Introduction Materials and Methods 24 Results and Discussion 26 Implications LITERATURE CITED 47 ACKNOWLEDGEMENTS 53 LITERATURE REVIEW Accuracy of Ultrasound Producers are continually trying to find ways to improve the end product. One method is through genetic selection. In the past, various breed associations have developed genetic improvement programs for beef cattle primarily based on growth, maternal, and reproductive traits. However, these programs have not focused on the end product until recently, mainly due to limited carcass data collection and the lack of premiums for superior carcasses. Two steps need to be taken in order for genetic improvement programs to be applied to carcass merit in cattle. First, seedstock producers must have an economic incentive from packers, retailers, or consumers to be interested in selection for carcass improvement. Second, the methods by which body composition measurements are taken need to be cost-effective, accurate, and reasonably quick. The use of ultrasound has provided results that can satisfy all of these needs (Wilson, 1992). Reporting Accuracies In the past, there have been many studies trying to validate the use of ultrasound as a means to measure body composition. Correlation coefficients have been the statistic used to report accuracy. Even though this is a useful tool, there are limitations when reporting correlations. These include 1) the idea that population variation affects correlation coefficients (i.e., a small variation in a population will decrease the coefficient, and a large variation will increase correlations); 2) correlation coefficients don't take bias into account (i.e., if a technician uses a scanning technique that over or underestimates the measurement, or if an interpretation technician over or underestimates the measurement while tracing the