In this paper we investigate whether a misestimate of proper motions could have been a source of substantial systematic errors in the statistical parallax determination of the absolute magnitude of RR Lyrae stars .
In an earlier paper , we showed that the statistical parallax method is extremely robust and rather insensitive to various systematic effects .
The main potential problem with this method would therefore arise from systematically bad observational inputs , primarily radial velocities and proper motions .
In that paper , we demonstrated that the radial velocities have not been systematically misestimated .
Here we turn our attention to proper motions .
We compare three different catalogs of proper motions — Lick , Hipparcos and the one compiled by Wan et al .
( WMJ ) .
We find that the WMJ catalog is too heterogeneous to be a reliable source .

We analyze the sample of 165 halo RR Lyrae stars with either Lick or Hipparcos proper motions .
For the stars with both Lick and Hipparcos proper motions we use the weighted means of reported values .
Various possible biases are investigated through vigorous Monte Carlo simulations and we evaluate small corrections due to Malmquist bias , anisotropic positions of the stars on the sky , and non-Gaussian distribution of stellar velocities .
The mean RR Lyrae absolute magnitude is M _ { V } = 0.74 \pm 0.12 at the mean metallicity of the sample \left < [ Fe / H ] \right > = -1.60 , only 0.01 mag brighter than the value obtained in the previous study which did not incorporate Hipparcos proper motions .
The faint absolute magnitudes of RR Lyrae stars confirmed by this analysis gives strong support to the short distance scale .

\keywords astrometry — distance scale — catalogs — Galaxy : kinematics and dynamics — methods : analytical , statistical — stars : variables : RR Lyrae