A note on some psychometric properties of Sternberg task performance: Modifications to content

Abstract This study investigated one commonly used experimental task on several psychometric dimensions. An investigation of the Sternberg (S. Sternberg, 1966, 1969) task was carried out to study both the measurement and psychometric quality of the task and the effects of variations to content as one source of nonattribute variance on various critical psychometric and measurement properties. The Sternberg task was chosen because of the ease of modifying stimulus content without greatly changing the cognitive process requirements or the task itself. The task was explored with respect to consistency across responses, stability over time, and stability across different stimulus contents. Individual performance was found to be moderately homogeneous across responses, fairly unstable over time (1 week), and fairly stable across stimulus contents (letters, numbers, words, and common symbols). Stabilities of raw scores, slopes, and intercepts were each found to be affected by stimulus content. Results are discussed in the context of need for careful attention to individual variation in performance on experimental tasks. Furthermore, the importance of psychometric analysis of experimental tasks used in individual differences research is stressed.