A substantial intellectual movement has been growing in the social sciences around the adoption of mechanism-and process-based explanations as complements to variable-based explanations, or even as substitutes for them. But once we have recognized the validity and dignity of studying mechanisms and processes, what is the next step? Recently, both political scientists' and sociologists' discussions have begun to turn away from correlation to mechanism-based approaches to causation. But there is still a widespread assumption that mechanisms are unobservable. We maintain that ways can be developed to observe the presence or absence of mechanisms either directly or indirectly. In this paper, by way of example, we put forward four methods-two direct and two indirect-for measuring mechanisms of contention.