This article presents a technical comparison of two methods of selecting juries. The research site was the Long Beach Superior Court in California. LISREL and computer-generated graphics were the analytic methods used. The analysis demonstrates that, if a representative list of prospective jurors is to be obtained, court officials must recognize the effects of racial and socioeconomic factors on residential segregation, and the selection procedure must take these factors into account. Minority and ethnic groups are unequally distributed within the jurisdiction studied. Traditional methods of jury selection, which are based on simple random sampling, are inadequate because they generate unrepresentative juror pools. In contrast, a cluster sampling method with the probability proportionate to size (PPS) incorporates residential segregation and generates a pool of jurors that reflects the cross-section of the community. Two test statistics, chi-square and Z, substantiated that cluster sampling is clearly superior to the currently employed simple random-selection method. Assurance of a cross-sectional representation of potential jurors is congruent with the requirement established by the Federal Jury Selection and Service Act of 1968. At the state level, as in California, the basic notion of this cross-sectionality is also congruent with the California Code of Civil Procedure of 1981.