Although the existing literature on immigrant mental health is extensive, major substantive and methodological gaps remain. Substantively, there is little population-based research that focuses on the mental health consequences of migration for Vietnamese Americans. More generally, although a wide range of mental health problems among immigrants has been identified, the potential causal or mediating mechanisms underlying these problems remain elusive. This latter substantive shortcoming is related to a key methodological challenge involving the potentially confounding effects of selection on migration-related outcomes. This article addresses these challenges by employing a "natural experiment" design, involving comparisons among three population-based samples of Vietnamese immigrants, never-leavers, and returnees (N=709). Data were collected in Ho Chi Minh City and in New Orleans between 2003 and 2005. The study investigates the long-term impact of international migration on Vietnamese mental health, and the potential mediating effects of social networks and physical health on these migration-related outcomes. The results reveal both mental health advantages and disadvantages among Vietnamese immigrants relative to the two groups of Vietnamese nationals. Selection can be ruled out for some of these differences, and both social networks and physical health are found to play important explanatory roles.