Civil procedure rests on a canon of core values: speedy, private, and inexpensive adjudication. In the name of these values, cases are summarily judged, claims are precluded, relief to parties is denied. But these values are not what litigants care about. Courts and doctrinal scholars have long discussed procedural values without knowing which procedural values matter to actual litigants. This Article fills that void. Using an original dataset based on 1200 mailed surveys to a broad range of litigants and judges, it finds a consensus among surveyed groups for not valuing highly speed, cost, and privacy. Beyond this consensus, this Article also reveals significant conflict between groups over which of the remaining procedural values are most important. Federal judges prioritize fairness and participation over all else. Large corporations emphasize the values of accuracy and finality. Pro se litigants stress the importance of accessibility and simplicity. Together, these empirical findings raise pressing normative concerns: whenever we favor one procedural value over another (as we often must), we also favor some litigants over others. Conflict over procedural values is conflict over which litigants to support and which to hinder. This Article is the first in legal scholarship to study procedural values empirically. It does so by introducing to legal scholarship a novel combination of methodological tools (multi-dimensional scaling and circular regressions). These tools are useful in any area of law where there is conflict over ends and values. That is to say everywhere.