THE EFFECT OF VARIANCE FUNCTION ESTIMATION ON NONLINEAR CALIBRATION INFERENCE IN IMMUNOASSAY DATA

Often with data from immunoassays, the concentration-response relation ship is nonlinear and intraassay response variance is heterogeneous. E stimation of the standard curve is usually based on a nonlinear hetero scedastic regression model for concentration-response, where variance is modeled as a function of mean response and additional variance para meters. This paper discusses calibration inference for immunoassay dat a which exhibit this nonlinear heteroscedastic mean-variance relations hip. An assessment of the effect of variance function estimation in th ree types of approximate large-sample confidence intervals for unknown concentrations is given by theoretical and empirical investigation an d application to two examples. A major finding is that the accuracy of such calibration intervals depends critically on the nature of respon se variance and the quality with which variance parameters are estimat ed.