Chemical evolution models for the local galactic disk are combined wit
h stellar evolutionary tracks to simulate real stellar samples in a Mo
nte Carlo-like approach. These synthetical samples enable to study the
significance of selection bias in major observational constraints on
chemical evolution models: the [O/H] distribution of long-lived stars
and the age-[O/H] relation (AOR). It is found that selection effects d
o not much alleviate the G-dwarf problem and that there is essentially
no selection bias in the AOR. It is difficult, however, to fit both c
onstraints by the same simple chemical evolution model.