The relatively new technique of alternating differential scanning calorimetry (ADSC) relies upon the measurement of a phase angle between the heating rate and heat flow in order to separate the real and imaginary parts of the complex heat capacity. A significant problem in this context is the separation of an 'instrumental' phase angle, arising from the heat transfer between instrument and sample, from any 'sample' phase angle which may result from a transition. In this work, the theoretically anticipated dependence of phase angle on the frequency of modulation has been investigated experimentally for polycarbonate in dynamic scans through the glass transition region using the Mettler-Toledo ADSC. There is a good agreement between the experimental results and the predictions of simple models for both heat transfer and relaxation in the transition region. This provides support for the procedure originally suggested by Schick and co-workers for the correction of the phase angle to allow for the instrumental effects.