Data collected with a three-component laser Doppler anemometer system is used to investigate velocity bias. The data is collected in the viscous layer of a fully developed turbulent open water flow at a Reynolds number of 14,766 based on the flow depth. The data collected at a relatively low data rate is analyzed using different correction methods including: straight forward arithmetical averaging, inverse velocity and transit time weighting. The streamwise mean velocity components, the RMS values and the kinematic Reynolds stress are computed using the various weighting methods and are compared to a three-component inverse velocity bias correction model which is taken to represent the "true" values. The three-component inverse velocity bias correction results are in general accordance with the expected behavior in open channel flow and are comparable to the results reported by other researchers employing different experimental techniques. The results of this study show that the bias is sensitive to the correction method used and the theory that the mean streamwise velocity error (without correction) is proportional to the square of the turbulence intensity is confirmed experimentally. Averaging the data without correcting it produced the largest bias while the results from the different inverse velocity techniques were approximately the same although the level of the bias varied with the turbulence variable that was analyzed. As reported in the literature, the transit time weighting method requires accurate determination of the residence time. The relatively poor performance of the transit time method in the present comparison is attributed to the poor accuracy in the measurement of the residence time.