Some sources of error and possible bias in Danscan ultrasonic measurements of cattle

Abstract

ABSTRACTPhotographs of ultrasonic images of 10 animals were taken. Two replicate photographs were taken at each of four body positions (10th rib, 13th rib, 3rd lumbar and hindquarter) on both sides of the body in the morning and the afternoon of the day of scanning. Several measurements were taken on each photograph by an experienced interpreter. Replicate photographs failed to account for all possible sources of measurement error. Since neither time of day nor side of the body affected the mean value, taking observations at different times of the day or on both sides of the body could permit most sources of measurement error to be taken into account. Variation due to errors of measurement and differences among animals are presented. The residual error distribution contained several extreme outliers.It was concluded that a better understanding of all the sources of bias and error will be needed if ultrasonic measurements are to be more widely used.