Much interest has recently been focussed on the possibility of the involvement of unstable DNA in the etiology of schizophrenia and bipolar affective disorder (BPAD), following several publications that report increases in frequency of large CAG/CTG repeats in affected individuals. Using the Repeat Expansion Detection (RED) technique, we have performed a matched control pair analysis for both disorders. No significant differences in CAG/CTG repeat sizes were observed for 52 bipolar affecteds and matched controls (P = 0.15), and borderline significance was observed for 54 schizophrenia affecteds and matched controls (P = 0.05), using a (CTG)10 oligonucleotide (one-tailed t-tests for paired samples). Furthermore, using a (CTG)17 oligonucleotide, no significant differences were observed for 58 bipolar affecteds and 55 schizophrenia affecteds compared to 81 unmatched controls. No significant sex effect was observed for either group, and no significant differences in repeat size were found for responders and non-responders to drug treatments. More importantly, there was no significant correlation (either positive or negative) between age of onset of disease and size of repeat. We thus cannot conclude that CAG/CTG trinucleotides are involved in psychotic disorders and that either the differences observed in similar studies may be the result of population stratification, or that the increased frequency of larger repeats amongst affected individuals is a much smaller effect than previously thought.