General ability index and cognitive proficiency index for the WISC-IV: Empirical norms vs. statistical norms

The General Ability Index (GAI) was developed within the WISC-III to measure general cognitive ability. The GAI can be used as a substitute for the FSIQ. Recently, the Cognitive Proficiency Index (CPI) was developed. The GAI and CPI tables for the French WISC-IV were recently created using statistical approximation. However, it has been suggested that tables created with statistical approximation were less proficient than tables created using samples. Objective: The objective of this study was to compare GAI and CPI normative tables of the French WISC-IV which use a statistical approximation procedure, to tables based on a sample. Method: To explore the validity of the GAI and ICC scores, we studied 182 children aged from 8 to 12-years-old. Results: Correlations between both types of tables were 0,997 for GAI and 0,999 for CPI, and the absolute mean difference was 5.38 points and 2.24, respectively. For the children presenting a GAI sub(S lower than 89, GAI) sub(E) suggested that it is likely that their scores are lower about 8-10 points than that indicated by GAI sub(S. Conclusion: Results of this study support the use of GAI and CPI to provide different views of children's cognitive abilities.)Original Abstract: L'indice d'aptitude general (IAG) a ete developpe dans le WISC-III pour estimer le niveau cognitif general. Il s'agissait de proposer une alternative au QI Total. Plus recemment, l'indice de competence cognitive (ICC) a ete propose. Jusqu'a il y a peu, seules les normes americaines de ces indices etaient disponibles ; les normes francaises IAG et ICC pour le WISC-IV ont ete recemment developpees a partir d'une procedure d'approximation statistique. Toutefois, les normes developpees de cette maniere seraient moins appropriees que les normes construites sur la base d'un echantillon. Cette etude vise a comparer les normes IAG et ICC developpees a partir d'une procedure d'approximation statistique a celles creees a partir d'un echantillon de 182 enfants ages de huit a 12 ans. Les correlations entre les deux types de normes sont de 0,997 pour IAG et de 0,999 pour ICC ; les differences absolues moyennes sont respectivement de 5,38 et de 2,24 points pour IAG et ICC. Enfin, pour les enfants presentant un score IAG sub(S inferieur a 89, les normes IAG) sub(E) semblent indiquer que leurs competences seraient inferieures d'environ 8-10 points a celui indique par le score IAG sub(S. Les resultats de cette etude indiquent que les scores IAG et ICC peuvent con