A group composed of 124 classroom teachers, reading teachers, and graduate students in reading specialization programs participated in a study that examined how teachers select and use information in judging students' probable performance in reading groups. Each subject read two case studies (one describing a boy with social problems but no reading problems, the other describing a boy with reading problems and no social difficulties) and rated whether the student described in each should be placed in a low, middle, or high reading group. After this rating, subjects were allowed to select one of five areas in which they would like to have additional information about the student. Following this they gave the student a second rating, requested information from another area, and completed a third rating. Results indicated that both initial ratings and initial information requests were related to type of case presented, that the patterns of information requested differed by problem type, and that final ratings and shifts in ratings were related to the positive or negative value of the information received. (Copies of the two cases are included.) (FL)