While disaster studies researchers usually view risk as a function of hazard, exposure, and vulnerability, few studies have systematically examined the relationships among the various physical and socioeconomic determinants underlying disasters, and fewer have done so through seismic risk analysis. In the context of the 1999 Chi-Chi earthquake in Taiwan, this study constructs three statistical models to test different determinants that affect disaster fatality at the village level, including seismic hazard, exposure of population and fragile buildings, and demographic and socioeconomic vulnerability. The Poisson regression model is used to estimate the impact of these factors on fatalities. Research results indicate that although all of the determinants have an impact on seismic fatality, some indicators of vulnerability, such as gender ratio, percentages of young and aged population, income and its standard deviation, are the important determinants deteriorating seismic risk. These findings have strong social implications for policy interventions to mitigate such disasters.