Research
Research Interests
Early in my career, my research focused on nonparametric and robust
statistical methods; especially linear combinations of order statistics
(L-statistics) and robust rank tests. When I learned that these methods
were used to analyze time series data in electrical engineering, we
investigated the potential impact of dependence on the distribution of
these statistics and found that the effect was quite noticeable. Later I
became interested in the measurement of economic inequality, which led
to several papers concerning the Lorenz curve and Gini index. This
naturally led to developing methods for analyzing salary data for
evidence of discrimination and my continuing work on problems arising in
legal applications of statistics.
I was fortunate to have many fine Ph.D. students. As several of them
are in the area of biostatistics, we have conducted research on topics
that are intellectually related to robust methods or techniques used in
economics and law. In genetic epidemiology, the underlying mode of
inheritance is often not known very precisely so we adapted the
efficiency robustness approach to obtain tests with high power over a
class of models. The problem of assessing whether an omitted variable
could change a statistical conclusion arises in case-control studies and
in the analysis of employment data as both applications deal with
observational studies rather than randomized designed ones. As the
issues involved in assessing minority health disparities are quite
similar to those arising in studying economic discrimination, we are
adapting some regression techniques used in EEO cases for the analysis
of data from complex surveys used in studying health inequalities.
Consulting
After my year as a Visiting Advisor to the Office of Statistical Policy in the Office of Management and Budget, I was asked to continue some research related to the estimation of unemployment rates in local areas as a consultant to the Bureau of Labor Statistics. A few years later I received two calls from lawyers, one concerned a consumer fraud case related to pyramid schemes and the other dealt with an equal employment case. Those consulting jobs generated several publications. For about five years I served as a one day a week consultant to the successor office of the Statistical Policy group I visited when I came to Washington. In 1985-86 I received a Guggenheim Fellowship and visited the Univ. of California at Davis. That year was devoted to research and writing. Since then I have sporadically consulted on law cases, which raised statistical issues generating new research. Most recently, a former student asked me to serve as an expert in a securities law case. That work
stimulated us to write several papers and also develop a small R-package, called Lawstat, so others can use the procedures we developed.
Software
lawstat:
an R package for biostatistics, public policy and law
Wallace Hui, Yulia R. Gel, Joseph L. Gastwirth, Weiwen Miao An R software package on statistical tests widely utilized in biostatistics, public policy and law. Along with the well known tests for equality of means and variances, randomness, measures of relative variability etc, the package contains new robust tests of symmetry, omnibus and directional tests of normality, and their graphical counterparts such as Robust QQ plot; a robust trend tests for variances etc. All implemented tests and methods are illustrated by simulations and real-life examples from legal statistics, economics and biostatistics.
Wallace Hui, Yulia R. Gel, Joseph L. Gastwirth, Weiwen Miao An R software package on statistical tests widely utilized in biostatistics, public policy and law. Along with the well known tests for equality of means and variances, randomness, measures of relative variability etc, the package contains new robust tests of symmetry, omnibus and directional tests of normality, and their graphical counterparts such as Robust QQ plot; a robust trend tests for variances etc. All implemented tests and methods are illustrated by simulations and real-life examples from legal statistics, economics and biostatistics.
Related publications:
- Gastwirth, J.L. (2006), A Sixty Million Dollar Statistical Issue Arising in the Interpretation and Calculation of a Measure of Relative Disparity: Zuni Public School District 89 v. U.S. Department of Education, Working Paper.
- Gel, Y., and Joseph L. Gastwirth (2006), A Robust Modification of the Jarque-Bera Test of Normality, Submitted.
- Gel, Y., W. Miao and Joseph L. Gastwirth (2006), Robust Directed Tests of Normality Against Heavy Tailed Alternatives, To appear in the Computational Statistics and Data Analysis, 12 pages.
- Miao, W., Y. Gel and Joseph L. Gastwirth (2006), A New Test of Symmetry about an Unknown Median, Random Walk, Sequential Analysis and Related Topics - A Festschrift in Honor of Yuan-Shih Chow. Agnes Hsiung, Cun-Hui Zhang, and Zhiliang Ying, Eds. World Scientific Publisher, Singapore (forthcoming), 13 pages.
- Gel, Y., W. Miao and Joseph L. Gastwirth (2005), The Importance of Checking the Assumptions Underlying Statistical Analysis: Graphical Methods for Assessing Normality, 46 Jurimetrics J., 3-26.
