Dieter Rasch

Moti Lal Tiku

This volume contains most of the invited and contributed papers presented at the Conference on Robustness of Statistical Methods and Nonparametric Statistics held in the castle oj'Schwerin, Mai 29 - June 4 1983. This conference was organized by the Mathematical Society of the GDR in cooperation with the Society of Physical and Mathematical Biology of the GDR, the GDR-Region of the International Biometric Society and the Academy of Agricultural Sciences of the GDR. All papers included were thoroughly reviewed by scientist listed under the heading "Editorial Collabora tories·'. Some contributions, we are sorry to report, were not recommended for publi cation by the rf'vif'wers and do not appear in these proceedings. The editors thank the reviewers for their valuable comments and suggestions. The conference was organizf'd bv a Programme Committee, its chairman was Prof. Dr. Dieter Rasch (Research Centre of Animal Production, Dummerstorf-Rostock). The members of the Programme Committee were Prof. Dr., Johannes Adam (Martin-Luther-University Halle) Prof. Dr. Heinz Ahrens (Academy of Sciences of the GDR, Berlin) Doz. Dr. Jana Jureckova (Charles University Praha) Prof. Dr. Moti Lal Tiku (McMaster University, Hamilton, Ontario) The aim of the conference was to discuss several aspects of robustness but mainly to present new results regarding the robustness of classical statistical methods especially tests, confidence estimations, and selection procedures, and to compare their perfor mance with nonparametric procedures. Robustness in this sens~ is understood as intensivity against. violation of the normal assumption.

Robustness of Statistical Methods and Nonparametric Statistics

In Statistical Theory And Practice, A Certain Distribution Is Usually Assumed And Then Optimal Solutions Sought. Since Deviations From An Assumed Distribution Are Very Common, One Cannot Feel Comfortable With Assuming A Particular Distribution And Believing It To Be Exactly Correct. That Brings The Robustness Issue In Focus. In This Book, We Have Given Statistical Procedures Which Are Robust To Plausible Deviations From An Assumed Mode. The Method Of Modified Maximum Likelihood Estimation Is Used In Formulating These Procedures. The Modified Maximum Likelihood Estimators Are Explicit Functions Of Sample Observations And Are Easy To Compute. They Are Asymptotically Fully Efficient And Are As Efficient As The Maximum Likelihood Estimators For Small Sample Sizes. The Maximum Likelihood Estimators Have Computational Problems And Are, Therefore, Elusive. A Broad Range Of Topics Are Covered In This Book. Solutions Are Given Which Are Easy To Implement And Are Efficient. The Solutions Are Also Robust To Data Anomalies: Outliers, Inliers, Mixtures And Data Contaminations. Numerous Real Life Applications Of The Methodology Are Given.

Moti Lal Tiku

About

Books

Join the club