We used an alternating treatments design to compare the effects of active student response error correction and no‐response error correction during sight word instruction. Six students with developmental disabilities were provided one‐to‐one daily sight word instruction on eight sets of 20 unknown words. Each set of 20 words was divided randomly into two equal groups. Student errors during instruction on one group of words were immediately followed by the teacher modeling the word and the student repeating it (active student response instruction). Errors on the other group of words were immediately followed by the teacher modeling the word while the student attended to the word card (no‐response instruction). For all 6 students, the active student response error‐correction procedure resulted in more words read correctly during instruction, same‐day tests, next‐day tests, 2‐week maintenance tests, and generality tests (words rea