Abstract
The Trail Making Test Part B (TMT-B) is widely used in clinical and research settings as a measure of executive function. Standard administration allows a maximal time score (i.e., floor score) of 300 s. This practice potentially masks performance variability among cognitively impaired individuals who cannot complete the task. For example, performances that are nearly complete receive the same 300-s score as a performance of only a few moves. Such performance differences may have utility in research and clinical settings. To address this, we propose a new TMT-B efficiency metric designed to capture clinically relevant performance variability below the standard administration floor. Our metric takes into account time, correct moves, and errors of commission and omission. We demonstrate that the metric has concurrent validity, permits statistical analysis of performances that fall below the test floor, and captures clinically relevant performance variability missed by alternative methods.