the use of both video and business simulation would require substantial preparation, which it was felt would be remembered by our informants even though some time had passed since the courses were delivered. Therefore, in the absence of specific knowledge of the use of video or a business simulation it seemed wisest to assume that general English material were used.
In the Japanese data the complete range of qualifications of the instructors included general Bachelor of Arts, any Master's degree, certificate in teaching, or training in specific teaching materials. Of these, only BA and training in specific teaching materials occurred in the major sample. Since the training in specific teaching materials would require substantial preparation and would be remembered bu our informants, it seemed prudent to impute BA as the requirement when the requirement was not known.
No imputation was necessary for the purpose of the courses, but not all of the class sizes were known. For only two companies were class sizes known to be less than 10 or greater than 20; all the rest of the class sizes were known to be intermediate or were unknown. In this case, it was considered safest to impute an intermediate size to classes of unknown size.
Information on the training conditions affecting the Korean data was far more complete than for the Japanese data. Hence, No special imputation procedures were used for the Korean data.
With the imputations described above, the Japanese sample required a complete 3x2x3x2 complete design. BusSim, GenEng and video were the levels of the first factor, which was the materials factor. The codes for BA and PRG were the levels of the second factor, which was the instructor requirement. The course purpose factor included levels for GenEd, StaffDev. Mid and Large made up the fourth factor, which was class size.