A systematic statistical model has been used to calibrate the turbulence saturation coefficients in a theory based tokamak transport model. Confidence contours are used to assess the accuracy of calibration. From the confidence contours, it is seen that the ballooning branch coefficient of this model is empirically more uncertain than that for the drift wave branch. Sensitivity studies show that the estimators obtained here are relatively insensitive to the level of prior knowledge assumed about the accuracy of theory and experiment, but the confidence regions shrink with increasing prior knowledge. The calibrated transport model is used to predict the profiles of electron density and of electron temperature and ion temperature in a large tokamak plasma. This demonstrates the potential utility of the calibrated transport model