Purpose The Cobb angle as an objective measure is used to determine the progression of deformity, and is the basis in the planning of conservative and surgical treatment. However, studies have shown that the Cobb angle has two limitations: an inter-and intraobserver variability of the measurement is approximately 3-5 degrees, and high variability regarding the definition of the end vertebra. Scoliosis is a three-dimensional (3D) pathology, and 3D pathologies cannot be completely assessed by two-dimensional (2D) methods, like 2D radiography. The objective of this study was to determine the intraobserver and interobserver reliability of end vertebra definition and Cobb angle measurement using X-rays and 3D computer tomography (CT) reconstructions in scoliotic spines. Methods To assess interoberver variation the Cobb angle and the end vertebra were assessed by five observers in 55 patients using X-rays and 3D CT reconstructions. Definition of end vertebra and measurement of the Cobb angle was repeated two times with a three-week interval. Intraclass correlation coefficients (ICC) were used to determine the interobserver and intraobserver reliabilities. 95% prediction limits were provided for measurement errors. Results Intraclass correlation coefficient (ICC) showed excellent reliability for both methods. The measured Cobb angle was on average 9.2 degrees larger in the 3D CT group (72.8°, range 30-144) than on 2D radiography (63.6°, range 24-152). Conclusions In scoliosis treatment it is very essential to determine the curve magnitude, which is larger in a 3D measurement compared to 2D radiography.