A Bayesian approach to multivariate and conditional calibration

Abstract In this paper the Bayesian approach of Hunter and Lamboy (1981) is extended from univariate to multivariate calibration. A Bayesian competitor of the conditional procedure derived by Fox (1989) is also given. Since the exact posterior distribution of the unknown controlled variable ξ cannot be derived in closed form, a Monte Carlo simulation study has been performed. The methods developed are applied to a real problem.