This paper investigates the magnitude of the mismeasurement that occurs when only a few education categories are used in the construction of a constant quality index for labor input. By employing a very comprehensive data set it is found that the error resulting from the omission of information...