Job evaluation is used to establish pay for many workers in the United States and has been used to measure the extent of pay discrimination. However, job evaluations are subject to measurement error that can bias these estimates. Using computed reliability ratios to adjust for measurement error in a study of Iowa state government jobs, we find that measurement errors exaggerate the implied extent of discrimination against predominantly female jobs by 34%-44%. Measurement errors also exaggerate the number of independent job factors which affect pay. A practical procedure for making these corrections is illustrated.