Discriminatory Power Analysis: An Exemplar Investigation

Abstract

Abstract Aims: This study examined the importance of reporting discriminatory power analysis when determining the validity of using standardized assessment tools with extrapolated normative data for secondary school students. Methods: The participants were 80 secondary school-aged students, selected based on typically developing phonological awareness and literacy skills. The Queensland University Inventory of Literacy (QUIL; Dodd, Holm, Oerlemans, & McCormick, 1996), a formal test of phonological awareness skills for students in years 1 to 7, was used as an exemplar test. Raw scores were obtained for each subtest prior to discriminatory power and normal distribution analyses. Results: When the raw scores were converted to standard scores and percentile ranks the results appeared to provide clinically applicable data. The discriminatory power analysis, however, determined that only 3 of the 10 subtests met sensitivity standards (i.e., were normally distributed). Conclusion: This suggests that discriminatory power, in addition to validity and reliability should be reported to allow informed use of standardized assessment tools.