本研究從認知風格與資訊量兩個層面對圖形符號判讀效率進行探討，首先以群組藏圖測驗將受測者之認知風格分爲「場獨立型」與「場依賴型」兩個類別，再由資訊理論爲圖形符號的資訊量加以定量，並利用集群分析分爲「高資訊量」與「低資訊量」兩個等級，最後以圖形符號問卷來檢測正確率與反應時間，從中歸納判讀模式之樣態特徵，其研究成果可提供圖形符號的評估依據與優化視覺解讀的設計參考，結論共歸納爲下述三點： 1.在圖形符號的判讀效率中，資訊量與認知風格間的交互作用未達顯著標準；就資訊量而言，不同等級間的判讀效率有明顯的差異存在，低資訊量圖形符號的反應時間耗費較少且擁有較高的正確率，而認知風格之間雖無明顯的差異存在，但在高資訊負載下，認知風格被視爲影響判讀效率的潛在變項。 2.低資訊量圖形符號慣於被視爲單一個體來判讀，而外形線索是面對陌生圖形符號時的優先判讀策略，故以外形意義爲主的低資訊量圖形符號，擁有較佳的判讀效率表現；指引意義爲主的低資訊量圖形符號，則需藉助適當的設計來導引判讀傾向。 3.在高資訊量圖形符號中，正常比例的人形肖像被視爲圖形符號意義的組織中心，是優先選擇的重點訊息，其輔助訊息多由服裝與動作姿態提供，判讀錯誤將導致圖形符號意義無法正確地建構，而象徵圖像的認知基礎也是影響判讀效率的重要因素。

This study investigates the recognition efficiency of graphical symbols by examining the information loads and cognitive styles. The cognitive styles of participants consist of two categories: F.I. style and F.D. style through GEFT. Information loads of graphical symbols are categorized into two levels: high information load and low information load, through the cluster analysis. Furthermore, a questionnaire is conducted for measuring accuracy and average time. The conclusions, which should be considered as an evaluation and optimization reference to graphical symbols, are as followed: 1. The interaction between cognitive styles and information loads of graphical symbols are not significantly different. The information loads of graphical symbols affect the recognition efficiency significantly since low information loads need less time to be recognized and is higher in accuracy. The recognition efficiency between cognitive styles are not significantly different. However, cognitive styles could be a main variable in higher information loads. 2. Graphical symbols of low information loads are recognized as individual entities; hence, the accuracy of recognition is high when graphical symbols exhibited iconic meanings, which is also a main recognition strategy. It needs design guidance to improve the recognition efficiency of graphical symbols of low information loads with indexical meanings. 3. When a proportioned human figure is presented in graphical symbols of high information loads, participants tend to describe the relationships between the figures and objects in the graphical symbols using third persons' point of views; in addition, movements and clothings are used as auxiliary information in aiding recognition. When symbolic meanings of graphic are unrecognizable, the recognition accuracy is affected significantly.