Big data are increasingly seen as a tool for assessing public sentiment without the intrusion or cost of a traditional sample survey. Social networking sites in particular provide masses of information about individuals that can be analyzed at little cost. Yet the value of social networking data depends on tools that can translate between the information provided and societal parameters. The current project assesses how the behavior of posting on social media sites relates to attitudes measures acquired from survey respondents. Mapping social media data onto contemporary surveys helps us understand the processes generating both data types and reveals biases that could potentially undermine conclusions from plague big data analytics.