Abstract
INTRODUCTION: Cognitive reserve is the ability to better tolerate brain damage through pre-existing and compensatory cognitive resources. One assessment method is the Rami CRQ-Cognitive Reserve Questionnaire. The objective was to carry out an analysis of the informative quality of the CRQ from the item response theory (IRT), in order to provide more precise data on the reliability of internal consistency. Convergent validity was also tested with measures of attention, working memory, and fluency. SUBJECTS AND METHODS: 210 Argentines from the general population (mean age, 66.8 years) participated. The CRQ was administered together with the digits test and three fluency tasks. A graded response model was fitted from IRT with estimation of discrimination parameters (a) and difficulty (b), and a CRQ information curve was created. Bivariate and partial correlations were made. RESULTS: The IRT indicated high discrimination for the CRQ items 'Education' and 'Occupation level' (both for the 8-item version and the 6-item version). In the CRQ of 8 items, low discrimination was obtained for 'Musical training' and 'Intellectual games'. In both versions of the CRQ, the curve indicates greater informational value at a low level of the construct. There was a correlation with the digits test and with fluency tasks, even when controlling for age. CONCLUSIONS: This study is the first analysis of CRQ from IRT, concluding that the instrument is more reliable when applied to subjects with less reserve. The CRQ has acceptable convergent validity.