Single use conservative confidence regions in multivariate controlled calibration

We consider a multivariate linear model for multivariate controlled calibration, and construct some conservative confidence regions, which are nonempty and invariant under nonsingular transformations. The computation of our confidence region is easier compared to some of the existing procedures. We illustrate the results using two examples. The simulation results show the closeness of the coverage probability of our confidence regions to the assumed confidence level.