초록
Evaluation plays an essential role in design activity. Many theorist have agreed that designing involves problem solving or decision making. In evaluation, designers attempt to determine the value of a particular proposal arrived at by synthesis. And the results of designing, the product, is evaluated twice, objectively and subjectively.Alternatives in $$\mu$ti-objective decision problems generally possess numerous attributes by which they can be described and compared. The evaluation factors include all attributes that have levels specified by quantitative and qualitative objectives. However since qualitiative factors are difficult to quantify as $$\mu$eral estmates, these factors have tended to be ignored without regard for their importance to human content. Therefore we need some study that convert qualitative attributes into quantitative scale values. Following to Thurstone' s Psychological scaling methods (The method of successive intervals), attribute values of TV set are assigned by rating scale methods. The method of successive intervals, like the method of equal-appearing intervals, requires but a single judgement from each subject for each statement to be scaled. It is, therefore, a convinient method to use when the number of to be scaled is large. We make the assumtion that those cu$$\mu$ative proportion distributions are normal for each statement when they are projected on the unknown psychological continuum. In this study, we have determined the scale values of 42 statements of TV set by the method of successive intervals. Then we can apply a test of internal consistency similar to that used with the method of paired comparisons. We have as our absolute average deviation, -1.748/252= -0.0069. We have reason to believe that our scale values are consistent with the empirical data, because these discrepanicies are very small.