許庭嘉Hsu, Ting-Chia劉南岑Liu, Nan-Cen2019-09-032016-07-142019-09-032016http://etds.lib.ntnu.edu.tw/cgi-bin/gs32/gsweb.cgi?o=dstdcdr&s=id=%22G060371063H%22.&%22.id.&http://rportal.lib.ntnu.edu.tw:80/handle/20.500.12235/96408　　本研究使用眼動儀來紀錄網頁設計同儕互評的過程，將每一網頁依照框架設計與內容素材設定視覺興趣區(ROI)，共有49位視覺型風格大專生參與本實驗。本研究對所有受試者進行同儕互評態度調查，發現受測者對於同儕互評抱持正面態度與操作理解。從同儕互評量化成績與專家評鑑之間的比較，發現在較簡潔的高評價網頁中，同儕互評與專家評鑑有超過七成以上的一致性，然而在評分素材較多樣化的網頁可能造成專家和學生的評量一致性降低。搭配同儕互評的量化成績與同儕互評的態度進行比對分析，簡潔的網頁設計即清晰度較佳會更有助於對同儕互評抱持較低負面態度，提高視覺型學生同儕互評時的專注度，受測者之專注度與同儕互評的負面態度是呈現顯著負相關。進一步透過分析視覺型學習風格之前27%高評價評分者與後27%之低評價評分者於網頁同儕互評過程之眼動行為序列分析，發現在得到最高評價的網站中，高評價評分者與專家評鑑結果有較多一致及有效的眼動行為出現，顯示在該網站中低評價評分者反而是較不具有效度的同儕互評。反之，在得到最低評價的網站中，低評價評分者與專家評鑑過程有較多有效的眼動行為出現。從熱區圖總結顯示當視覺型學習風格學生有較專注的眼動行為時，同儕互評的結果會與專家評鑑的結果非常接近，因此由同儕互評過程中所測量的眼動數據確實可以為同儕互評結果的信效度提供的第三方的驗證。本研究也從眼動熱區圖分析中發現有規劃設計重點的網站會讓使用者更了解設計者欲傳遞的重點資訊，為網頁設計者提供眼動證據上的回饋。This study employed eye-tracking machine to record the process of peer assessment. This study divided each web page into several region of interests (ROI) based on the frame design and material in the content. Forty-nine undergraduate students with the visual learning style participated in the experiment. This study investigated the peer assessment attitudes of the participants and found that they possessed highly positive attitudes and understanding toward peer assessment. After comparing the results of peer assessments and the evaluation from experts, high consistency was happened when the design of the web page was concise; however, the consistency was decreased when the web page contained too much diverse material. After comparing the peer assessment attitudes of the participants and the web page design scores marked by them, the study found that the web page with concise design attracted the visual-style students’ attention more, and those students possessed lower negative attitude toward peer assessment, showing a significant negative correlation. In addition, the study further analyzed the visual-style students’ serial behaviors in the process of peer assessment for each web page design. After comparing the evidences of each student’s eye movements and his/her evaluation results, this study found that the students who grade higher or lower scores have different eye movements, and the fixations and behaviors of the assessors grading higher scores was highly consistent with that of the experts in the web site which was scored the best design, implying few assessors grading lower scores were relatively invalid peer assessment. On the contrary, the fixations and behaviors of the assessors grading lower scores were highly consistent with that of the experts in the web site which was scored the worst design. Consequently, from the hot spot evidences of eye fixations, when the students more concentrated on the peer assessment, their evaluated results would be more closed to the evaluated results of the expert. Finally, the study found that the hot spots of eye fixations were the same with the key points planed by the student designers of the website which scored the best, which provided the student designer additional important eye-tracking feedbacks from the peer assessment activities.同儕互評眼球追蹤視覺興趣區熱區圖行為序列peer assessmenteye trackingregion of interesthot spotsserial behavior