The study investigated whether mental age in children, as assessed by the IQ in the Draw-A-Person (DAP) test (Naglieri, 1988), can be improved by practice. In addition, it was tested whether children needed novel content to keep up their performance level during test repetition. The DAP test was given to 6-, 8-, 10-, and 12-year-old children (N ϭ 80) 3 times. In addition, they drew a police figure 2 times, with task sequence counterbalanced. Repeated drawings resulted in significant omission of detail and deterioration of scores, but the novel task instruction temporarily recovered the IQ scores. This did not occur in the reverse sequence of the tasks, with the less specified DAP instruction given in the 2nd half of the drawing series. Furthermore, structural regressions in the human figure drawing could be observed in individual cases. However, 12-year-olds did not need the external introduction of novel content to maintain their IQ score. Correlations showed that this age group redefined their drawing plan on each repetition, and often created unique figures on each occasion, even if the instruction had stayed the same.