Morphometric X-ray absorptiometry (MXA) is a low-radiation technique used for the identification of vertebral deformities. The reliability of MXA measurements is dependent on the precision of the technique, and this is influenced by system error, variability associated with morphometric analysis, and variability within study populations. Short-term precision in vivo affects the identification of prevalent deformities, and may vary according to the age and health status of the study population. The aims of this study were to measure short-term precision in vitro and in vivo for MXA, and to compare intraoperator variability for repeat marking of radiographs in morphometric radiography (MRX) and scan images in MXA. For short-term precision in vitro (based on 30 consecutive MXA phantom scans), the coefficient of variation (CV) was approximately 1%. Intraoperator variability for repeat marking of baseline radiographs and MXA scans (marked without the compare facility) in a population-based group of 32 postmenopausal women, was significantly lower for MRX (CV for vertebral heights = 1.5%) than for MXA (CV = 2.9%). The CV for duplicate same-day MXA scans performed in subjects with osteoporosis (n = 20) was 2.2% for vertebral heights between T8 and L4 using the compare facility, and 4.8% when scans were analyzed without the compare facility. We conclude that short-term precision for MXA is good for vertebrae T8 to L4, but the technique is limited due to poor image resolution in the upper thoracic vertebrae. Intraoperator variability for repeat marking was lower for MRX, but the error associated with marking MXA scans remained below the minimum change required for identification of a vertebral deformity.