with the instructor's help. Following the orientation, the two raters evaluated a random sample of direct and minimal generalization audiotapes (9 transcripts for each of the 3 participants). The training enabled the two raters to achieve a high level of agreement in their evaluations of the transcripts (92% agreement rate). Following training, the raters reliably evaluated the remaining transcripts. The Kappa coefficient of inter-rater agreement for all participant transcripts was .91 (Cohen, 1960).
Consistent with the objectives of the investigation, results demonstrated the utility of the "self-advocacy" training package. Figures 1-3 indicate the extent to which the three participants acquired, maintained, and generalized the target behaviors over the course of the investigation. Data are reported in terms of each student's demonstration of the 17 target behaviors in the series of direct, minimal generalization, extended generalization, and pre/posttests. For efficiency in reporting, results regarding the percentage of the 17 target behaviors acquired are charted across time and assessments rather than presented in 17 multiple baseline charts for each target behavior. Multiple baseline graphs for each target behavior, available from the senior author, indicate that the behaviors were acquired primarily after their introduction in a lesson and that previously acquired targets were maintained over time. Because several of the target behaviors were common sense (i.e., naturally occurring), participants occasionally demonstrated some behaviors before they were taught. For example, participants used about 35% of the 17 request behaviors after lesson one rather than just the three taught in that lesson (18%). Examples of behaviors that occurred out of the training sequence included thanking the instructor and mentioning that the student needed a classroom accommodation.
Each of the 11 points in Figures 1 and 2 presents the results of either direct or minimal generalization tests pertinent to the percentage of the 17 target behaviors that the student manifested. Sessions indicated on the abscissa include the two pretests, the seven advocacy lessons (introduction, disclosure, solution, restatement, agreement, summary, and closure), and the two posttests.
In the pre-training direct tests, participants demonstrated from 12% to 41% of the 17 advocacy behaviors. Gradual increments in use of the 17 target skills are documented in Figure 1, to the point of 80% to 100% demonstration during the two posttest assessments. Person C was the only participant, who deviated slightly from a consistent increase in use of the skills across the lessons (e.g., note decrease in use of skills from lessons six to seven). Clearly, participants were acquiring and maintaining the target behaviors over time and were able to use them tin advocating for classroom accommodations in familiar situations (i.e., the English and algebra courses used in the role plays in each of the seven lessons).
Representing the results of the minimal generalization test, Figure 2 presents a similar pattern of increased use of the self-advocacy skills across time. Again, in the pretests, participants used about 18% to 35% of the target behaviors. Following each of the seven