This study develops a model of different types of political regime changes and their effect on life integrity violations. The data covers 147 countries from 1977-1993. Basic bivariate analyses and multivariate pooled cross-sectional time series analyses employing Ordinary Least Squares regression with panel-corrected standard errors are used. The results show that political regime change in general has no effect on state-sponsored violence. Looking at different types of regime changes, the regression analysis indicates that change from democracy to anocracy is positively correlated with levels of repression at the level of p < .001. A change toward democracy from autocracy is negatively related to human rights violations at the level of p < .01, once relevant control variables are considered.