The overall objectives of this study were to 1) evaluate the impact of including crossbred relative records on estimated breeding value selection accuracy for crossbred performance and to 2) estimate the subsequent extent of animal re-ranking and differences in selection decisions. Number born alive (NBA) and litters per sow per year (LSY) records were obtained from a large swine breeding company which maintains a database consisting of nucleus, multiplier, commercial, and rotational cross herds. The dataset included 187,674 litters from 46,749 Landrace (LR); 117,782 litters from 31,665 Large White (LW); and 295,163 litters from 76,925 LR Ã Â LW (F1) females, respectively. Variance components for each breed were calculated in REMLF90 using bivariate models with one purebred trait and the respective F1 trait. Number born alive and LSY crossbred breeding values and accuracies were estimated for purebred sows separately with purebred information only (EBVPLS) and with combined crossbred and purebred information (EBVCCPS) in BLUPF90. Purebred sow contemporary groups with >10 sows each were obtained according to herd weaning schedule from the most recent six months of data from each herd. Sows were ranked within contemporary groups for EBVPLS and EBVCCPS. Spearman rank correlations between EBV rankings and selection decision differences were calculated for each contemporary group, boars used for artificial insemination, and a multiplier herd (HP) with seven commercial herds which contained F1 progeny with > 98% pedigree information. Selection accuracy increases, contemporary group Spearman rank correlations, and percentages of change in selection decisions were evaluated separately for NBA and LSY EBV using mixed model equations. Including crossbred relative records in crossbred EBV calculations improved selection accuracy for boars and sows from all herd types. Sow re-ranking and differences in selection decisions within contemporary groups occurred within each herd type, including nucleus and multiplier herds in which selection decisions have the greatest impact. The greatest impact on sow re-ranking was observed in the HP herd, implying that including crossbred records in EBV calculations could have a greater impact than estimated in the present study if complete pedigree information is provided. Genetic progress in crossbred sows could be improved by including performance records from pedigreed crossbred sows in purebred breeding animal EBV calculations for crossbred performance.