Core affect is an elementary affective state expressed through subjective feelings. Nonetheless, despite extensive empirical evidence in the field, researchers still disagree about its dimensionality. Thus, the present thesis aims to verify the validity evidence of existing models of core affect, overcoming the methodological issues of previous studies, and establishing the dimensionality of core affect. First, theoretical contributions are presented, and both conceptual (e.g. what is core affect?) and methodological issues (e.g. how core affect is measured?) are discussed. Following that, two empirical studies are presented. The first study explores the dimensionality of core affect and provides validity evidence of a new core affect measure. In the second study, a robust-to-biases core affect measure is developed and tested. In addition, the relationship between core affect, contextual variables (e.g. mood) and personality traits are studied in a longitudinal design. Items formats and their consequences in the measurement of core affect (e.g. rating scales, forced-choice items) are debated. Theoretical and methodological advances are discussed at last, as well as limitations and future directions.