This study evaluated the effectiveness and efficiency of constant time delay (a near errorless learning procedure) in a small-group instructional arrangement. Three fifth-grade students with learning disabilities were taught to recognize multisyllabic basal vocabulary words. A multiple-probe design across behaviors (word sets) was used to evaluate the procedure. Following instruction on each word set, students were assessed on their ability to recognize their own target words, recognize observational words, spell both target and observational words, and define both target and observational words. The results indicated that the constant time-delay procedure was reliably implemented and was effective in establishing criterion-level performance for all students with extremely low error percentages.