Estimation of relative potency in multivariate parallel line bioassays

Abstract

Finney (1978) has proposed using a composite response to estimate the relative potency of a test treatment with respect to a standard treatment, in the case of a multivariate parallel line bioassay. However, he assumes the coefficients in the composite response to be constants though they are estimated from the data itself. We provide two different methods of obtaining confidence intervals for this relative potency by using the exact variance. A method based on the use of delta technique is also proposed.

Keywords:

• Relative potency
• multivariate parallel line bioassays
• anderson's classification statistic
• composite response