改進卡片分類法以提升網站尋獲度之研究

Abstract

網站設計者最困擾的問題之一在於如何提高網站的內部尋獲度，亦即一網站讓使用者找到其所需資訊之能力；過去研究顯示，有效之資訊架構有助於網站內部尋獲度與使用性之提昇。在以使用者為中心的網站設計概念下，卡片分類法（card sorting）一直是建構較佳尋獲度與使用性網站經濟、有效的工具。Paul在2007年提出「修正型德菲式卡片分類法」，提供前位受試者之分類結果予受試者，經提出回饋後再將其分類結果提供予後一位受試者，最終為多位受試者一致意見之結果。然而前位受試者的背景因素及其所分類之結果，是否對接續受試者的網站內容分類認知會造成影響，進而導致最終網站架構的差異？本研究計畫將受試者之學科背景、職業身份別、網站使用經驗等因素，考量於「修正型德菲式卡片分類法」實驗中，以探討受試者因素對於卡片分類結果及爾後網站尋獲度與使用性之影響。 利用卡片分類法進行網站資訊架構分析實驗時，受試者的分類認知會受環境、時間等其他因素影響，如果受試者能於事後瞭解網站項目的意涵，並再對分類內容作部分調整與確認自己的認知結果，或許能得到更符合個人認知的分類結果。因此，為更深入瞭解受試者對於每一網站架構項目之間的相關程度，本研究計畫預計透過相似性評定(similarity rating)作為確認受試者想法之工具。藉由相似性評定後之受試者網站內容分類結果，對於所建構網站之尋獲度與使用性是否與傳統卡片分類所建構之網站有差異存在，亦為本研究計畫探討的議題。 過去針對網站分類的認知資料分析多半以群集分析、因素分析為主，藉由分析結果建構符合使用者分類認知之網站架構。然而徑路搜尋分析是能以較客觀的量測方式表徵知識結構，本研究計畫欲嘗試以不同於因素分析、群集分析的知識結構表徵工具-- 徑路搜尋量尺，藉由使用者於網站項目之分類認知資料來建構網站架構，並探討其網站尋獲度、使用性之效益，以實證此工具適於改善或建構網站架構。

One of the most difficult issues to website designers is to enhance the internal findability of websites. That is to refine the ability of websites for users to locate what they need. Past researches show that well information architectures help websites increasing their internal findability and usability. Based on the concept of user-centered design, card sorting has been recognized as an economical and effective tool to construct the websites with better findability and usability. In 2007, Paul proposed the method named “refined modified-Delphi card sorting” which the final test result was the consensus of a group of participants. Participants give their feedbacks based on the former participant’s, and then pass their own result to next one. What remains questionable is that would the latter participant’s recognition about website items affected by the former participant’s disciplines or experiences. And finally it will lead to the differences between websites constructions. This research project will discover what of these factors will have impacts on the classifying results and the effectiveness of website findability and usability. Cluster analysis or factor analysis are used to be the methods in analyzing precognitive information of website classification and constructing website structures according to users’ classification Within the test of website’s information architecture analysis by card sorting, participants’ classification cognitions would be affected by the ambiance, time or other factors. Test results would be more fit to personal cognitions of classification if the test takers recognize the meanings of each website items after the test, and then revise their own classified results. Thus, in order to understand test takers’ recognitions of relationships between website items, we can use similarity rating to assist in and further confirm participants’ thoughts. After applying similarity rating to participants, this research project will explore that whether the card sorting results will generate websites with better performance in their findability and usability. cognitions. However, the path-finding method is a more objective method to represent users’ knowledge structures. In this study, we will use path-finding scaling method to construct website structures with users’ classification cognitions. Also, researchers discuss the findability and usability of the constructed websites to prove the method is suitable for refining website architectures.