MINOS is a long baseline neutrino oscillation experiment that uses two detectors separated by 734 km. The readout systems used for the two detectors are different and have to be independently calibrated. To verify and make a direct comparison of the calibrated response of the two readout systems, test beam data were acquired using a smaller calibration detector. This detector was simultaneously instrumented with both readout systems and exposed to the CERN PS T7 test beam. Differences in the calibrated response of the two systems are shown to arise from differences in response non-linearity, photomultiplier crosstalk, and threshold effects at the few percent level. These differences are reproduced by the Monte Carlo (MC) simulation to better than 1% and a scheme that corrects for these differences by calibrating the MC to match the data in each detector separately is presented. The overall difference in calorimetric response between the two readout systems is shown to be consistent with zero to a precision of 1.3% in data and 0.3% in MC with no significant energy dependence.