The primary objective of this study was to evaluate the validity of using data obtained from slaughtered pigs for farm-level epidemiologic studies of Salmonella. The study involved groups of pigs from five farms. Salmonella isolates were obtained from on-farm samples, and a total of 370 on-farm and an additional 486 isolates from samples collected after commercial slaughter were subsequently tested. Preharvest samples included feces of individual animals from defined groups of nursery and finishing pigs on commercial farms and swabs from trucks. Postslaughter samples were cecal contents and mesenteric lymph node samples. The concordance between Salmonella serovars isolated from on-farm samples and those serovars isolated after slaughter varied widely among farms. Results of paired lymph node and cecal cultures were strongly associated (odds ratio, 7.0), but the agreement between on-farm and postslaughter results at the pig level was poor (κ = 0.34). The results support recent findings that risk of exposure to Salmonella during transport and lairage remains a concern under contemporary industry conditions. The findings further imply that slaughter plant studies based on phenotyping of Salmonella alone (such as serovars) may not reliably indicate the Salmonella status of commercial swine farms.