We suggest that a consistency check of the redshift correlation function between three redshift samples of Abell clusters is able to distinguish whether the positive redshift correlations, found by Sutherland (hereafter Sutherland effect), at large redshift and small angular separations are mainly the result of foreground/background cluster contaminations or of real clustering. Our results do not support that contaminations are the origin of this effect. Instead, the observed features of the redshift correlation function are compatible with strong, richness-dependent clustering of small numbers of clusters which often results in supercluster elongations along the line of sight. To test this suggestion we have constructed a set of simulated cluster catalogs which have the same two-point correlation function and the same selection effects as the real samples. We found that the Sutherland effect is quite common in the simulated catalogs, which are completely free of contamination effects. The X^2^ test shows that the observed redshift correlation functions can be reproduced in the simulations more than 15%- 30% of the time, thus supporting that contaminations of the Abell cluster catalog are not necessarily the cause of this effect. Our results support that the rich cluster-cluster correlation function, as estimated by Bahcall & Soneira, is not significantly enhanced by such effects, and therefore the biased CDM model is still in disagreement with the observed clustering of clusters.