Empirical legal scholars have traditionally modeled trial court judicial opinion writing by assuming that judges act rationally, seeking to maximize their influence by writing opinions in politically important cases. To test such views, we collected data from a thousand cases in four different jurisdictions. We recorded information about every judicial action over each case’s life, ranging from the demographic characteristics, workload, and experience of the writing judge; to information about the case, including its jurisdictional basis, complexity, attorney characteristics, and motivating legal theory; to information about the individual orders themselves, including the relevant procedural posture and the winning party. Our data reveal opinions to be rare events in the litigation process: only 3% of all orders, and only 17% of orders applying facts to law, are fully reasoned. Using a hierarchical linear model, we conclude that judges do not write opinions to curry favor with the public or with powerful audiences, nor do they write more when they are less experienced, seeking to advance their careers, or in more interesting case types. Instead, opinion writing is significantly affected by procedure: we predict that judges are three times more likely to write an opinion on a summary judgment motion than a discovery motion, all else held equal. Judges similarly write more in cases that are later appealed, and in commercial cases, while writing less in tort and prisoner cases. Finally, jurisdictional culture is very important. These findings challenge the conventional wisdom and suggest the need for further research on the behavioral aspects of opinion writing.