dylan Small (2015), Strong control of the familywise error rate in observational studies that discover effect modification by exploratory methods, Biometrika, 102, pp. 767782.

Edward I. George, Veronika Rockova, Paul R. Rosenbaum, Ville Satopää, Jeffrey H. Silber (Under Revision), Hospital Mortality Rate Estimation for Public Reporting. Journal of the American Statistical Association, Applications.

Abstract:   For the estimation of many parameters, borrowing statistical strength across related observations via hierarchical modeling has been one of the most powerful statistical developments over the past 60 years. The phenomenon is especially evident in classical frameworks where dominating ensemble information shrinks noisy parameter estimates to an overall mean. This is exactly what happened with Medicare's Hospital Compare random effects model, which asserted that 99.5% of hospital mortality rates for acute myocardial infarction (AMI) were ''no different than the U.S. national rate''. In this talk we shall see that their conclusions stemmed from seemingly innocuous, though controversial, assumptions about ensemble means and variances that were at odds with the data. As an alternative, we propose hierarchical random effects models with flexible prior structure that emancipate the means and variances and yield dramatically different conclusions. The superior calibration of our models is demonstrated with comparisons based on predictive Bayes factors and predictive matched samples. Finally, direct rather than indirect standardization is seen to be superior for public reporting.  

Paul R. Rosenbaum, J.H. Silber, RR Kelz, DJ Gaskin, JM Ludwig, R.N. Ross, B. Nikram, Alexander Hill, Min Wang, Orit EvenShoshan, Lee Fleisher (2015), Examining Causes of Racial Disparities in General Surgical Mortality: Hospital Quality Versus Patient Risk, Medical Care, 53, pp. 619629.

Past Courses

STAT112 INTRODUCTORY STATISTICS

Further development of the material in STAT 111, in particular the analysis of variance, multiple regression, nonparametric procedures and the analysis of categorical data. Data analysis via statistical packages.

STAT500 APPLIED REG & ANALY VAR

An applied graduate level course in multiple regression and analysis of variance for students who have completed an undergraduate course in basic statistical methods. Emphasis is on practical methods of data analysis and their interpretation. Covers model building, general linear hypothesis, residual analysis, leverage and influence, oneway anova, twoway anova, factorial anova. Primarily for doctoral students in the managerial, behavioral, social and health sciences.

STAT501 INT TO NONP & LOGLIN MOD

An applied graduate level course for students who have completed an undergraduate course in basic statistical methods. Covers two unrelated topics: loglinear and logit models for discrete data and nonparametric methods for nonnormal data. Emphasis is on practical methods of data analysis and their interpretation. Primarily for doctoral students in the managerial, behavioral, social and health sciences. May be taken before STAT 500 with permission of instructor.

STAT924 ADVANCED EXP DESIGN

Factorial designs, confounding, incomplete blocks, fractional factorials, random and mixed models, response surfaces.

STAT933 ANALY CATEGORICAL DATA

Likelihood equations for loglinear models, properties of maximum likelihood estimates, exact and approximate conditional inference, computing algorithms, weighted least squares methods, and conditional independence and loglinear models. Applied topics, including interpretation of loglinear and logit model parameters, smoothing of tables, goodnessoffit, and incomplete contingency tables.

STAT950 QUANT CONSULTING SEM

The seminar offers the opportunity for small teams of MBAs to work on "real life" quantitative consulting projects. These projects are drawn from both business and University sources. The emphasis is on providing a relevant and comprehensible solution to the client's problem. Inclass brainstorming sessions, client presentations and written reports give students the opportunity to test for the existence of an intersection between their quantitative and communication skills.

STAT991 SEM IN ADV APPL OF STAT

This seminar will be taken by doctoral candidates after the completion of most of their coursework. Topics vary from year to year and are chosen from advance probability, statistical inference, robust methods, and decision theory with principal emphasis on applications.

George W. Snedecor Award from COPSS, 2003 Fellow, American Statistical Association, 1992

Knowledge @ Wharton

Grading Hospital Quality with a Level Playing Field, Knowledge @ Wharton 05/20/2014