Prior work from this laboratory has reported anomalous differences in the viscoelastic responses between temperature-jump formed glasses and carbon dioxide pressure-jump or relative humidity formed glasses. In the present work, we investigate the anomalous behaviour further by examining the structural response of an epoxy resin after pre-annealing treatments. In particular, we have measured the volume change of amine-cured diglycidyl ether of bisphenol A after thermal and carbon dioxide pressure (PCO2) treatments. Our results show that contrary to prior interpretations in the literature, a plasticizer quench is different from a temperature hyperquench. Consistent with our prior work, the CO2-formed glass is more stable than the temperature-formed glass in spite of the former having a higher excess volume. Our new results show that the stability persists to above the nominal glass temperature, contrary to what happens in a temperature hyperquench. (C) 2004 Elsevier Ltd. All rights reserved.