Working with local researchers, WorldBank staff recently analyzed a random sample of cases filedin the first instance courts of Argentina and Mexico. (Firstinstance or trial courts make the initial rulings on casesbased on both facts and law. Higher instance or appealscourts are often restricted to questions of law and so maynot reconsider the facts of a case.) The Mexico study wasconducted in the Federal District, the largest ofMexico's 32 local and state jurisdictions, and reviewed464 debt collection cases brought under a special procedurethat provides for rapid dispute resolution. In Argentina astratified sample of criminal, civil, and labor cases wasdrawn from trial courts in the national capital, BuenosAires (600 cases), and in the province of Santa Fe (450cases). In both countries the identities of the parties, thenature of the cases, the amounts at issue, the times todisposition, and other data from the case files were codedand analyzed. Aggregate statistics kept by the judiciary andinformation from interviews and focus groups were used tohelp interpret the case file findings. Both studies revealedthat when it comes to the operation of Latin Americanjustice systems, much of what experts "know" isincorrect. Because this conventional wisdom often informsjudicial reform projects, it can encourage changes aimed atsolving nonexistent problems-while ignoring real ones.