學業情緒圖案系統的基礎研究：圖案編制、在女性高中生樣本之典型圖案篩選、部件凝視分析及與學業功能關聯

The Design, Selection and Component Fixation Analyses of Achievement Emotion Images and Their Relationships with The Academic Functions of Female High School Students

張家約(Chia-Yueh Chang)

指導教授 ： 林珊如

國立交通大學/人文社會學院/教育研究所/碩士(2017年)

摘要

過去研究指出，學生的學業情緒和學習動機、努力、成就目標、學習策略和成就等關係密切 (Pekrun, Goetz, Titz, & Perry, 2002)。然而目前教學現場對評估學生情緒的方式選擇不多，最常用來評估學生學業情緒的方法是老師自身的教學經驗，或花額外的時間讓學生填寫問卷。然而通常填寫問卷的時間冗長以至於學生填答意願低，且在改編國外問卷時，情緒用詞翻譯有極大的選詞困難，如「relief」的中文意譯在心理學界常見的是「解脫」，但對當代台灣學生而言，如果翻譯成「放鬆」會較貼近其感受、易於理解，這樣的例子不勝枚舉。近年來人們使用圖形來表達情緒的機會愈發增加，Line及Facebook等社群網站推出的表情符號貼圖(emoticons)功能方便使用者發布情緒；同時，情緒圖案具有跨文化的特色，透過圖案便能表達情緒。
本研究根據 Pekrun等人 (2002)學業情緒的理論，將學業情緒歸納成九種，分別為愉快、希望、自豪、放鬆、生氣、焦慮、羞恥、無望、無聊，針對這九種情緒設計表情圖像，每一個情緒設計5張圖案，共45張情緒圖案，首先，受試者先以電腦問卷評定每張情緒圖案，目的為篩選出九張典型學業情緒圖案。再以眼動實驗對九種情緒的典型圖案進行情緒辨識作業，使用眼動追蹤技術 (Eye Link 1000)蒐集並分析資料，目的為探索閱讀情緒圖案時，表情部件的重要性對比。最後，本研究編制「數學科學業情緒量表-圖案版」，並分析數學科學業情緒分數與學業功能 (數學自我效能及數學成就)之間的相關。
本研究邀請205位高中一年級女生參與研究，其中52位參加眼動實驗。首先，為篩選出典型學業情緒圖案，每張圖案都有兩項指標作為判別依準：強度高 (情緒圖案對應單一情緒形容詞的適配度強)及熵值低 (某一情緒圖案與某一情緒形容詞適配評定在受試者間的一致性高)。其次，使用眼動技術分析對比情緒圖案的四個興趣區 (眼睛、嘴巴、肢體動作及背景圖樣)的重要性，收集凝視次數 (fixation count)及凝視時間 (dwell time)兩個眼動指標，採納相依樣本二因子變異數分析分析眼動資料。研究結果如下。
1. 從45張情緒圖案能成功地篩選出9張典型學業情緒圖案，並發現正向情緒圖案(希望、自豪、放鬆)容易與愉快圖案混淆，焦慮圖案跟羞恥圖案較不易區分，無望圖案跟無聊圖案也較不易分辨。這可能是愉快與正向情緒、焦慮與羞恥表情、無望與無聊表情不易區分，亦可能是有些情緒圖案人物的表情或姿勢相像所致。
2. 將九種典型情緒圖案分為: 正向與負向情緒類、活化與鈍化情緒類、基本情緒與其他類。無論哪一類圖案，受試者辨識情緒圖案的注意力主要都集中於眼睛跟嘴巴，而肢體動作和背景圖樣被凝視的次數及時間則較少。此結果與過去研究相符，受試者辨識情緒時臉孔表情比肢體更重要 (Beaudry, Roy-Charland, Perron, & Tapp, 2014; Eisenbarth & Alpers, 2011; Shields, Engelhardt, & Ietswaart, 2012)。
3. 針對九張典型學業情緒圖案在各興趣區的眼動型態，不同情緒圖案的凝視時間跟凝視次數結果有不一樣的模式：希望跟羞恥圖案凝視眼睛的次數最多次，自豪圖案凝視眼睛的時間最久。生氣圖案凝視嘴巴的時間最多次也最久，顯示生氣的嘴巴圖案對判別很重要；羞恥圖案凝視嘴巴的次數最少，但凝視嘴巴的時間最長，在背景圖樣上的凝視時間最短。焦慮圖案凝視肢體的次數多，凝視時間也長；無望圖案凝視肢體的次數較多，但凝視時間較短，顯示焦慮跟無望圖案或許需要肢體動作來輔助判斷情緒。放鬆圖案凝視背景的次數較多，凝視時間也較長，可見看放鬆情緒圖案時，背景圖樣似乎輔助判斷。這反映出每個情緒圖案都有其獨特的特色。
4. 把九張典型學業情緒圖案應用於數學科，結果顯示數學科學業情緒量表圖案版具有良好的信度。此外，所有數學學習情境中的正向情緒圖案評分皆與數學自我效能和數學成績顯著正相關，與過去研究相符 (Pekrun et al., 2002; Valiente, Swanson, & Eisenberg, 2012)，顯示本量表具有適當的效度。
本研究設計了學業情緒圖案系統，篩選出九個典型學業情緒圖案來分類學習情境中學生的情感，教師可以以一種有趣的方式使用它來開發學習材料和貼圖，讓教師和學生能更直接自然的溝通。此外本研究旨在提供一個良好的量測學業情緒的工具給第一線的教學者，期望能提升情緒表達時的準確度，降低情緒表達時的選詞困難。

關鍵字

學業情緒 ； 學習動機 ； 情緒圖案 ； 熵值 ； 眼動 ； 凝視分析 ； 高中女生

並列摘要

Studies indicated that academic emotions are closely related to learning motivation, effort, achievement goal, learning strategy, and achievement (Pekrun et al., 2002). However, pragmatic methods to assess students’ emotion are rare in the context of teaching. Some common ways of assessing students' academic emotion are teacher observation/rating or student self-report. However, it took lot of time to finish a survey, which might reduce students’ willingness to respond. Additionally, translating English terms of academic emtion to Chinese is very difficult because several Chinese terms could reveal subtle different perspects of an English emtion usgage. For example, "relief" usually been translated into "解脫" by Chinese psychologists, but an alternative tranlastion of "放鬆" might better fit the common usage of Taiwan adolescents in school settings. Thus in this study, the author tried to take an innovative way to convey academic emotions by replacing Chinese phrases with emotion images (emoticons). Nowadays people often use emoticons to express emotion. Social media, such as Line and Facebook, provides emoticon-stickers to help users express their emotions. The meaning of an emotion image is often compatible or consistent across cultures; therefore people from diffent language systems can express emotions through images.
This study designed achievement emotion images for nine emotions (enjoyment, hope, pride, relief, anger, anxiety, shame, hopelessness and boredom) identified by Pekrun’s Achievement Emotion theory (2002). For each emotion, 5 images were drawn and evaluated. Digital questionnaires and eye tracking technique were used in this study. First, participants used digital questionnaires to evaluate each achievement emotion images; the purpose was to select 9 typical emotion images out from 45. Then Eye Link 1000 was used to collect eye movement data (i.e., fixation duration and counts) while emotion recognition and the goal was to compare differneces of attention placed on four components/AOIs in emotional images (eye, mouth, gesture and decoration). The 9 typical emotion images were used to form the “Acadecmic Emotion Scale for Math Learning– Emoticon version (AESMAL-E)” and an examination of reliability and validity was conducted. The relationship between score of AESE and academic functions (mathematical self-efficacy and mathematical achievement) in math were examined.
Two hundred and five 10th grade female students were invited to participate, and among them the eye movement data of 52 students were recorded. In selecting a typical academic emotion image, each image was evaluated by two criteria: level of strength (An emotion image highly corresponds to one and only one emotion adjective) and level of entropy (High group consistency on correspondence rating between an emotion image and an adjective). For each academic emotion, one out from five images was selected as the typical if it had the highest level of strength and lowest level of entrophy. The eye movement technique was adopted to identify the relative importance of 4 AOIs. Fixation count and dwell time data were analyzed by conducting two-way repeated measure analysis of variance. The results are as follows.
1. Nine achievement emotion images selected from 45 images were verified as typical achievement emotion images with high strength and low entrophy. The students inclined to confuse the enjoyment images with other positive emotion images (i.e., hope, pride, relief). It was also difficult to distinguish between images of anxiety and shame and between images of hopelessness and boredom. The aforementioned could result from that it was difficult to distinguish between emotions, such as enjoyment versus positive emotion, anxiety versus shame expression, hopelessness versus boredom. The other reason might be that some emotion images are similiar in facial expressions or gestures.
2. Nine academic emotion images could be categorized as positive versus negative, activating versus deactivating, and basic versus other emotion images. For most images, participants fixated on AOIs of eyes and mouth; on the other hand, gesture and decoration drew less attention. These results are consistent with previous studies showing that "facial expressions" are more important than gesture when the participants recognize emotions (Beaudry et al., 2014; Eisenbarth & Alpers, 2011; Shields et al., 2012).
3. RM-ANOVA showed that for each AOI, the fixation counts and dwell time of each emotion images have different patterns: Among nine emotion images, the image of hope and shame received most fixation counts on the component ‘eyes’ while the image of pride received longest fixation duration. The image of anger received most fixation counts and longer fixation duration on the component ‘mouth’, showing the importance of mouth in identifyingthe emotion of anger. For the image of shame, participants had less fixation counts but longer fixation duration than the other images while gazing at mouth. Moreover, participants had shorter fixation duration while gazing at the AOI of decoration. For the image of anxiety, the AOI of gesture received more fixation counts and longer fixation than other images. For the image of hopelessness, the AOI of gesture received more fixation counts but shorter fixation duration than other images, showing that gesture might be benefical in recognizing anxiety, as well as hopelessness. For the image of relief, the AOI of decoration received more fixation counts and longer fixation duration, showing that decoration may be helpful in recognizing the emotion of relief. In summary, the results reflected that each image might have distinct characteristics.
4. The AESMAL-E scale showed quality internal reliability. Futhermore, all positive emotion images were positively correlated with math achievement and self-efficacy, which is consistent with previous studies (Pekrun et al., 2002; Valiente et al., 2012) showing acceptable validity.
This study designed nine achievement emotion images to classify students’ emotion. The images could be used by teachers to develop learning materials and stickers that provide natural way of communication about academic emotion between teachers and students. Also, this study provided an alternative and interesting tool of measuring emotions in learning settings.

並列關鍵字

Achievement emotions ； learning motivation ； emoticon ； entropy ； eye movement ； fixation analyses ； female high school students

參考文獻

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs: NJ: Prentice-Hall.

Beaudry, O., Roy-Charland, A., Perron, M., & Tapp, R. (2014). Featural processing in recognition of emotional facial expressions. Cognition and Emotion, 28(3), 416-432. doi: 10.1080/02699931.2013.833500

Beedie, C., Terry, P., & Lane, A. (2005). Distinctions between emotion and mood. Cognition and Emotion, 19(6), 847-878. doi: 10.1080/02699930541000057

Bristor, V. J., & Drake, S. V. (1994). Linking the language arts and content areas through visual technology. Technological Horizons In Education, 22(2), 74.

Calvo, M. G., & Nummenmaa, L. (2008). Detection of emotional faces: Salient physical features guide effective visual search. Journal of Experimental Psychology-General, 137(3), 471-494. doi: 10.1037/a0012771