This study compares the intra- and interobserver variance of Cobb angle measurements of primary and secondary curves on digital radiographs versus traditional radiographs. Four orthopaedic surgeons of varying experience measured the Cobb angles from a standard posteroanterior thoracolumbar scoliosis radiograph (25 digital, 25 traditional) on two occasions 2 weeks apart. The intra- and interobserver variances were calculated and compared for major versus minor curves and the digital versus traditional radiographs. There was no statistical difference in the mean error index, the variability in choosing the end vertebra on successive measurements, between the digital and traditional groups. Similarly, there was no significant difference in the intraobserver or interobserver variance between the digital and traditional groups. Digital radiographs are comparable to the use of traditional radiographs for following patients with adolescent idiopathic scoliosis. Furthermore, increasing years of experience appears to result in fewer errors and more consistency using the Cobb method.