The experience of race in the United States is shaped by both self-identification and ascription. One aspect reflects personal history, ancestry, and socialization while the other draws largely on appearance. Yet, most data collection efforts treat the two aspects of race as interchangeable, assuming that the relationship between each and an individual's life chances will be the same. This study demonstrates that incorporating racial self-identification and other-classification in analyses of inequality reveals more complex patterns of advantage and disadvantage than can be seen using standard methods. These findings have implications for how racial data should be collected and suggest new directions for studying racial inequality in the United States and around the world.