The direct and generalized effects of a performance feedback system introduced sequentially in two elementary classrooms were assessed during two daily story writing periods. Following baseline, a treatment package consisting of explicit timing, self-recording, and public posting was instituted only during the first story writing period with the target behavior being words per minute. Next, the target was changed to the number of action words written. Following a return to baseline several follow-up assessments were made. The treatment package produced a sustained increase in both target behaviors which generalized to the second story period. Results are mor eequivocal for the percentage of action words. Observations suggest that student performances comments mediated the generalization and maintenance of effects.