The Bracken School Readiness Assessment (BSRA) has been used in large studies such as the Millennium Cohort Study (MCS). Important conclusions might be done regarding its reliability for the prediction of children's school readiness taking advantage of such large-scale evaluation. Although BSRA has being largely used, few are the studies at item-level under latent approach investigating its psychometric features. Using data from 14,899 2-3-year olds who participated in the MCS, we used Bayesian confirmatory factor analysis to examine multidimensionality of the subtests of the BSRA and their consistencies, specificities, and reliabilities. We found clear indications of multidimensionality. From the 88 items, 10 showed low reliability. Future research may consider excluding these low reliability items to improve the psychometric properties of the BSRA and its use as multidimensional measurement tool.