In 1988 the Strategic Highway Research Program purchased four falling weight deflectometers (FWDs). During the acceptance testing it became evident that an improved procedure for calibration was needed to determine whether the specifications for the precision and the accuracy of the sensors were achieved. The authors were responsible for developing the procedure. This paper reports on the steps taken during the development of the calibration protocol. The reasons underlying the equipment and the procedures chosen are discussed. The sources of error in FWD measurements are identified, and ways that have been used to reduce those errors are reported. The goal was to reduce the systematic (bias) error to less than 0.3% through calibration. This level of error ensures that the random measurement error is larger than the systematic error for all pavement deflections less than 600 μm [24 mils (1 mil = 0.001 in.)]. Experience has shown that most highway pavements deflect less than 600 μm. The effects of FWD measurement errors on backcalculated pavement moduli are briefly reviewed. Verification of the protocol by several means showed that the calibration goal was achieved. Subsequent experience with the calibration protocol has shown that it has been effective and that it ensures high quality in the FWD data.