• 沒有找到結果。

使用情感運算技術評估數位遊戲式學習之成效

N/A
N/A
Protected

Academic year: 2021

Share "使用情感運算技術評估數位遊戲式學習之成效"

Copied!
144
0
0

加載中.... (立即查看全文)

全文

(1)國立臺中教育大學數位內容科技學系碩士班 碩士論文 指導教授: 吳智鴻 博士. 使用情感運算技術評估 數位遊戲式學習之成效 Using the Affective Computing Technique to Evaluate the Outcome of Digital Game-Based Learning. 研究生: 曾奕霖. 撰. 中 華 民 國 ㄧ百零二 年 六 月.

(2)

(3) 致. 謝. 三年來的研究所求學生活,一路走來相當踏實,。今日得以順利完成碩士學位,要歸 功於這一路協助、鼓勵我的貴人。. 首先要感謝的是我的指導教授---吳智鴻教授,在研究所三年過程中,老師對學術的專 業以及繁忙業務中仍撥空不辭辛勞的關心與指導,讓我在研究的過程中,能掌握正確的方 向,啟發思考的觀點,教授所提醒教誨的每一個學術環節,在貫徹執行之後,往往能體會 到更深一層的涵意與價值。. 其次要感謝黃悅民教授、林豪鏘教授及王怡舜教授等口試委員,提供了相當精闢的見 解,使論文的內容架構能夠更周延、縝密。老師的真知灼見,讓我看見自己思維的缺陷, 寶貴的意見和修正方向,使得論文更臻完善,使我獲益良多. 再來感謝在研究所求學期間,長儒學長、歐陽還有班上所有同學的協助支持,以及精 神上的關心與鼓勵,讓我能完成研究過程中的種種挑戰。另外還有感謝協助實驗進行的認 識或不認識的朋友,讓本研究得以順利完成。. 許多曾經幫助及鼓勵我的人,無法一一提及,在此致上最誠摯的謝意,謹以本論文夾 帶著說不盡的感激、感動與喜悅,獻給所有關心我、幫助我完成碩士學位的貴人,因為有 你們,所以成就了我。. 無盡感激,難以言表。. 曾奕霖 謹誌於 國立臺中教育大學 數位內容科技學系 中華民國 102 年 7 月.

(4) 使用情感運算技術評估數位遊戲式學習之成效. 國立臺中教育大學數位內容科技學系. 研究生:曾奕霖. 指導教授:吳智鴻博士. 中文摘要 數位遊戲式學習被認為一種有效的學習工具,但是文獻中提供的證據仍是有限甚至有 矛盾的看法。為了驗證用數位遊戲式學習學習牛頓運動定律的可能性,本研究使用準實驗 設計法比較數位遊戲式學習與傳統靜態數位教材在學習動機、學習情感體驗、學習成效及 問題解決能力上的差異。在第一階段的實驗中,使用情感運算技術測量學生學習時的生理 訊號,並分析他們的學習狀態,而在學習階段結束後,根據前測及後測之成績評估學生學 習成效。在第二階段的實驗中,本研究探討擁有不同工作記憶能力之學習風格的學生(行動 型/思考型,感官型/直覺型,循序思考型/全盤思考型)、不同學習環境(數位遊戲式學習/傳 統靜態數位教材)、不同主修科目(非理工科/理工科)和不同解決問題能力的學生(高分組/低 分組)在解決物理問題時思考策略上的差異。 從實驗結果得到五項發現,第一,數位遊戲式學習的組別從情威運算技術的證據上顯 示擁有較高的認知負荷;第二,數位遊戲式學習的組別雖然有比較高的學習成就,但與傳 統教學組相比沒有顯著的差異;第三,高工作記憶能力的學習風格組、數位遊戲式學習組 以及解決問題能力的高分組比較知道去搜尋題目中的解題關鍵因素;第四,成功解決問題 的學生和失敗的學生相比,擁有不同的視覺思考模式。第五,可以透過學習者解題時所偵. I.

(5) 測到的眼動變數做為區分學習風格(行動型/思考型)的依據。 本研究認為,教育者需要提供足夠的數位遊戲式學習環境的條件,才能充份的發揮數 位遊戲式學習的優勢,另外,我們發現教育者可以透過情感運算技術來測量學生的學習風 格,這對於適性化學習系統的開發將有很大的幫助。關於給後續研究的建議,我們認為未 來的研究可以增加更多的生理訊號來幫助測量學習者的學習狀態,另外,還可以提供更完 善的數位遊戲式學習環境,提昇學習的效能。. 關鍵字:情感運算、眼動、腦波、心跳一致性、數位遊戲式學習、序列分析. II.

(6) Using the Affective Computing Technique to Evaluate the Performance of Digital Game-Based Learning. Department of Digital Content and Technology National Taichung University of Education. Student: Yi-Lin Tzeng. Advisor: Dr. Chih-Hung Wu. ABSTRACT Digital Game-Based Learning (DGBL) is thought to be an effective tool for learning, but the empirical evidence to support this assumption is still limited and contradictory. For verifying the possibility of playing digital game to learn the Newton's laws of motion, this study used a quasi-experimental design to examine the effectiveness of Digital Game-Based Learning (DGBL) and traditional static e-learning on students’ learning attention, affective experiences, cognitive load, academic achievement and problem solving skills. In phase 1, when student learning, their physiology signals were measured by affective computing technique for analyzing their learning states. After learning, a posttest of learners was conducted to find the differences in academic achievement between DGBL and static e-learning. In phase 2, this study found that learner had difference problem solving strategies between different working memory capacity learning styles (active vs. reflective, sensing vs. intuitive, sequential vs. global), learning environments (DGBL vs. traditional static e-learning), major (non-science vs. science) and problem-solving performance (low performance vs. high performance). The results showed five major findings. The first finding, the DGBL group has more. III.

(7) cognitive load via proof of affective computing technique. The second finding, the DGBL group has better academic achievement but have no significant differences between DGBL and static group. The third finding, the high working memory capacity learning style group, the DGBL group and high problem-solving group easily to find out the key factors. The fourth finding, Successful problem solvers inspected the factors in a different pattern from unsuccessful problem solvers. The fifth finding, learning style (active vs. reflective) showed significant various ways of eye movement when learners solving problem. This study suggested future educators can provide the enough feature of DGBL environment. In addition, we found that educator can measure leaners’ learning style by affective computing technique. This is helpful for developing an adaptive learning system. In the future work, we propose the future studies can adopt more physiology signals to measure learning state of learners and provide better DGBL environment.. Keyword: Affective Computing, eye movement, brain wave, heart rhythm coherence, Digital Game-Based Learning, Sequential Analysis.. IV.

(8) CONTENT. 摘 要…………………………………………………………………………………………….Ⅰ ABSTRACT…………………………………….……………………………………….……….Ⅲ CONTENT……………………………………………………………………………….………Ⅴ LIST OF TABLES……………………………………………………………………………….Ⅷ LIST OF FIGURES…………………………..……………………………………………….…Ⅹ CHAPTER Ⅰ INTRODUCTION .............................................................................................................. 1 1.1 Research motivation ........................................................................................................................... 1 1.2 Research purposes .............................................................................................................................. 5 1.3 Research gaps ..................................................................................................................................... 6 CHAPTER Ⅱ LITERATURE REVIEW ................................................................................................... 8 2.1 Digital Game-Based Learning ............................................................................................................ 8 2.1.1 The advantages of Digital Game-Based Learning ....................................................................... 9 2.1.2 Learning motivation, affective experiences, cognitive load and Digital Game-Based Learning ............................................................................................................................................................ 10 2.1.3 Academic achievement and Digital Game-Based Learning ...................................................... 11 2.1.4 Gender and Digital Game-Based Learning................................................................................ 12 2.2. Problem solving strategy ................................................................................................................. 13 2.3. Affective Computing in Learning.................................................................................................... 15 2.3.1 Eye movement variables ............................................................................................................ 20 2.3.2 Brain wave variables ................................................................................................................. 22 2.3.3 Heart rate variables .................................................................................................................... 23 Chapter Ⅲ METHOD .............................................................................................................................. 26 3.1. Research hypotheses ........................................................................................................................ 26 3.2. Participants ...................................................................................................................................... 33 3.3. Materials .......................................................................................................................................... 35 3.3.1 Pre-test ....................................................................................................................................... 35 3.3.2 Felder–Silverman learning style questionnaire ......................................................................... 36 3.3.3 Digital Game-Based Learning materials ................................................................................... 38 3.3.4 Static e-learning materials ......................................................................................................... 40 3.3.5 Post-test materials ...................................................................................................................... 40. V.

(9) 3.4. Procedure ......................................................................................................................................... 43 3.5. Apparatus......................................................................................................................................... 45 3.5.1 Eye Tracker ............................................................................................................................... 45 3.5.2 emWave ..................................................................................................................................... 46 3.5.3 Neurosky.................................................................................................................................... 47 3.6. Physiology signals analysis ............................................................................................................. 48 3.6.1 Eye movement variables ............................................................................................................ 48 3.6.2 Emotion variables ...................................................................................................................... 50 3.6.3 Brain wave variables ................................................................................................................. 52 3.6.4 Data analysis .............................................................................................................................. 52 Chapter Ⅳ DATA ANALYSIS ................................................................................................................ 58 4.1 DGBL group have more cognitive load than traditional static e-learning group ............................. 58 4.1.1 Physiology signals representation for learning motivation, affective experiences and cognitive load ..................................................................................................................................................... 58 4.1.2 Finding....................................................................................................................................... 59 4.2 DGBL group have better academic achievement but have no significant differences ..................... 61 4.3 High working memory capacity learning style group knew better where to look the key factors. .. 63 4.3.1 Active vs. reflective ................................................................................................................... 63 4.3.2 Sensing vs. intuitive................................................................................................................... 67 4.3.3 Sequential vs. global .................................................................................................................. 70 4.3.4 Finding....................................................................................................................................... 73 4.4 DGBL group knew better where to look the key factors .................................................................. 74 4.4.1 Attention distributions on each LookZone ................................................................................ 74 4.4.2 Hot Zone image ......................................................................................................................... 77 4.4.2 Finding....................................................................................................................................... 77 4.5 Science major group couldn’t know better where to look the key factors, but Non-science majors’ learners need more clues to solve problem ............................................................................................. 79 4.5.1 Attention distributions on each LookZone ................................................................................ 79 4.5.2 Hot Zone image ......................................................................................................................... 82 4.5.3 Finding....................................................................................................................................... 82 4.6 Successful problem solvers are able to recognize and concentrate on relevant cues ....................... 84 4.6.1 Attention distributions on each LookZone ................................................................................ 84 4.6.2 Hot zone images ........................................................................................................................ 86 4.6.3 Finding....................................................................................................................................... 86. VI.

(10) 4.7 Successful problem solvers inspected the factors in a different pattern from unsuccessful problem solvers. .................................................................................................................................................... 88 4.7.1 Sequential analysis .................................................................................................................... 88 4.7.2 Finding....................................................................................................................................... 91 4.8 Learning style couldn’t be measure by physiology signals when learners learning ......................... 92 4.9 Correlation analysis of physiology signals ..................................................................................... 100 4.9.1 Total Correlation...................................................................................................................... 100 4.9.2 Correlation analysis divide into static e-learning and DGBL .................................................. 100 4.10 Summary of hypotheses verified .................................................................................................. 102 4.11 Discussion..................................................................................................................................... 103 Chapter Ⅴ CONCLUSION AND IMPLICATIONS ............................................................................. 109 5.1 Conclusion ...................................................................................................................................... 109 5.2 Implications .................................................................................................................................... 111 5.2.1 The necessary feature of DGBL environment ......................................................................... 111 5.1.2 The relationship between affective computing and learning style .......................................... 112 5.3 Future Study ................................................................................................................................... 112 Reference .................................................................................................................................................. 114 Appendix I ................................................................................................................................................ 123 Appendix Ⅱ ............................................................................................................................................ 128. VII.

(11) LIST OF TABLES Table 1.Multi physiological feature system review ...................................................................... 18 Table 2.Eye movement parameters ............................................................................................... 21 Table 3.Emotion and attention state in EEG ................................................................................. 23 Table 4.Power spectrum components............................................................................................ 24 Table 5.Emotion and attention state in HRV ................................................................................. 25 Table 6.Hypotheses and reference in PHASE 1 ............................................................................ 27 Table 7.Hypotheses and reference in PHASE 2 ............................................................................ 29 Table 8.The relationship between learning styles, working memory and learning performance . 31 Table 9.Descriptive statistics......................................................................................................... 34 Table 10.Definitions for the eye-movement measures .................................................................. 49 Table 11.Physiology signals variables in PHASE 1 ...................................................................... 53 Table 12.Physiology signals variables in PHASE 2 ..................................................................... 56 Table 13.ANOVA Analysis of physiology signals between Static and DGBL Learner ................ 59 Table 14.Analysis of covariance for academic achievement ........................................................ 62 Table 15.Visual attention distributions between active and reflective group (problem 1)............ 64 Table 16.Visual attention distributions between active and reflective group (problem 2)............ 65 Table 17.Visual attention distributions between sensing and intuitive group (problem 1) ........... 67 Table 18.Visual attention distributions between sensing and intuitive group (problem 2) ........... 68 Table 19.Visual attention distributions between sequential and global group (problem 1) .......... 70 Table 20.Visual attention distributions between sequential and global group (problem 2) .......... 71 Table 21.Visual attention distributions differences between two WMC groups. .......................... 73 Table 22.Visual attention distributions between two learning environment (problem 1) ............. 75 Table 23.Visual attention distributions between two learning environment (problem 2) ............. 76 Table 24.Visual attention distributions differences between two learning environment .............. 78 Table 25.Visual attention distributions between Non-science and Science major (problem 1) .... 80 Table 26.Visual attention distributions between Non-science and Science learners (problem 2) 81 Table 27.Visual attention distributions differences between two learners’ major......................... 83 Table 28.Visual attention distributions between Low and High performance (problem 1). ......... 84 Table 29.Visual attention distributions between Low and High performance (problem 2) .......... 85 Table 30.Visual attention distributions differences between two learners’ performance .............. 87. VIII.

(12) Table 31.Adjusted residuals (z scores) for high and low performance group (problem 1) ........... 89 Table 32.Adjusted residuals (z scores) for high performance group (problem 2). ....................... 90 Table 33.Adjusted residuals (z scores) for low performance group (problem 2).......................... 90 Table 34.t-test of physiology signals between active and reflective group (static e-learning) ..... 94 Table 35.t-test of physiology signals between active and reflective group (DGBL) .................... 94 Table 36.t-test of physiology signals between sensing and intuitive group (static e-learning)..... 95 Table 37.t-test of physiology signals between sensing and intuitive group (DGBL) ................... 95 Table 38.t-test of physiology signals between sequential and global group (static e-learning) .... 96 Table 39.t-test of physiology signals between sequential and global group (DGBL) .................. 96 Table 40.t-test of physiology signals between non-science and science group (static e-learning)97 Table 41.t-test of physiology signals between non-science and science group (DGBL) .............. 97 Table 42.t-test of physiology signals between P1 low and high performance group (static) ........ 98 Table 43.t-test of physiology signals between P1 low and high performance group (DGBL) ..... 98 Table 44.t-test of physiology signals between P2 low and high performance group (static)........ 99 Table 45.t-test of physiology signals between P2 low and high performance group (DGBL) ..... 99 Table 46.Person Correlation Matrix ............................................................................................ 100 Table 47.Person Correlation Matrix for static e-learning group ................................................. 101 Table 48.Person Correlation Matrix for DGBL group ................................................................ 101 Table 49.Summary of hypotheses verified .................................................................................. 102 Table 50.physiology signals significant difference during learning between each group .......... 107 Table 51.physiology signals significant difference when solving problem between each group 108. IX.

(13) LIST OF FIGURES Figure 1.Research framework in PHASE 1 .................................................................................. 27 Figure 2.Research framework in PHASE 2 .................................................................................. 30 Figure 3.Basic information for participants .................................................................................. 34 Figure 4.Learning style classification for participants .................................................................. 35 Figure 5.Post-test performance for participants ............................................................................ 35 Figure 6.Easy level of the SURGE game ...................................................................................... 39 Figure 7.Hard level of the SURGE game ..................................................................................... 39 Figure 8.Static e-learning materials .............................................................................................. 40 Figure 9.Post-test problem 1 (easy) .............................................................................................. 42 Figure 10.Post-test problem 2 (difficult)....................................................................................... 42 Figure 11.Participant attached physiological sensors ................................................................... 43 Figure 12.Eye tracker, NeuroSky and emWave had signal input .................................................. 44 Figure 13.The procedure of experiment ........................................................................................ 45 Figure 14.emWave hardware ........................................................................................................ 46 Figure 15.emWave interface ......................................................................................................... 47 Figure 16.Neurosky hardware ....................................................................................................... 47 Figure 17.The definition scheme for LookZones in problem1 ..................................................... 55 Figure 18.The definition scheme for LookZones in problem 2 .................................................... 55 Figure 19.The pre-test and post-test measurement between two learning environment ............... 62 Figure 20.Significant differences LookZone between active and reflective group (problem 1). .. 64 Figure 21.Significant differences LookZone between active and reflective group (problem 2). .. 65 Figure 22.HotZone image between active and reflective group (problem 1) ............................... 66 Figure 23.HotZone image for active and reflective (problem 2) .................................................. 66 Figure 24 Significant differences LookZone between sensing and intuitive group (problem 2). . 68 Figure 25.HotZone image between sensing and intuitive group (problem 1) .............................. 69 Figure 26.HotZone image between sensing and intuitive group (problem 2) .............................. 69 Figure 27.HotZone image between sequential and global group (problem 1).............................. 72 Figure 28.HotZone image between sequential and global group (problem 2).............................. 72 Figure 29.Significant differences LookZone between DGBL and Static e-learning (problem 1). 75 Figure 30 Significant differences LookZone between DGBL and Static e-learning (problem 2). 76. X.

(14) Figure 31.HotZone image for learning environment (problem 1) ................................................ 77 Figure 32.HotZone image for learning environment (problem 2) ................................................ 77 Figure 33.Significant differences LookZone between Non-science and Science (problem 1) ..... 80 Figure 34 Significant differences LookZone between Non-science and Science (problem 2) ..... 81 Figure 35.HotZone image between Non-science and Science learners (problem1) ..................... 82 Figure 36.HotZone image between Non-science and Science learners (problem 2) .................... 82 Figure 37.Significant differences LookZone between Low and High performance (problem 2) . 85 Figure 38.HotZone image between Low and High performance learners (problem1) ................. 86 Figure 39.HotZone image between Low and High performance learners (problem 2) ................ 86 Figure 40.Sequential analyses between high and low performance group (problem 1) ............... 89 Figure 41.Sequential analyses between high and low performance group (problem 2) ............... 90 Figure 42. Academic achievement between learning environment and learning style (Active vs. Reflective) ............................................................................................................. 92 Figure 43. Academic achievement between learning environment and learning style (Sensing vs. Intuitive) ................................................................................................................ 92 Figure 44. Academic achievement between learning environment and learning style (Sequential vs. Global) ............................................................................................................. 93 Figure 45. Academic achievement between learning environment and learners’ major (Non-science vs. Science) ..................................................................................... 93. XI.

(15) CHAPTER Ⅰ INTRODUCTION In section 1.1, this study searched the papers for Digital Game-Based Learning and summarized some issue may be able to research by affective computing technique. In section 1.2, according to section 1.1, research questions were explored for verifying the possibility of playing digital game to learn the physics problem. In addition, 2 research gaps were found and described in section 1.3. 1.1 Research motivation Digital Game-Based Learning (DGBL) is thought to be an effective tool for learning (Kebritchi & Hirumi, 2008) that can promote enhanced learning experiences (T. M. Connolly, Stansfield, & Hainey, 2007) and student motivation (Papastergiou, 2009). DGBL can be defined as “the use of a computer game-based approach to deliver, support, and enhance teaching, learning, assessment, and evaluation” (T. M. Connolly et al., 2007). There is also widespread acknowledgment of the advantages that the use of games has in elementary and secondary education (Ebner & Holzinger, 2007). Games that encompass educational objectives and subject matter are believed to hold the potential to render learning of academic subjects more learner-centered, easier, more enjoyable, and more interesting. Although games are believed to be motivational and educationally effective, the empirical evidence to support this assumption is still limited and contradictory (Marina, 2009). The education games and commercial games are different. Many popular commercial games offer interesting pedagogical opportunities for physics education with their focus on physics-based problem solving that involves careful manipulation force and motion. Specific titles of note in this genre have included, for example, Angry Bird. While these commercial. 1.

(16) physics games provide students with a strong intuitive ‘feel’ for physics concepts, they don’t appear to (and were not designed to) help students make the leap from tacit understanding to more formalized knowledge (Clark et al., 2011). Game-based experiences thus appear to require scaffolding in order for students to make the connections between the game and the more formalized knowledge required in a school-based context. These findings suggest that simply having players engage with physics based games is not sufficient to help them learn physics. Many studies have investigated the effects of Digital Game-Based Learning (DGBL) on learning and motivation (Erhel & Jamet, 2013). For example, Huang, Huang, and Tschopp (2010) based on the data collected by ARCS-based Instructional Materials Motivational Survey (IMMS), a regression analysis revealed a significant model between motivational processing (attention, relevance, and confidence) and the outcome processing (satisfaction). Clark et al. (2011) checked for how similar or different are the learning and affective experiences of students playing the game in two different countries (i.e., Taiwan and the United States). In the other research, Huang (2011) found that learners’ cognitive capacities were in high demand in the online gamed-based learning. Since researchers on cognitive load have concluded that an overloaded cognitive capacity can de-motivate learners. This study found the all research used questionnaire survey to measure the attention score, the score was not an impersonal measure. Thus, the first goal of this study is to measure the attention score by affective computing technique and compare the attention, affective experiences and cognitive load of learners in digital-game based and traditional learning environments. However, major reviews of digital games seeking to explore the issue of academic achievement have reported contradictory or ambiguous findings (Papastergiou, 2009). A meta-analysis of students’ learning performance (Randel, Morris, Wetzel, & Whitehill, 1992). 2.

(17) reported ambiguous results, with the majority of studies (38) indicating no difference between game-based and traditional teaching methods 27 studies advocating game-based learning, and 3 studies supporting traditional methods of instruction. Unfortunately, the effectiveness of DGBL on students’ academic achievement is still unproven in a robust empirical research setting. Thus, the second goal of this study is to compare the academic achievement of learners in digital-game based and traditional learning environments. Problem-solving is a 21st century skill which is essential for learning, work, and daily life (L. A. Annetta, 2008). Problem solving can be defined as the ability to find causes, find solutions, and avoid problems (Chan & Wu, 2007). This study found the 4 factors at least can affect the problem solving strategy: (1) Learning environments Digital games provide a meaningful framework for solving problems (L. A. Annetta, 2008), since students are placed in scenarios in which they must synthesize diverse information and analyze strategies, leading to a greater understanding of the causal links between decision-making behaviors (Ebner & Holzinger, 2007). Therefore, digital games can be seen as a good tool for understanding the link between cause and effect (Kiili, 2005). Although research on problem solving in DGBL has been conducted (Dickey, 2006; Robertson & Howells, 2008), its effectiveness in fostering problem solving abilities has not received sufficient attention from empirical research. (2) Learners’ learning style Graf, Lin, and Kinshuk (2008) investigated the relationship between learning styles and working memory capacity, they demonstrated that learners with high working memory capacity tend to prefer a reflective, intuitive, and sequential learning style whereas learners with low working memory capacity tend to prefer an active, sensing, visual, and global learning style. In. 3.

(18) addition, previous research indicates that individuals with higher levels of working memory capacity perform better on learning tasks because they have more cognitive resources (Daneman & Carpenter, 1980; R.E. Mayer, 2001) It is likely that working memory capacity affects cognitive efficiency due to the processing and storage requirements necessary to solve mental problems (Hoffman & Schraw, 2009). According these research findings, this paper want to investigate the relationship between learning style and problem solving abilities. (3) Prior knowledge F. Y. Yang, Chang, Chien, Chien, and Tseng (2013) investigated university learners’ visual attention during a PowerPoint (PPT) presentation on the topic of “Dinosaurs” in a real classroom. The results showed that the earth-science majors’ students were better at information decoding and integration than non-earth-science majors’ students. The discussion points out that the interaction between types of graphics and information processing behaviors is also mediated by prior knowledge. (4) Academic achievement Tsai, Hou, Lai, Liu, and Yang (2012) employed an eye-tracking technique to examine students’ visual attention when solving a multiple-choice science problem. The results showed that successful problem solvers focused more on relevant key factors, while unsuccessful problem solvers experienced difficulties in decoding the problem, in recognizing the relevant factors, and in self-regulating of concentration. Thus, the third goal of this study is to compare the science problem solving strategy of learners in different learning environments, learning style, major and academic achievement.. 4.

(19) 1.2 Research purposes For verifying the possibility of playing digital game to learn the physics problem, this study used a quasi-experimental design to examine the effectiveness of Digital Game-Based Learning (DGBL) and traditional static e-learning on students’ learning attention, affective experiences, cognitive load, academic achievement and problem solving skills. When student learning, their eye movement data, brain wave and heart beat were measured for analyzing their learning states. After learning, learners took a posttest to find the difference in academic achievement between DGBL and static e-learning. The other hand, this study used the eye movement data to find the difference problem solving strategy between different working memory capacity learning style, learning environments, major and problem-solving performance. Thus, the following research purposes were explored: 1. Using the affective computing technique to evaluate the learning attention, affective experiences and cognitive load of Digital Game-Based Learning. 2. Taking pre-test and posttest to evaluate the effects of the game environment on students’ achievement. 3. Using eye movement variables to find the differences of problem solving strategy between different working memory capacity learning style (active vs. reflective, sensing vs. intuitive, sequential vs. global), learning environments (DGBL vs. traditional static e-learning), major (non-science vs. science) and problem-solving performance (low performance vs. high performance).. 5.

(20) 1.3 Research gaps a. Few studies used affective technique to evaluate DGBL issue Affective computing technique has become the learning research trends. But there is few studies used physiology signals recognition in DGBL issue. In this study, the affective computing technique was used to measure the students’ learning attention. In the literature, several studies for physiology signals recognition in learning have used eye tracking technology to observing the visual attention (Dimigen, Sommer, Hohlfeld, Jacobs, & Kliegl, 2011; Latanov, Konovalova, & Yermachenko, 2008; Lin, Imamiya, & Mao, 2008; Schmid, Schmid Mast, Bombari, Mast, & Lobmaier, 2011), also have used EEG and Heart rate variability in ECG to measure the learning emotion(Chen & Wang, 2011; C. Zhang, Zheng, & Yu, 2009) and mental workload (Patel, Lal, Kavanagh, & Rossiter, 2011; Zhao, Zhao, Liu, & Zheng, 2012). b. Few studies of DGBL and learning style which used eye-tracking technology were situated in the problem solving strategy. Few studies of DGBL and learning style which used eye-tracking technology were situated in the problem solving strategy. But several studies of multimedia learning were used. For example, Tsai et al. (2012) employed an eye-tracking technique to examine students’ visual attention when solving a multiple-choice science problem. F. Y. Yang et al. (2013) investigated university learners’ visual attention during a PowerPoint (PPT) presentation on the topic of “Dinosaurs” in a real classroom. The results showed that the earth-science majors’ students displayed higher visual attention than non-earth-science students to the text zones, but few differences were found for the picture zones. When the students viewed those slides containing scientific hypotheses, the difference in attention distributions between the text and pictures. 6.

(21) reduced. Further analyses of fixation densities and saccade paths showed that the earth-science majors’ students were better at information decoding and integration. Thus, eye-tracking studies have provided insights on how students pay attention to learning material. This study further explores how learning environment and style affect the science problem solving strategy of students.. 7.

(22) CHAPTER Ⅱ LITERATURE REVIEW In chapter 2, the literature of Digital Game-Based Learning, problem solving strategy and affective computing in learning were organized. In section 2.1, this paper described that advantage of Digital Game-Based Learning. And then, the learning attention, affective experiences, cognitive load, academic achievement and gender in Digital Game-Based Learning issue were discussed. In section 2.2, this paper summarized 4 factors at least can affect the development of problem solving strategy, included learning environments, learners’ learning style, prior knowledge and academic achievement. In section 2.3, these papers of affective computing in learning were organized. In addition, this paper found that eye movement, EEG and ECG have become the research trends, so we introduced that researches and variables of eye movement, EEG and ECG.. 2.1 Digital Game-Based Learning Digital entertainment games have become one of the most popular leisure activities globally. The effect of Digital Game-Based Learning in promoting meaningful learning might be due to opportunities for “learning by doing” (Pannese & Carlesi, 2007). According to T. M. Connolly et al. (2007), GBL can be defined as “the use of a computer game-based approach to deliver, support, and enhance teaching, learning, assessment, and evaluation”. Students use games to explore, discover, and question, ultimately constructing concepts and relationships in authentic contexts. These “learning by doing” and “active learning” concepts are important constructivist principles which underlie game-based learning (Y. T. C. Yang, 2012). Richard E. Mayer and Johnson (2010) considered that a DGBL environment should feature 1) a set of rules and constraints, (2) a set of dynamic responses to the learners’ actions, (3). 8.

(23) appropriate challenges enabling learners to experience a feeling of self-efficacy, and (4) gradual, learning outcome oriented increases in difficulty.. 2.1.1 The advantages of Digital Game-Based Learning Games that encompass educational objectives and subject matter are believed to hold the potential to render learning of academic subjects more learner-centered, easier, more enjoyable, more interesting, and, thus, more effective (Prensky, 2001). There is also widespread acknowledgment of the advantages that the use of games has in elementary and secondary education (Ebner & Holzinger, 2007). Specifically, games constitute potentially powerful learning environments for a number of reasons. Kebritchi and Hirumi (2008) identified the following five reasons for defining GBL as an effective tool for learning: (a) GBL uses action instead of explanation; (b) GBL creates personal motivation and satisfaction; (c) GBL accommodates multiple learning styles and skills; (d) GBL reinforces mastery of skills; and (e) GBL provides an interactive and decision-making context. According to O’Neil, Wainess, and Baker (2005), computer games are useful for instructional purposes and they also provide multiple benefits: (a) complex and diverse approaches to learning processes and outcomes; (b) interactivity; (c) ability to address cognitive as well as affective learning issues; and (d) motivation for learning. Oblinger (2004) identified the following five reasons: (a) they can support multi-sensory, active, experiential, problem-based learning, (b) they favored activation of prior knowledge given that players must use previously learned information in order to advance, (c) they provide immediate feedback enabling players to test hypotheses and learn from their actions, (d) they. 9.

(24) encompass opportunities for self-assessment through the mechanisms of scoring and reaching different levels, and (e) they increasingly become social environments involving communities of players.. 2.1.2 Learning motivation, affective experiences, cognitive load and Digital Game-Based Learning Many studies have investigated the effects of Digital Game-Based Learning (DGBL) on learning and motivation (Erhel & Jamet, 2013). Thomas M. Connolly, Boyle, MacArthur, Hainey, and Boyle (2012) found the use of games to teach educational content inevitably raises the question of their compatibility with deep learning. This has prompted many researchers to investigate the actual benefits of digital games, in terms of learning and motivation. For instance, Virvou, Katsionis, and Manos (2005) designed VR-ENGAGE computer game for teaching geography to fourth grade students. It was concluded that computer games could promote motivation, especially for at-risk students or for students with motivational problems. Digital Game-Based Learning provides that learners have sufficient level of curiosity to explore the learning task, sustain the learners’ motivation. A study used a attention subscale to measure the attention score when learners study by digital game (Huang, 2011). Learning motivation is dependent of four perceptual components: attention, relevance, confidence and satisfaction (Keller, 2008). The study surveyed 264 undergraduate students after playing the Trade Ruler online game. Based on the data collected by ARCS-based Instructional Materials Motivational Survey (IMMS), a regression analysis revealed a significant model between motivational processing (attention, relevance, and confidence) and the outcome processing (satisfaction). Based on the observed ARCS scores in this study, learners started out with a successful motivational processing that consisted of a high attention level, a low relevance. 10.

(25) level, and a high confidence level. At the end of the learning process, however, they reported a relatively low level of satisfaction (Huang et al., 2010). Derbali and Frasson (2010) investigated players’ motivation during serious game play. It is based on a theoretical model of motivation (John Keller’s ARCS model of motivation) and EEG measures, and the results showed that power spectral analysis showed EEG waves patterns correlated with increase of motivation during different parts of serious game play. Thus, this study tried to measure the attention score by EEG. Clark et al. (2011) checked for how similar or different are the learning and affective experiences of students playing the game in two different countries (i.e., Taiwan and the United States). In the other research, Huang (2011) found that learners’ cognitive capacities were in high demand in the online gamed-based learning. Since researchers on cognitive load have concluded that an overloaded cognitive capacity can de-motivate learners. This study found the all research used questionnaire survey to measure the motivation score, the score was not an impersonal measure. Thus, the first goal of this study is to measure the attention score by affective computing technique and compare the attention, affective experiences and cognitive load of learners in digital-game based and traditional learning environments.. 2.1.3 Academic achievement and Digital Game-Based Learning The major reviews of digital games seeking to explore the issue of academic achievement have reported contradictory or ambiguous findings (Papastergiou, 2009). A meta-analysis of students’ learning performance Randel et al. (1992) reported ambiguous results, with the majority of studies (38) indicating no difference between game-based and traditional teaching. 11.

(26) methods 27 studies advocating game-based learning, and 3 studies supporting traditional methods of instruction. Unfortunately, the effectiveness of DGBL on students’ academic achievement is still unproven in a robust empirical research setting. Thus, the third goal of this study is to compare the academic achievement of learners in digital-game based and traditional learning environments. Vogel, Greenwood-Ericksen, Cannon-Bowers, and Bowers (2006) in a literature review based on 32 empirical studies, reported that interactive games were more effective than traditional classroom instruction on learners’ academic learning gains and cognitive skill development. Otherwise, many studies found that no differences in student learning can be found between learning environments that involve games and those without game elements (Leonard A. Annetta, Minogue, Holmes, & Cheng, 2009; Papastergiou, 2009) Wrzesien and Alcañiz Raya (2010) found sixth graders reported higher motivation and engagement levels as a result of playing a science-based game; however, there was no evidence to show that the game led to significant learning advancements over the traditional class. Base on above literature, this study wanted to compare the academic achievement of learners in digital-game based and traditional learning environments.. 2.1.4 Gender and Digital Game-Based Learning Regarding gender issues, as shown in the study, despite the fact that the boys of the sample exhibited significantly greater involvement with, liking of and experience in computer gaming outside school as well as significantly greater initial knowledge of the embedded subject matter, and greater interaction among them during the intervention, the learning gains that boys and girls achieved through the use of the game did not differ significantly. Furthermore, no significant gender differences were found in students’ views on the overall appeal, quality of user interface, and educational value of the game used. Papastergiou (2009)Data analyses showed that the. 12.

(27) gaming approach was both more effective in promoting students’ knowledge of computer memory concepts and more motivational than the non-gaming approach. Despite boys’ greater involvement with, liking of and experience in computer gaming, and their greater initial computer memory knowledge, the learning gains that boys and girls achieved through the use of the game did not differ significantly , and the game was found to be equally motivational for boys and girls. Ke and Grabowski (2007) also found that gender did not influence the learning effectiveness and motivational appeal of games for school children. Thus, this study didn’t check the relationship between gender and Digital Game-Based Learning.. 2.2. Problem solving strategy Problem-solving is a 21st century skill which is essential for learning, work, and daily life (L. A. Annetta, 2008). Problem solving can be defined as the ability to find causes, find solutions, and avoid problems (Chan & Wu, 2007). It has been argued that digital games provide a meaningful framework for solving problems (L. A. Annetta, 2008), since students are placed in scenarios in which they must synthesize diverse information and analyze strategies, leading to a greater understanding of the causal links between decision-making behaviors (Ebner & Holzinger, 2007). Therefore, digital games can be seen as a good tool for understanding the link between cause and effect (Kiili, 2005). Thus, apart from knowledge acquisition, game playing can also favored the development of various skills, such as critical thinking and problem-solving skills. Problem solving is one of the integral approaches to achieving effective and meaningful learning (D. H. Jonassen, 2004). Learners, while solving a problem, have to understand the problem, devise a plan and test the plan to solve it. In other words, they have to analyze the strategies which can possibly solve the problems by themselves, and thus, are more likely to generate creative solutions and achieve effective learning. As a result, problem solving has been. 13.

(28) extensively applied to many subject domains such as science (Linn, Clark, & Slotta, 2003), mathematics (D. Jonassen, 2003) and design (Jermann & Dillenbourg, 2008) as a means of promoting learning in these domains. This study found the 4 factors at least can affect the problem solving strategy: (1) Learning environments Digital games provide a meaningful framework for solving problems (L. A. Annetta, 2008), since students are placed in scenarios in which they must synthesize diverse information and analyze strategies, leading to a greater understanding of the causal links between decision-making behaviors (Ebner & Holzinger, 2007). Therefore, digital games can be seen as a good tool for understanding the link between cause and effect (Kiili, 2005). Although research on problem solving in DGBL has been conducted (Dickey, 2006; Robertson & Howells, 2008), its effectiveness in fostering problem solving abilities has not received sufficient attention from empirical research. (2) Learners’ learning style Graf et al. (2008) investigated the relationship between learning styles and working memory capacity, they demonstrated that learners with high working memory capacity tend to prefer a reflective, intuitive, and sequential learning style whereas learners with low working memory capacity tend to prefer an active, sensing, visual, and global learning style. In addition, previous research indicates that individuals with higher levels of working memory capacity perform better on learning tasks because they have more cognitive resources (Daneman & Carpenter, 1980; R.E. Mayer, 2001) It is likely that working memory capacity affects cognitive efficiency due to the processing and storage requirements necessary to solve mental problems (Hoffman & Schraw, 2009). According these research findings, this paper want to investigate the relationship between learning style and problem solving abilities.. 14.

(29) (3) Prior knowledge F. Y. Yang et al. (2013) investigated university learners’ visual attention during a PowerPoint (PPT) presentation on the topic of “Dinosaurs” in a real classroom. The results showed that the earth-science majors’ students were better at information decoding and integration than non-earth-science majors’ students. The discussion points out that the interaction between types of graphics and information processing behaviors is also mediated by prior knowledge. (4) Academic achievement Tsai et al. (2012) employed an eye-tracking technique to examine students’ visual attention when solving a multiple-choice science problem. The results showed that successful problem solvers focused more on relevant key factors, while unsuccessful problem solvers experienced difficulties in decoding the problem, in recognizing the relevant factors, and in self-regulating of concentration. Thus, the third goal of this study is to compare the science problem solving strategy of learners in different learning environments, learning style, major and academic achievement.. 2.3. Affective Computing in Learning Since Affective Computing was proposed, there has been a burst of research that focuses on creating technologies that can monitor and appropriately respond to the affective states of the user (Picard, 1997). Because this new Artificial Intelligence area, computers able to recognize human emotions in different ways. Why human emotion is an important research area? The latest scientific findings indicate that emotions play an essential role in decision-making, perception, learning and more (Ben Ammar, Neji, Alimi, & Gouardères, 2010).. 15.

(30) Eye movement, EEG and ECG have become the research trends Some of the main physiological signals highly adopted for human emotion assessment are: Electrocardiogram (ECG), Electromyogram (EMG), Skin Conductive Resistance (SCR), and Blood Volume Pressure (BVP). Several approaches have been found the correlation between the emotional changes and EEG signals (Chanel, Ansari-Asl, & Pun, 2007). In this study, we summarize the multi physiological feature system researches as shown in Table 1. According the table, we found the physiological signals of eye movement, EEG and ECG have become the research trends. But it not exist a system combined these signals to recognize the affective of human.. Attention and emotion effect student learning and performance Our aim is develop a learning affective recognition system, which can get, recognize and analyze attention and emotion state when students learning. Attention can effects student learning and performance is as everyone knows. And distractor processing with emotional information also has implications for theories of attention (Srinivasan & Gupta, 2010). Thus, if human emotions are essential for human thinking and learning processes, a successful learning environment have to intensively consider this fact. In recent years, researchers have increasingly called into question the relevance of these basic emotions to the learning process (Baker, D'Mello, Rodrigo, & Graesser, 2010). Much of what is known about achievement emotions stems from research with students participating in traditional educational settings. Positive emotions, such as enjoyment, hope, and pride, have been positively associated with intrinsic motivation, effort, self-regulation, and more sophisticated learning strategies (Pekrun, Goetz, Frenzel, Barchfeld, & Perry, 2011), whereas. 16.

(31) negative emotions such as anger/frustration, shame, anxiety, and boredom have been associated with reduced effort, lower performance, increased external regulation, and decreased self-regulated learning strategies(Daniels et al., 2009).. The Physiological input signals this study selected The physiological input signals of eye movement, EEG and ECG were selected to input our learning affective recognition system. According the past studies, several techniques need to be combined to estimate the state of attention and emotion. Eye movements provides information about location of attention and the nature, sequence and timing of cognitive operations (Lin et al., 2008). With the emergence of Electroencephalography (EEG) technology, learner's brain characteristics could be accessed directly and the outcome may well hand-in-hand supported the conventional test. recognize a learner's Learning Style (Rashid et al., 2011). And the arousal state of the brain (Q. Zhang & Lee, 2012), alertness, cognition, and memory (Chris Berka et al., 2004; C. Berka et al., 2007) also can be measure. Heart rate variability from ECG, has gained widespread acceptance as a sensitive indicator of mental workload(Lin et al., 2008). And positive emotions may change the HF components of HRV (von Borell et al., 2007).. 17.

(32) Table 1.Multi physiological feature system review According the table, we found the physiological signals of eye movement, EEG and ECG have become the research trends. But it not exist a system combined these signals to recognize the affective of human. Emotion recognition features Research object. Reference. Eye. EEG. ECG. emotion recognition. (Kim, Ban, & Kim, 2004). neonatal seizures. (Greene, Boylan, Reilly, de Chazal, & Connolly, 2007). emotion recognition. (Ruffman, Henry, Livingstone, & Phillips, 2008). emotion recognition. (Lin et al., 2008). ✕. visual search task. (Latanov et al., 2008). ✕. emotion recognition. (Q. Zhang & Lee, 2010). emotional distractors. (Srinivasan & Gupta, 2010). emotion recognition. (B. Yang & Lugger, 2010). emotion recognition. (Murugappan, Ramachandran, & Sazali, 2010). brain computer interface. (Lee, Woo, Kim, Whang, & Park, 2010). reading process. (Dimigen et al., 2011). ✕. emotion recognition. (Schmid et al., 2011). ✕. learning state. (Chen & Wang, 2011). ✕. driver fatigue. (Patel et al., 2011). ✕. driver fatigue. (Zhao et al., 2012). ✕. emotion recognition. (Q. Zhang & Lee, 2012). ✕. epilepsy state. (Valderrama et al., 2012). ✕. ✕. Learning state. My research. ✕. ✕. Facial. speech. ✕ ✕. ✕. ✕ ✕ ✕. ✕. ✕. ✕ ✕. ✕ ✕ ✕ ✕. ✕ 18. ✕ ✕. ✕. SCR.

(33) 19.

(34) 2.3.1 Eye movement variables The eye tracking technique has been typically adopted to examine human visual attention based on the eye-mind assumption (Just & Carpenter, 1980). In general, eye fixation location reflects attention and eye fixation duration reflects processing difficulty and amount of attention (the longer the information is fixated, the more complex it is or the deeper it is processed). The fixation locations and duration reflect the individuals’ reading strategies and prior knowledge or experience (Hyönä, Lorch, & Kaakinen, 2002). Besides, scan path patterns exhibit individuals’ cognitive strategies utilized in goal-oriented tasks (Gandini, Lemaire, & Dufau, 2008). Thus, a record of eye movements provides information about location of attention and the nature, sequence and timing of cognitive operations (Lin et al., 2008). Three eye movement parameters—fixation, saccade and scan path—are often investigated to indicate users’ cognitive behaviors (Lin et al., 2008). Fixation is defined as a relatively motionless gaze that usually lasts for 200–500 ms, during which information about a visual stimulus is extracted (Jukka, 2010). Saccades are rapid and ballistic movements of gaze between fixations with a velocity of about 500。, directing the viewer’s eye to a visual target. Information processing is suppressed during a saccade, though some peripheral information may be available. Scan paths are defined as a sequence of fixations and saccades, indicating a movement of attention. About an eye tracking study for solving multiple-choice science problem, the researcher tested their hypotheses and got some discussion as following (Tsai et al., 2012). (1) There is a significant difference in students’ scan sequences among factors for solving the problems between successful and unsuccessful problem solvers. Successful problem solvers tend to shift their visual attention from irrelevant factors to relevant factors, while unsuccessful problem. 20.

(35) solvers tend to shift their visual attention from relevant to irrelevant factors and to the problem statement. It is obvious that the two groups shift attentions in opposite directions. (2) Students pay more attention to chosen options than to rejected alternatives, and tend to spend more time inspecting relevant factors than irrelevant one. Based on these studies, the strategy of solving problem will be analysis by eye movement data.. Parameters Fixation (visual attention). Table 2.Eye movement parameters Research variables fixation durations (Tsai et al., 2012) hot zone image (Tsai et al., 2012) number of fixation (Schmutz, Roth, Seckler, & Opwis, 2010). Scan path (strategy). sequential analysis (Tsai et al., 2012) qualitative analysis (Schmutz et al., 2010). 21.

(36) 2.3.2 Brain wave variables The EEG signal is a voltage signal that can be measured on the surface of the scalp, arising from large areas of coordinated neural activity. This neural activity varies as a function of development, mental state, and cognitive activity, and the EEG signal can measurably detect such variation. EEG is generally described in terms of its frequency band. The amplitude of the EEG shows a great deal of variability depending on external stimulation as well as internal mental states.. The relationship between EEG and emotion The best known correlates of emotionality found with EEG involve prefrontal asymmetry. That is, more active left frontal region indicates a positive reaction, and more active right anterior lobe indicates negative affection (Q. Zhang & Lee, 2012). The two that are most important for arousal state of brain are the alpha (8–12 Hz) and beta (12–30 Hz) frequencies. Alpha waves are typically for an alert/relaxed mental state, while beta activity is most prominent in the frontal cortex over other areas during intense focused mental activity (Kandel, Schwartz, & Jessell, 2000). Therefore, the beta/alpha ratio could be an indication of the arousal state of the brain. By taking arousal axis into consideration, we can sub-categorize positive and negative emotions into pleasure/joy and sadness/anger, respectively.. The relationship between EEG and attention Gamma synchrony between frontal and posterior regions enhance in many cognitive and neural processes including attention, perception, memory, and information processing (Fries,. 22.

(37) Nikolić, & Singer, 2007). The electroencephalogram (EEG) can monitor indexes of alertness, cognition, and memory (Chris Berka et al., 2004; C. Berka et al., 2007). Table 3.Emotion and attention state in EEG State. Variables. Category. Emotion. Beta / Alpha ratio. Positive or Negative emotions. Attention. Gamma wave. High or low mental activity. 2.3.3 Heart rate variables Heart rate variability (HRV) is regarded as an indicator of the autonomic regulation activity of heart rate, specifically sympathetic and parasympathetic activity (Tran, Wijesuriya, Tarvainen, Karjalainen, & Craig, 2009). Analysis of the HR and the HRV is increasingly used to investigate cardiovascular regulation both under physiological and pathological conditions (Malliani, Pagani, Lombardi, & Cerutti, 1991). The advantage of the frequency domain analysis is its ability to break HRV time series data into different spectrum viz. ULF, VLF, LF and HF and thereby providing information regarding each spectrum individually. Information related to attention can be obtained by analyzing LF and HF bands and deriving LF/HF ratio. The LF/HF ratio is considered to be a measure of sympathovagal balance (American College of Cardiology/American Heart Association, 1999). Spectral changes in low-frequency (LF; 0.04–0.15 Hz) and high-frequency (HF; 0.15–0.4 Hz) components of HRV have been reported to be associated with distressing conditions such as hemorrhagic shock, acute myocardial infarction, elevated anxiety, and depressed mood.. 23.

(38) Table 4.Power spectrum components Frequency (Range). Influence. Reference. Ultra-low (0.0–0.0033 Hz). unknown. (Batchinsky et al., 2007). Very low (0.0033–0.04 Hz). unknown. (Batchinsky et al., 2007). Low (0.04–0.15 Hz). Vagal and the sympathetic nervous (Stein & Kleiger, 1999) system. High (0.15–0.40 Hz). Vagal activity. (Stein & Kleiger, 1999). The relationship between HRV and emotion Positive emotions may increase the HF components of HRV, indicating that parasympathetic nerve activity and negative emotion may increase LF components of the sympathetic nervous system (von Borell et al., 2007). Heartbeat as a random point process, its rate is dependent on the activity level of the autonomic nervous system, which in turn is dependent on emotional stimuli (Kim et al., 2004). In previous research, heart rate variability (HRV) was found a relationship with subjective well-being – as indicated by positive habitual mood (Geisler, Vennewald, Kubiak, & Weber, 2010).. The relationship between HRV and attention Spectral analysis of HRV has gained widespread acceptance as a sensitive indicator of mental workload(Lin et al., 2008). For example, the LF component of the HRV power spectrum systematically decreases as mental demands increase (Lin et al., 2008). And Low LF/HF ratio could serve as an indicator of alert (Patel et al., 2011).. 24.

(39) Table 5.Emotion and attention state in HRV State. Variables. Category. Emotion. HF and LF. Positive emotions / Negative emotions. Attention. LF/HF ratio. Alert or Relax. The analysis method of HRV The analysis method of HRV can be evaluated using time-based measures or frequency domain measures (Patel, et al., 2011). Time domain measures are common and the simplest to perform and can be assessed with calculation of the standard deviation of R–R (inter-beat) intervals. Frequency domain analysis is based on mathematical transformations (i.e., Fast Fourier Transforms) of the signals from time domain to frequency domain (expressed in cycles per beat with varying amplitudes and frequencies). Power spectral density (PSD) analysis provides the basic information of how power (i.e., variance) distributes as a function of frequency (American Heart Association, 1996). Methods for calculating PSD can be classified as nonparametric and parametric. Nonparametric (which is normally conducted using FFT) has the advantage of high processing speed and simplicity of algorithm whereas parametric (which is normally conducted using autoregressive model) has the advantage of smoother spectral components that can be distinguished independent of preselected frequency bands.. 25.

(40) Chapter Ⅲ METHOD For verifying the possibility of playing digital game to learn the physics problem, this study explored 3 research questions: (1) DGBL group have better learning motivation, affective experiences and more cognitive load (the proof of physical signals) than traditional static e-learning group? (2) DGBL group have better academic achievement than traditional static e-learning group? (3) There have significant difference in science problem solving strategy of learners in different learning environments, learning style, major and academic achievement.. 3.1. Research hypotheses PHASE 1 For using affective computing technique to evaluate and compare the attention score, affective experiences and cognitive load of learners in digital-game based learning and traditional static e-learning environments, the attention in ARCS model was measured by NeuroSky, the affective experiences was measured by emWave and the cognitive load was measured by eye-tracker. Academic achievement was measured by pre-tests (Force Concept Inventory, FCI) and post-test (Mechanics Baseline Test, MBT). In addition, the following two different learning environments were designed: (1) Digital Game-Based Learning, learners studied the physics problem by SURGE physics game; (2) traditional static e-learning, the learners studied by text description, coordinates and formula. To fairly compare how learning environments affect learning attention, emotions, strategy and, the two learning environments had the same learning content and learning objectives, that is, the same learning materials are presented in different methods. Table 6 shows the hypotheses and reference. Figure 1 shows the relationship framework in PHASE 1.. 26.

(41) Table 6.Hypotheses and reference in PHASE 1 Hypotheses. Reference. H1. DGBL group have better attention score than traditional static (Derbali & Frasson, 2010) e-learning group and have significant difference. H2. DGBL group have better affective experiences than traditional (Baldaro et al., 2004; static e-learning group and have significant difference. Ravaja, Turpeinen, Saari, Puttonen, & Keltikangas-Jarvinen, 2008) H3. DGBL group have more cognitive load than traditional static (Huang, 2011) e-learning group and have significant difference. H4. DGBL group have better academic achievement than traditional (Vogel et al., 2006) static e-learning group.. Attention (brain wave) Attention score. H1. Affective experiences (heart beat) Occupied percentage of positive emotion. Learning method. H2 Cognitive load (eye movement). 1. Digital game-based 2. Traditional static e-Learning. H3 H4. Total fixation duration (TFD) Number of fixations (NF) Average fixation duration(AFD) Percentage of viewing time (PVT) Frequency of saccade path (FSP) Sum of saccade paths (SSP). Academic achievement Post-test score. Figure 1.Research framework in PHASE 1. 27.

(42) H1. DGBL group have better attention score than traditional static e-learning group and have significant difference. Brain wave measurement tool manufactured by Neurosky was adopted, which was one type of non-invasive brain wave measurement instrument. Derbali and Frasson (2010) investigated players’ motivation during serious game play. It is based on a theoretical model of motivation (John Keller’s ARCS model of motivation) and EEG measures, and the results showed that power spectral analysis showed EEG waves patterns correlated with increase of motivation during different parts of serious game play. H2. DGBL group have better affective experiences than traditional static e-learning group and have significant difference. Ravaja, Turpeinen, Saari, Puttonen, and Keltikangas-Jarvinen (2008) found the game events did in fact lead to emotion state. This indicated that physiological and emotional changes take place while playing games (Baldaro et al., 2004). Based on these studies, we can assume Digital Game-Based Learning can increase learners’ positive emotion. H3. DGBL group have more cognitive load than traditional static e-learning group and have significant difference. Learners’ cognitive capacities were in high demand in the online GBLE. Since researchers on cognitive load have concluded that an overloaded cognitive capacity can de-motivate learners, Huang (2011) argues that the target online GBLE might overload learners’ cognitive capacity thus lead to a fairly unsatisfactory learning experience. H4. DGBL group have better academic achievement than traditional static e-learning group.. 28.

(43) Based on 32 empirical studies, Vogel et al. (2006) reported that interactive games were more effective than traditional classroom instruction on learners’ academic learning gains and cognitive skill development.. PHASE 2 The goal of PHASE 2 is to compare the science problem solving strategy of learners in different working memory capacity learning style (active vs. reflective, sensing vs. intuitive, sequential vs. global), learning environments (DGBL vs. traditional static e-learning), major (non-science vs. science) and performance (low performance vs. high performance). In this study, 1 easy problem and 1 difficult problem in post-test were selected from Mechanics Baseline Test (MBT). The test covers concepts in basic principles (Newtons' First, Second, and Third Laws, superposition principle) and special forces (gravity and friction).. Table 7.Hypotheses and reference in PHASE 2 PHASE 2 H1. High working memory capacity learning style group knew better (Graf et al., 2008; where to look the key factors than low working memory capacity Hoffman & Schraw, learning style group. 2009) H2. DGBL group knew better where to look the key factors than (Liu, Cheng, traditional static e-learning group Huang, 2011). &. H3. Science major group knew better where to look the key factors (F. Y. Yang et al., than non-science major group. 2013) H4. Successful problem solvers are able to recognize and concentrate (Tsai et al., 2012) on relevant cues more than unsuccessful problem solvers H5. Successful problem solvers inspected the factors in a different (Tsai et al., 2012) pattern from unsuccessful problem solvers.. 29.

參考文獻

相關文件

The purposes of this research are to find the factors of raising pets and to study whether the gender, age, identity, marital status, children status, educational level and

This research project tries to examine the issues of domestic green hotel service design, integrating QFD and VAHP, as well as to construct green criteria and their

Therefore, the purpose of this study is to propose a model, named as the Interhub Heterogeneous Fleet Routing Problem (IHFRP), to deal with the route design

This study focuses on the consumer’s impression in exploring the quality attributes of digital photo frame (DPF), in which the Kano’s model is used to classify the quality

This study based on the computer attitudes, the digital learning attitude and the digital game attitude and tried to find out the factors affecting digital game-based

The effect of gender on motivation and student achievement in digital game-based learning: A case study of a contented-based classroom. Using game-based learning to

Currently, much of the literature about Flipped Classroom used a quasi-experimental method and has demonstrate its merits: the improvement of the relationship

This study combined the concepts of service science, orienteering problem, geographic information systems and genetic algorithms to design the feeder bus route.. In