Before slaughter, 44 Hereford-sired steers were measured ultrasonically for backfat (UFAT) and longissimus muscle area (ULMA) between the 12th and 13th ribs by three technicians (TECH) using two different machines (MACH) on two consecutive days (DAY). Each TECH interpreted (INT) his own images in addition to other TECH images. The absolute values of the difference between the 2 DAY's ultrasound measurements for ULMA (magnitude of LMAR) and UFAT (magnitude of FATR) were analyzed with a model including fixed effects of MACH and TECH with a random effect of steer and all interactions. For both magnitude of LMAR and magnitude of FATR, MACH x TECH was significant (P < .10). Correlations between the 2 DAY's measurements ranged from .36 to .90 and .69 to .90 for ULMA and UFAT, respectively. Simple statistics to quickly evaluate TECH and MACH were developed. Root mean squared errors (RMSE) and error standard deviations (ESD) between repeated measurements ranged from 3.89 to 11.32 and 3.93 to 11.34 cm2 for ULMA and .12 to .20 cm and .12 to .20 cm for UFAT, respectively. For accuracy, the absolute values of the difference between the ultrasound and carcass measurement for fat (magnitude of FATD) and longissimus muscle area (magnitude of LMAD) were analyzed with a model accounting for fixed effects of DAY, TECH, and MACH and a random effect of steer with all higher-order interactions. For magnitude of LMAD, TECH x MACH was a significant source of variation (P < .001). Also, a similar model was fit that included the fixed effects of TECH, MACH, and INT and a random effect of steer with all interactions. The MACH x INT interaction was found to be significant for magnitude of LMAD (P < .05). From this research, TECH and MACH differences do exist. Ultrasound is a valid means of measuring carcass traits in live steers if appropriate personnel and equipment are selected.