Compared 2 types of elaboration training (imagery and sentence elaboration) over 2 levels of verbalization (none and overt). 32 elementary, educable mentally retarded children were randomly assigned to 1 of the 4 elaboration treatment conditions of the resulting 2 * 2 factorial design. Results indicate that verbalization greatly influenced the degree of effectiveness of elaboration training. However, no support was found for the secondary hypothesis that overt verbalization would result in relatively greater facilitation with one type of elaboration (imagery) than with the other type (sentence elaboration). 2 sets of explanations are examined. In addition, educational applications of elaboration are discussed, with special attention given to the implications of the verbalization findings. (25 ref.)