“Big data” alone cannot contribute to science. Strong theory matters as well. And qualitative methods must play a central role, as we argue in our recent paper. This study examines the extent to which surveys and qualitative methods can improve the ability of large-scale data research to arrive at reliable discoveries in the social sciences. Taking seriously the data quality problems that accompany large-scale data, including misclassification, outdated records, unreliable sourcing, and ambiguity, we examine the role of field research at the level of design, interpretation, collection, and theory building with large-scale data.