This paper describes the non-standard method that was used to evaluate the various processes and procedures employed by more than 20 PLATO courseware development groups and to evaluate this method. The method used was the Flanagan's critical incident technique, modified to include decisions as well as incidents. Sources of information for the evaluation included personal accounts of a number of staff members, interviews with other staff members, and papers and reports. More than 125 case histories of critical incidents or decisions were classified, and the processes and procedures that led to successful, unsuccessful, or equivocal results were analyzed. Discussion of this evaluation method includes its limitations, time requirements, advantages, suggestions for improvement, and its utilization. (VT)