The statistical properties of galaxy distance estimators corresponding
to the Tully-Fisher and D-n-sigma relations are studied, and a rigoro
us framework for identifying and removing the effects of Malmquist bia
s due to observational selection is developed. The prescription of Sch
echter for defining unbiased distance estimators is verified and exten
ded to more general-and more realistic-cases. Finally, the derivation
of ''optimal'' unbiased estimators of minimum dispersion, by utilizing
information from additional suitably correlated observables, is discu
ssed and the results applied to a calibrating sample from the Fornax c
luster, as used in the Mathewson spiral galaxy redshift survey. In thi
s illustrative example the optimal distance estimator derived from app
arent magnitude, isophotal diameter, and line width has an intrinsic s
catter which is similar to 22% smaller than that of the conventional T
ully-Fisher relation.