A study investigated the effects of teacher probes, a type of questioning, on children's written revisions. Subjects, 260 children from grades two through six in a midwestern school district of moderate size, were randomly assigned to one of two conditions (probe or comment). Subjects in both conditions were presented with a story starter that they completed after group discussion. When the stories were returned, subjects reread and revised their stories to answer the questions written on them. Children in the probe condition received questions directed at specific content in their stories, while subjects in the comment condition received generalized remarks such as "nice job" or "tell me more." The revised stories were then marked holistically on a five-point scale. The main effect of condition was significant, indicating that children's revisions by teacher probes can produce qualitatively better text as measured by holistic ratings. The main effect of grade level was not significant, indicating that probing does not appear to be more effective at one grade level than at another. (HTH)