Learning strategies have been widely studied, and they have been related to academic achievement in several studies. Among the measurement instruments developed, the 44-item version of the Learning Strategies Scale (ACRA) is, by far, one of the most widely used questionnaires in the Spanish speaking context. This instrument has been recently shortened to a 17-item version. Neither the 44-item nor the 17-item version have been tested in the Dominican Republic. The aim of this study is to study the psychometric properties of the 44-item and 17-item versions of the ACRA. Participants were 1712 Dominican secondary school students. Both structures of the ACRA were tested via CFA; alpha and CRI Index were calculated for internal consistency; and, to obtain evidence of the relations with other variables, two models of academic achievement prediction were tested. Results pointed adequate internal structure of both versions, but higher consistency estimates for the former. Both versions of the scale were invariant across gender and age, and no latent mean differences were found. In the prediction model of the 44-item, cognitive strategies was the only significant predictor of achievement; in the 17-item ACRA, it was emotional-social support. These predictive models were, again, invariant across gender and age. Taking into account these results, the 44-item version has shown better properties to asses learning strategies in the Dominican Republic. However, when our aim is to predict academic achievement, a mixed version of the 44- and the 17-item versions could be used.