Abstract The ability to properly assess the dimensionality of a set of item responses is essential in psychometrics. Many popular psychometric methods using item response theory (IRT) such as item calibration and test equating assume that there is a single latent trait underlying the responses. As well, the popularity of multidimensional models has increased. Furthermore, test dimensionality is directly linked to the structural aspect of the construct validity of a test. History has shown some confusion in the definition of dimensionality and in the methods to assess it. Recent literature relates both the definition of dimensionality and its methods of assessment to the concepts of local independence and essential independence. This entry provides an overview of these definitions and of the more current methods of dimensionality assessment.