Research disrupts the social world, often by making respondents aware that they are being observed or by instigating reflection upon particular aspects of life via the very act of asking questions. Building on insights from the first Hawthorne studies, reflexive ethnographers, and methodologists concerned with panel conditioning, we draw on six years of research within a community in southern Malawi to introduce a conceptual framework for theorizing disruption in observational research. We present a series of poignant-yet-typical tales from the field and two additional tools-the refresher-sample-as-comparison and study-focused ethnography-for measuring disruption empirically in a longitudinal study. We find evidence of study effects in many domains of life that relate directly to our scope of inquiry (i.e., union formation, fertility) and in some that extend beyond it (i.e., health). Moreover, some study effects were already known and discussed in the broader community, which was also affected by our research in unintended ways. We conclude that the assumption of non-interactivity in observational research is shaky at best, urging data-gatherers and users to think more seriously about the role of disruption in their work.