Longitudinal consistency of Matching Familiar Figures Test performance from early childhood to preadolescence.

Age-appropriate versions of the Matching Familiar Figures Test (MFFT) were administered to a longitudinally followed sample of children at ages 3, 4, 5, and 11. Uncorrected for attenuation, MFFT error scores were more consistent over time than MFFT latency scores for both girls and boys. When the longitudinal MFFT correlations were corrected for attenuation, the magnitude of the MFFT error coefficients increased considerably, whereas the magnitude of the MFFT latency coefficients remained basically unchanged. Thus MFFT latency scores seem to have relatively little long-term implication as compared to MFFT error scores. Across-time consistency of MFFT error scores was an appreciable function of performance IQ, supporting a competence view rather than a conceptual tempo view of what the MFFT measures. In girls, the size of the inverse relation between MFFT error and MFFT latency increased from age 3 to age 5 and then leveled off. In boys, this relationship remained unchanged between age 3 and age 5 but increased markedly after age 5. The several implications of these results for the validity of the MFFT as a measure of an enduring cognitive style are discussed.