The galaxy clustering correlation length r0 is studied using a variety
of galaxy samples drawn from Huchra's CfA-I galaxy redshift catalogue
(completeness limit m(B) less-than-or-equal-to 14.5) and the first Cf
A-II slice (m(B) less-than-or-equal-to 15.5). Estimates of r0 depend o
n the sample, and some samples show the well-known trend that r0 incre
ases with sample depth. We study this phenomenon both in complete volu
me limited samples, and in velocity limited samples with an a priori f
ixed absolute magnitude limit M(com) = -18.5 + 5 log h, using a select
ion function based on the sample luminosity function. The first sample
is certainly affected by luminosity segregation. To isolate this effe
ct we have studied samples with different depth and the same absolute
magnitude limit. The effect of the luminosity segregation is clearly n
oticed in biasing the value of r0. In the analysis of incomplete sampl
es, if the luminosity function were a universal function, we would not
expect strong luminosity dependent effects, therefore the observed va
riability of r0 in these samples is mainly due to the weight large sca
le inhomogeneities and the fact that we are not using a fair sample of
the Universe. We have also studied the influence of the corrections p
erformed by Einasto et al. (1984): galactic absorption and random pecu
liar velocities in clusters.