This research explores violent and property crime rates in nonmetropolitan counties. It is argued that crime rates are lower in these counties because of higher levels of social integration. We test the hypothesis that predictors of crime from social disorganization theory exert different effects on violent and property crimes at different levels of population change in nonmetropolitan counties. We use a spatial lag regression model to predict the 1989-1991 average violent and property crime rates for these counties, taken from the Uniform Crime Reports (UCR). The results show that a factor-analyzed index of resource disadvantage (poverty rate, income inequality, unemployment, percent female-headed households) has different effects on both violent and property crime at different levels of population change in nonmetropolitan counties. Contrary to expectations, we find that resource disadvantage exerts a greater positive effect on both violent and property crimes in nonmetropolitan counties that lost population between 1980 and 1990. Implications for theory and research are discussed. (Contains 3 tables, 2 figures and 8 footnotes.)