Abstract The study reassessed the configural and metric structures of the Brazilian version of the Control, Autonomy, Self-realization and Pleasure (CASP-19) quality-of-life scale. Data came from the EpiFloripa Ageing Study, which included 1,131 respondents from Southern Brazil. The original and two recently factorial solutions for the Brazilian CASP-19 were initially examined. Exploratory Factor Analyses and Exploratory Structural Equation Models were estimated in the first half of the sample, selected at random. In the second half, Confirmatory Factor Analyses determined the most tenable configural and metric model for the instrument. Neither the original nor the two Brazilian solutions were supported by our data. Instead, we suggest that two factors underlie CASP-19's configural structure: while the first one groups the control and autonomy dimensions, the second combines self-realization and pleasure. Except for four items, all others presented moderate to strong loadings, and only two showed a theoretically meaningful and sufficiently large residual correlation, which was worthy of inclusion in the final model. Cross-loadings were not detected. When assessed in a population-based sample of older respondents, this Brazilian version of the CASP-19 appeared to have two factors, moderate to strong loadings and a pair of redundant items. Future studies should evaluate the consistency of these findings, examine the scalar structure of the instrument, and assess configural, metric and scalar invariance across social groups.