不同卡片分類法效益比較之研究

Card sorting has always been recognized as the most economical and effective method used to elicit users' cognition. Therefore, it has been applied as the most common method to design libraries web information architectures based on user centered concepts. However, according to Nielsen's study, to conduct card sorting experiments, the most appropriate number of participants is 15 people. Paul proposed that the traditional card sorting method spends more cost at the participant number and analysis time. Thus, he suggested applying the modified Delphi card sorting method to resolve the situation. In the modified Delphi card sorting method, there are about 9 (8~10) participants to conduct experiments in a linear fashion. Beside the seed participant, the rest were to revise the card sorting result made by the preceding participant. The final obtained result is regarded as the one of all of participants. However, Shieh and Wu put forward the refined modified Delphi card sorting method that is similar to the Delphi card sorting method. The refined modified Delphi card sorting method has a total of nine participants. But it limits rounds to at most 3. However, when applying card sorting methods to website construction, which one is appropriate? What are the differences in their effectiveness and cost? This study applied algorithm analysis method to explore the time cost differences between various card sorting methods and their performances in findability. We also conduct corresponding experiments to collect related data as evidence for analysis. The result can be as the reference for the future applications of card sorting methods.