基於不同任務表現之故宮典藏資料搜尋、學習與意義建構歷程之實徵研究

Abstract

本研究以國立故宮博物院「器物典藏資料檢索系統」與「書畫典藏資料檢索系統」作為研究個案，研究設計三項學習任務，進行故宮典藏系統的使用者學習行為研究與網站評估。研究目的旨在評估故宮器物、書畫檢索系統，透過不同型態之學習任務，觀察使用者在檢索過程中是否具有意義建構的過程，並探討其搜尋行為與策略，是否達到學習成效。研究以螢幕側錄使用者的滑鼠點擊行為，並搭配眼動儀器觀察使用者的凝視軌跡，分析其資訊需求之行為特徵。透過量化分析與質性訪談，經由滑鼠點擊與眼動追蹤了解使用者在系統使用的搜尋行為與其資源選擇之情況，搭配後測訪談詢問使用者對任務的感受與自我評估，進而探究使用者於典藏系統之搜尋歷程。簡要說明研究結果如下：（1）學習指標與專家評分之任務表現呈現正相關，顯示任務表現影響學習成效；（2）透過滑鼠點擊與眼動追蹤觀察使用者行為，開放式任務較封閉式任務具有更明顯的表現差異與學習跡象；（3）先備知識不完全影響受試者的任務表現，真正影響表現差異的關鍵因素還是看任務過程中的投入程度；（4）故宮系統資源的豐富性與良好的搜尋機制也會影響使用者的任務表現。此外，結果也證實搜尋即學習的概念，具備良好資訊搜尋能力的學習者，在資訊搜尋歷程與學習任務中皆有成功的表現。

In this study, we explore users’ search and learning behaviors on the digital archives for antiques search system and digital archives for panting and calligraphy search system of the National Palace Museum in Taiwan. We will invite users with basic knowledge in the fields of history or arts as the research target. We design three learning-related tasks to observe users’ learning behaviors while using one of the digital archives. We aim to examine whether the meaning-making occurs during the search process and analyze the users’ search behaviors with associated strategies to achieve learning outcome. In this study, we adopt screen recording software to collect the users’ click stream data and use eye-tracking equipment to observe users’ eye movement patterns, gaze duration, and gaze fixation count information which are related to their online searching and exploration behaviors for fulfill the information needs. We apply quantitative analysis and qualitative interview to understand the users’ search behavior and resource selection for accomplish the tasks by using the system in detail. That is, we collect data from clicks behaviors, eye movement tracking, and ask the users’ feelings about the task execution process with self-assessment, for investigating the relationship between their information search process with learning and meaning-making outcome. Finally, the research results can be used to understand the effectiveness of user’s learning on the system based on their information needs status and task types. We also aim to provide suggestions for improving the functions of the system.