A round robin test was performed to determine the reliability of value
s for the glass transition temperature T-g as determined by DTA on pol
ymers. Ten different instruments were involved. The test material was
high molecular weight polystyrene. Values for T, (midpoint) were repor
ted in the range 107 degrees C+/-2 K. The respective heat flow curves
differed considerably in shape. In the literature a T-g of 100 degrees
C is often given for polystyrene. The discrepancy between this value
and the value of 107 degrees C found in the round robin test is due to
three differences: the thermal history of the sample, the evaluation
of the heat flow curves, and the effect of finite sample size.