Working with local researchers, World Bank staff recently analyzed a random sample of cases filed in the first instance courts of Argentina and Mexico. (First instance or trial courts make the initial rulings on cases based on both facts and law. Higher instance or appeals courts are often restricted to questions of law and so may not reconsider the facts of a case.) The Mexico study was conducted in the Federal District, the largest of Mexico's 32 local and state jurisdictions, and reviewed 464 debt collection cases brought under a special procedure that provides for rapid dispute resolution. In Argentina a stratified sample of criminal, civil, and labor cases was drawn from trial courts in the national capital, Buenos Aires (600 cases), and in the province of Santa Fe (450 cases). In both countries the identities of the parties, the nature of the cases, the amounts at issue, the times to disposition, and other data from the case files were coded and analyzed. Aggregate statistics kept by the judiciary and information from interviews and focus groups were used to help interpret the case file findings. Both studies revealed that when it comes to the operation of Latin American justice systems, much of what experts "know" is incorrect. Because this conventional wisdom often informs judicial reform projects, it can encourage changes aimed at solving nonexistent problems-while ignoring real ones.