On experimental design in linear regression models with random regressors

Summary

In a linear regression model with deterministic regressors and also in such a model with random regressors the estimation and designing problem is formulated. Differently to the known designing problem for deterministic regressors the designing problem for random regressors consists in a suitable choosing of expectations of the regressors. In the case of simple linear regression D‐, G‐and I‐optimum designs are robust relative to the considered model types „deterministic regressor” and “G