We have investigated the possibility of calibrating the PMTs of scintillation detectors, using the primary scintillation produced by X-rays to induce single photoelectron response of the PMT. The high-energy tail of this response, can be approximated to an exponential function, under some conditions. In these cases, it is possible to determine the average gain for each PMT biasing voltage from the inverse of the exponent of the exponential fit to the tail, which can be done even if the background and/or noise cover-up most of the distribution. We have compared our results with those obtained by the commonly used single electron response (SER) method, which uses a LED to induce a single photoelectron response of the PMT and determines the peak position of such response, relative to the pedestal peak (the electronic noise peak, which corresponds to 0 photoelectrons). The results of the exponential fit method agree with those obtained by the SER method when the average number of photoelectrons reaching the first dynode per light/scintillation pulse is around 1.0. The SER method has higher precision, while the exponential fit method has the advantage of being useful in situations where the PMT is already in situ, being difficult or even impossible to apply the SER method, e.g. in sealed scintillator/PMT devices.