Le diagnostic du potentiel d’apprentissage par LPAD: Une etude de la fidélité

The present study was intended to test some of the psychometric properties of Feuerstein’s Learning Potential Assessment Device (LPAD), in particular the reliability of the Set Variations II, considered to assess modificability or learning capacity. Thirty 10-year-old children were studied: an experimental group (N=20) submitted to two training sessions with Set Variations II and a control group (N=10). Both groups were administered three times the Standard Progressive Matrices (SPM). The experimenter’s interventions during training were semi-standardized, based on an examination guide, and the helps given were classified in four types. Analyses dealt with the number and type of helps provided as well as with errors during training. Correlations between these indices for the two sessions of training were rather low, indicating weak stability, in particular with respect to the helps provided. Correlations were also computed between the gains in SPM for the three administrations, indicating low stability in amount of progression. Further, it was also noted that the experimental group did not progress more than the control group. Conclusions are drawn with respect to reliability and validity of the Set Variations Test, as well as, on a more theoretical level, with respect to the concept of cognitive modifiability.