The purpose of this study was to investigate whether the activity distribution of P3a and P3b across the scalp changes with age or remains identical. 103 well-functioning adults, 20-90 years, performed a visual three-stimuli oddball ERP task yielding both a P3a and a P3b. The sample was divided into three age groups: 20-44 years, 45-69 years, and 70-90 years. The P3as and P3bs from 20 electrodes were subjected to principal component factor analyses. The unrotated factor matrices from the three different age groups were then compared using orthogonal Procrustes rotation to maximize congruence, and Tucker's coefficient of congruence was used as index of similarity. The results showed that less variance in latency than in amplitude was explained by three factors. Further, for amplitude, the factor structure remained the same for both P3a and P3b across different age groups, implying that the structure of the neurophysiological generation of the P3 amplitude operates similarly at different ages. For P3a and P3b latency the congruence coefficients were weaker, implying that when it comes to latency, the neural generators or parts of the generator circuits are differentially susceptible to the influence of age. It is speculated that a possible reduced temporal synchronicity across electrodes could result in less predictable shifts of the clustering of P3 latency, and may explain the presently observed factor variability with age. Implications of the findings and the utility of Procrustes rotation in ERP studies are discussed.