In order to evaluate the most effective delivery format for providing gifted/talented education, 247 elementary students in gifted programs in Evansville, Indiana, were divided into experimental students, who received instruction in self-contained classrooms of fellow gifted students, and control students, who received instruction in a pull-out program. Pre- and post-measures (about 8 months apart) on the Developing Cognitive Abilities Test (DCAT), writing samples, collage drawings, and post-test measures for the Indiana Statewide Test of Educational Progress (ISTEP) and the California Achievement Test (CAT) measured students' progress. Results from the DCAT and the collage drawing showed highly significant differences favoring the self-contained classroom format. Analysis of the writing samples showed some advantage for the control group. No substantive significant differences were found on the ISTEP or the CAT. Based on the number of significant differences that favored the self-contained classroom, it was concluded that this seemed to be the more effective format, but careful attention is recommended to determine how writing scores can be improved in that kind of format. It is also recommended that other delivery models, such as in-class enrichment, be considered. The major portion of the report is composed of statistical results from the assessment instruments and examples of student materials. Contains 10 references. (JDD)