All materials undergo some degree of compaction or expansion when exposed to radiation. Multi-component materials are more susceptible to this effect than single-component materials (e.g., fused silica). Nonetheless, the much lower expansion characteristics of multi-component materials -- such as the ultra-low expansion glass-ceramic Zerodur -- preserves the attractiveness of such materials for applications that require superior dimensional stability. In this study, we present a reanalysis of experimental data describing the compaction effects of electron radiation on Zerodur. These data include high-dose, high dose-rate bulk density measurements as well as lower-dose, interferometrically-measured surface figure changes. We show that previous attempts to deduce linear compaction from figure changes are in error and in fact have precluded earlier attempts to predict radiation effects for an arbitrary optical geometry. By interpreting surface figure measurements in light of a more relevant physical model -- a simplified bimetal equation -- we are able for the first time to accurately predict expected deformation as a function of prescribed dose for both laboratory and space-based experiments. Moreover, we show that a real discrepancy exists between compaction estimates from bulk density experiments and those from surface figure measurements.