The spinning rotor gauge is the most stable device available for measuring pressure between 10(-4) and 10(-1) Pa. The calibration of these gauges is determined by measuring an effective accommodation coefficient, which is used in conjunction with an equation derived from the kinetic theory of gases to calculate the pressure seen by the rotor gauge. To assess the reproducibility of the spinning rotor gauge (and hence its accommodation coefficient), eight spinning rotor gauges were calibrated numerous times simultaneously with the primary midrange vacuum standard at the National Institute of Standards and Technology over a 15 month period. Events encountered by a rotor during its normal operation were investigated for their effects on the accommodation coefficient of the rotor. The effects of baking as well as "crashing" of the rotors were also investigated. We found that the accommodation coefficient may be dependent on the orientation of the rotor's suspension as a consequence of anisotropically distributed blemishes on the rotor's surface. Consequently, the accommodation coefficient will not reproduce well when the rotor is resuspended because the rotational axis is not replicated. Data on the accommodation coefficient obtained from ordinary rotors as well as rotors specifically prepared with a band of scratches to exaggerate the anisotropic nature of the blemishes will be presented. (c) 2007 American Vacuum Society.