Data from four closely related Line 1 Hereford herds were used to estimate variance components and predict EPD for birth weight (BWT), weaning weight (WWT), and postweaning gain (PWG). Herds were located in diverse environments and differed in level of phenotypic performance. Within-herd BWT analyses considered effects of inbreeding of calf and dam, sex, age of dam (AOD), and contemporary group as fixed and direct and maternal additive genetic effects and permanent environmental effects due to dam as random. The model for WWT included these effects and age of calf. The model for PWG included inbreeding of calf, contemporary group, and direct additive genetic effects. Across-herd analyses were conducted with additional models. The first considered herd-specific inbreeding, sex, and AOD effects. A second model pooled these effects across herds, and a third included pooled sex and AOD effects but ignored inbreeding. Across-herd EPD, including and ignoring inbreeding, were predicted for WWT preadjusted with standard adjustments for Hereford cattle. Within-herd analyses indicated potential for heterogenous genetic and environmental variances across herds. Across-herd variance component estimates were consistent, regardless of the model. Estimates of genetic trends indicated potential for bias in genetic evaluations resulting from choice of model. Differences in magnitude of fixed effects between herds were observed. Genetic evaluations were different when pooled or herd-specific fixed effects were used. Allowance for individual herd differences in fixed effects in across-herd evaluations is suggested.