Statistical Effects from Hipparcos Astrometry

Abstract The Hipparcos astrometry is used mainly for the derivation of stellar physical quantities such as luminosity, masses and velocity. However, sample selections on data with observational errors or an intrinsic dispersion may lead to biased estimates, especially when the error distributions are non-Gaussian. We review the classical biases and the ways to avoid them through the use of statistical methods.