Anticipation describes an inheritance pattern within a pedigree in which disease severity increases, and/or age at onset decreases, in successive generations. This phenomenon has been described in different samples of schizophrenic subjects, and could explain many inconsistencies in the inheritability of schizophrenia. Anticipation is, however, subject to numerous and significant biases, partially controlled by different methodologies used in different studies. We analyzed the anticipation effect on an original sample of schizophrenic patients (n = 57) who had at least one other schizophrenic in their family belonging to another generation (father/mother, uncle/aunt, son/daughter). We tested the anticipation effect according to previously published methodologies, such as percentages of parent-child pairs showing negative versus positive anticipation, comparison of anticipation limited to parent-child or uncle-nephew pairs, anticipation analysis on the basis of families with unilineal origins only, and comparison of the age at onset-survival distribution of the two generations. The 31 schizophrenic subjects who belonged to the younger generation had a significantly earlier age at onset (24.58 years) than the 26 schizophrenic subjects who belonged to the older generation (36.46 years). Whatever the method used to control biases, we significantly found earlier age at onset for schizophrenic patients from the younger generation. There is strong evidence for the existence of the anticipation effect in schizophrenia in our sample, as well as in various others, which may elucidate numerous inconsistencies in clinical and epidemiological data which characterize schizophrenia. Looking for expanded trinucleotide repeats is thus the next step to detect the gene(s) that are potentially involved.