Within an epidemiological study regarding the correlation between skin pathologies and personal ultraviolet (UV) exposure due to solar radiation, 14 field campaigns using polysulphone (PS) dosemeters were carried out at three different Italian sites (urban, semi-rural and rural) in every season of the year. A polysulphone calibration curve for each field experiment was obtained by measuring the ambient UV dose under almost clear sky conditions and the corresponding change in the PS film absorbance, prior and post exposure. Ambient UV doses were measured by well-calibrated broad-band radiometers and by electronic dosemeters. The dose-response relation was represented by the typical best fit to a third-degree polynomial and it was parameterized by a coefficient multiplying a cubic polynomial function. It was observed that the fit curves differed from each other in the coefficient only. It was assessed that the multiplying coefficient was affected by the solar UV spectrum at the Earth's surface whilst the polynomial factor depended on the photoinduced reaction of the polysulphone film. The mismatch between the polysulphone spectral curve and the CIE erythemal action spectrum was responsible for the variability among polysulphone calibration curves. The variability of the coefficient was related to the total ozone amount and the solar zenith angle. A mathematical explanation of such a parameterization was also discussed.