The impact of two variations of teacher attention to oral reading errors and a notreatment control condition were compared in an alternating treatment design. For four moderately mentally retarded children, each uncorrected oral reading error was followed by: (a) immediate teacher attention, with the teacher correcting the error word as soon as it occured, (b) delayed teacher attention, with the teacher correcting the error word at the end of the sentence in which it occured or within 10 to 15 seconds if the subject paused following an error, or (c) no teacher attention. Both immediate and delayed teacher attention were effective in reducing the number of uncorrected oral reading errors and increasing the number of self-corrections when compared to the no-treatment control. However, the delayed condition had greater effect than the immediate condition on both measures. In a later remediation phase, when only the delayed treatment was given, uncorrected errors remained at low levels whereas self-corrections generally increased. Total errors whether subsequently self-corrected or not, with one exception, decreased during the study.