There are few quantitative comparisons in the literature between glass transitions (T-g) measured by differential scanning calorimetry (DSC) and by dynamic mechanical analysis (DMA). Also, in the case of DMA, two different operational definitions have been used to obtain the glass transition, namely, the loss modulus (E") and damping (tan delta) peak temperatures. We propose a new DMA definition of T-g and demonstrate that it agrees with DSC T-g measurements within +/-2 degrees C for both thermoplastic polystyrene and thermoset cross-linked acrylic polyurethane films with measurable tan delta peaks. The glass transitions for a single polystyrene standard and several cross-linked acrylic polyurethane films were measured by DSC. Additionally, E" and tan delta peak temperatures were measured by DMA as a function of frequency and temperature. Empirically, it was determined that the average of the E" and tan delta peak temperatures measured at 1 rad/s oscillation frequency corresponds to the glass transition measured by the ASTM E1356 DSC test method.