Psychometric Characteristics of the <i>n</I>-back Task: Construct Validity Across Age and Stimulus Type, Internal Consistency, Test-Retest and Alternate Forms Reliability

Abstract

Age-related impairments are observed in a variety of cognitive constructs, including significant deteriorations in working memory (WM) performance. To measure the WM, the commonly used task in cognitive aging is the n-back. Despite its widespread use, however, its psychometric properties have not been sufficiently studied. Previous research has typically examined construct validity by correlating n-back task measures with other WM tasks and has primarily assessed reliability through test-retest, occasionally including split-half reliability. In this study, we comprehensively examined the psychometric properties of the n-back task by evaluating its construct validity across different age groups (younger and older adults) and stimulus types (verbal and non-verbal) and by assessing reliability through three methods: test-retest, split-half, and alternate forms. Our findings showed similar performance in verbal and non-verbal n-back tasks for both younger and older adults in terms of discrimination ability and reaction time. However, age-related differences favored younger adults, providing evidence for construct validity. In addition, the n-back measures were stable over time, consistent for two halves, and comparable for different forms, demonstrating task. This study provides the first evidence supporting the validity of the n-back task with regard to age and stimulus type, alongside its reliability. These findings have important implications for clinical research and practice, suggesting that the n-back task can be confidently utilized in neuropsychological assessments.