The aim of this investigation was to compare the classic method of tracing by hand with a computerized method, where the lateral cephalograms were scanned at 300 dpi and digitized onscreen. The inter- and intra-observer errors were investigated for tracing and digitizing errors. Thirty lateral cephalograms were scanned into digital format at 300 dpi, displayed on a high-resolution monitor, and processed twice by two operators using Dolphin Imaging Software 9.0. The same radiographs were then traced and measured manually by the same two operators. Intraclass correlation coefficients (ICC) were used for detecting intra- and inter-rater agreement for each cephalometric variable. The results showed that each operator was consistent in the repeated measurements; all ICC were greater than or equal to 0.90 and none of the 95 per cent confidence limits on these ICC had a lower boundary less than 0.84. Inter-rater agreement also showed correlation coefficients greater than 0.75. The angles, maxillary height, maxillary depth, y-axis, FMA, and nasolabial, and the distance N perpendicular point A had a wider reliability interval and lower correlation than the other parameters tested. The findings demonstrated that the use of computer software for cephalometric analysis carried out on scanned images does not increase the measurement error when compared with hand tracing.