Abstract
Much confusion exists about gifted underachievement as a result of significant variations in how gifted underachievement has been identified and measured. From a review of the literature, five methods were found to be commonly used to identify gifted underachievement, two of which also measure the degree of gifted underachievement. In this project, the validity of the use of these methods was assessed using empirical convergence, criterion, and generalisation evidence obtained using data collected from a K–12 school in the Sydney metropolitan area (Australia). First, convergence was assessed using three different approaches to compare the results obtained using the five different identification/ measurement methods: (a) the differences in proportions of gifted students identified as exhibiting underachievement, (b) the correlation of identification/measurement results, and (c) the agreement of the identification/measurement results. The findings suggested that the different methods commonly used to identify/measure gifted underachievement cannot be considered convergent, and therefore should not be used interchangeably. Second, a latent class model was used to assess the criterion validity of the individual identification/ measurement methods. Two of these methods were found to have strong levels of criterion validity. Third, meta-analysis was carried out across 41 different combinations of expected achievement and actual achievement data obtained from the school, to indicate that generalisation was not possible across different combinations of expected/actual achievement data. A final assessment of the validity of each identification/measurement method was completed by synthesising the empirical evidence within the context of the planned uses of the methods. The final assessment demonstrated that the simple difference method may be the most valid method to use to identify and measure gifted underachievement.