RESUMO
BACKGROUND: The few studies of children's academic performance in developing countries have largely focused on effects of early risks and cognitive ability and usually neglected other factors such as attention and anxiety. Previously, we reported that scores on the Learning Behaviour Scale (LBS) and the Revised Behaviour Problem Checklist (RBPC) were correlated with academic scores and achievement test scores for village children in St Vincent, the West Indies. AIMS: We examined the stability of LBS and RBPC scores and their ability to predict academic and achievement scores in the same population. SAMPLE: Vincentian village children, ages 6-12, participated in the study: 65 participated in the one-year sample and 68 participated in the two-year sample. METHOD: Children completed a curriculum-based achievement test and the Raven Coloured Progressive Matrices. Teachers completed the LBS and an adjusted RBPC and reported children's academic scores in 1998 and 1999. RESULTS: LBS scores were stable over one year and RBPC scores were stable over two years. LBS, RBPC, and Raven scores predicted achievement and academic scores. For both academic scores and achievement test scores, the greatest improvement in prediction came when the RBPC's attention and anxiety subscales were added to regression models. CONCLUSIONS: Results provide additional support for the finding that Caribbean village children's academic performance is greatly influenced by attention and anxiety problems, not just their cognitive ability. Despite cultural differences, LBS and RBPC scores were as predictive of academic performance in this population as in American populations. The best way to improve academic performance for these children may be to reduce attention problems.