Research findings emphasize the need for programs for the gifted reader, particularly the need to determine which methods or programs best benefit the gifted student. The purpose of this study was to assess the effectiveness of the Junior Great Books (JGB) literature program on reading comprehension achievement of gifted fifth grade students due to the factors of a) discussion/non-discussion and b) levels of students' experience in the JGB program. This study was replicated to determine if similar results in reading comprehension achievement occurred by using different JGB stories. In addition, the study explored methodological issues of cross-level inferences to determine if different results were obtained when applying the General Linear Model to individual-level and aggregate-level data. The research was experimental in design. Seventy-eight fifth grade gifted students were randomly assigned to eight groups in two schools, four treatment groups with discussion and four control groups without discussion. Students in one school had no prior JGB experience; students in the other school had a range of one to three years JGB experience. Two different JGB literature stories, randomly selected, were read by all groups. Reading comprehension was measured, pre and post, by an instrument that emerged from the JGB stories using the cloze procedure. Additional data to measure students' achievement in the form of responses to openended questions about the JGB literature stories were collected. Observations of groups featuring discussions/non-discussions were taped and rated to determine whether appropriate formats were followed by the leaders. Individual-level data and aggregate—level data were analyzed using a two—way ANOVA with nesting using the General Linear Model of the Statistical Analysis System (SAS). lt was found, and supported by study replication, that there were no significant differences between those fifth grade gifted students who participated in the JGB s program with discussion and those who did not. Nor were there differences due to level of experience in the JGB program. Similar results were obtained when applying the General Linear Model to individual—level and aggregate-level data. Further investigation of the JGB program and related methods and procedures involved in the study were suggested. / Ph. D.