Sankhya: The Indian Journal of Statistics

2000, Volume 62, Series B, Pt. 1, 119--133

ROBUST PROCEDURES FOR BIOASSAYS AND BIOEQUIVALENCE STUDIES

By

HO KIM, Seoul National University, Korea

and

PRANAB K. SEN, University of North Carolina, Chapel Hill

SUMMARY. In bioassays, inference on relative potency and bioequivalence (of a new preparation relative to a standard one) constitute the main theme of studies. A conditional quantile process based on the k-nearest neighborhood method is proposed for stochastic dose designs. Minimum chi-square type estimators are considered and incorporated in the test for the fundamental assumption of the assay. Bias correction by jackknife methods is discussed. As an illustration, ARIC ultrasound data are analyzed.

AMS (1991) subject classification. 62G99, 62P10.

Key words and phrases. Bias correction; dose-response regression; conditional quantile process; relative potency; nearest neighbor estimator; nonparametrics; robustness.

Full paper (PDF)

This article in Mathematical Reviews