The dopant and trap density profiles in MOSFETs can be deduced from the static and transient variations of VT caused by changes in VBS. This letter reports a study of errors in VT measurements by three different methods. Relatively small errors in VT could cause serious inaccuracies in dopant profiles. It is found that the major source of errors is the non-constancy of the surface mobility and that errors in VT measured by a "constant VGS-VDS method" cannot be eliminated easily. However, with properly chosen VGS-VT or ID, a "constant ID-VDS method" can provide almost error-free measurements of VT. The theoretical predictions agree with experiments.