人們在何時何地、做什麼活動時能有最佳經驗(optimal experience)呢? Csikszentmihalyi (1975, 1990)以心流 (flow)指出一個人專心致志於某一活動，即使付出頗多認知資源，但個人的注意力、知覺、記憶、目標與自我依然呈現「極其順暢」、「毫不費力」的狀態，這就是個人認知與情感的最佳經驗。心流是在活動之中感受到的當下經驗，容易受到外在環境影響產生變化，研究者視為瞬時變化的經驗 (momentary experiences)，而非穩定的人格特質 (Seligman & Csikszentmihalyi, 2000)。但是有些人確實較常體驗到心流，因此研究者也同意心流具有個別差異，其特質包含：傾向發自內心想要探索生活周遭，積極地面對生活中的各種挑戰，主動學習更多技能，換句話說這些人對自己所作的事具有強烈的內在動機，相對地他們較不太注重外在環境與條件；研究者稱他們為具有自成目標的人格特質或傾向 (autotelic personality; Nakamura & Csikszentmihalyi, 2002)。過去大多數心流研究以西方成年人、尤其專業人士為研究對象，對青少年一般生活情境也略有著墨，但較少研究探索學生在教室內的最佳經驗，更少有以台灣學生為樣本的實證研究。以「國小」與「心流」為關鍵字，在「國家圖書館期刊文獻資訊網」與「全國碩博士論文網」搜尋，僅得到九篇國內相關論文（陳彥瑋，2006；鄭婷文，2009；郭士豪，2011；鄭富美，2011；蕭伊茹，2013；劉向軒，2013；劉家均，2014；李杰駿，2014；陳和昌，2014），篇數極少，不足以呈現臺灣國小學生學習活動經驗的全貌，也無法比較學生在不同科目和時段的心流經驗。這些論文多採用橫斷式問卷調查或實驗法，探討單一學科（數學、英語、音樂、美術、生活科技、閱讀）的心流為目標。此外，本研究僅搜尋到兩篇期刊論文分析台灣「大學生」的最佳經驗(Tseng, 2010a, 2010b)。
本論文探索學習情境中的最佳經驗，針對學習活動及學校情境拓展心流理論與修正量測方法。關於理論部分，文獻顯示過去研究對於心流定義是分歧的，而本研究從中選擇的心流（最佳經驗）操作型定義是適用在學習情境、特別是適合年紀較小的國小學生群體。過去心流相關研究大多採取渠道論(experience fluctuation model; EFM) 將心流定義為多種經驗中(如：無聊、焦慮…)的一類，這樣的心流是「類別變項」，研究者觀察受試進入心流渠道或不在心流渠道的差異。同時大多數研究者都依據理論去設計特定的任務，努力地引導出心流，加以觀察。本研究不同於以往之處，在於 (1) 純粹記錄學生參與三種教學策略的課堂、課後參與四種活動所體驗到的最佳經驗，並不設計任務去誘發心流；本研究目的在於觀察真實學校課堂、課後活動當下，小學生是否有相對優質的學習經驗。(2) 本研究定義的最佳經驗是在事件當下的連續變項，因此可以比較在那些環境脈絡因素（教學策略）與個人知覺的環境特性 （心流條件）時，能得到較強或較弱的最佳經驗。心流是一種複雜而瞬時變化的內在經驗歷程，需要同時考慮個體和環境的內外在條件，才能嚴謹的檢驗此一歷程。在研究工具與方法部分，過去專門為取得心流資料而設計的資料蒐集方法稱為「經驗取樣法」 (experience sampling method; ESM)，但以學生為樣本會因為學校的制度與規範使 ESM 變成教學中極大的干擾。因此本研究決定採用不會侵犯教學活動的「一日經驗重建法」(Day Reconstruction Method; DRM; Kahneman et al., 2004)。DRM是一種改良的調查法，能蒐集個人如何使用時間 (在某些情境，進行某些活動)，以及在各活動歷程裡個人所感受到的主觀經驗。受試者在早晨還沒進行活動前，以日記回想前一天的日常活動，將時間區分為幾個完整的事件時段，接著以問卷蒐集各事件時段中個人感受到的主觀經驗。DRM蒐集到的資料性質是一種短期密集追蹤資料 (intensive longitudinal data) ，在心理計量上具有重複量測和持續追蹤（panel data，亦即「同一受試者」反覆被量測多次） 的特性，有別於一般蒐集的長期資料經常是跨年/學期（間距較大）的問卷調查。
本論文是三個研究的組合，並以「一日經驗重建法」蒐集學生一天從起床到就寢間連續不間斷的密集追蹤資料。前二個研究以同一組小學生為樣本 (N = 191)，將國小學生一天的資料區分為「在學校」與「放學後」兩份，並採變項為核心的統計取向 (variable-centered approach)，探討在不同的課程與課後活動情境中，能影響最佳經驗的因素為何。第三個研究則以大學生為樣本(N = 271)，採人群為核心的研究取向 (person-centered approach)，根據大學生所參與的日常生活活動類型（課業活動、生產性活動、社交活動、主動休閒、被動休閒、維持性活動）與活動中的主觀經驗品質（最佳經驗、心流條件、負向情緒）深入分析是否找得到自成目標傾向的群組。
第一個研究分析國小上課時間中，各課堂的環境脈絡 (contextual factor) 與學生知覺環境特性 (perceptual factors) 是否能預測學生在課堂活動當下的最佳經驗。此研究的環境脈絡為「課堂老師採用的教學策略」，分別為講授(lecture)~以教師為中心、座位工作(seat work)~以任務為中心、及活動(activity)~以學生為中心，而學生知覺的環境特性則包含「挑戰與技巧的平衡程度」（亦即引發最佳經驗的前提條件）及「延伸性心流條件」（目標清楚、自主性、課堂重要性、專心），二者共同預測最佳經驗（依變項），同時控制了個人階層中的個人背景及心理適應程度。研究者把個人層級的訊息控制住，乃是為求事件情節(episodic)訊息的清楚性，並預期在控制個人背景與心理適應程度後，小學各課堂情節所呈現的高低心流條件（「挑戰與技巧的平衡程度」及「延伸性心流條件」）應該能預測學生體驗到的最佳經驗之高低。結果顯示國小學生在下課比起在所有課堂都更能知覺到最佳經驗，課堂經驗品質依序為: 活動式教學、座位工作與直接講授教學。此外，國小學生最佳經驗在事件階層的變異大於個體之內的變異，符合心流理論之預測：個人特性引發的經驗品質變異不大，但活動與環境營造的經驗品質變異較大。。值得注意的是，當國小學生知覺到「高技能」與「低挑戰」時比較有機會體驗到良好的經驗品質，與美國高中生心流研究(Clarke & Haworth, 1994; Csikszentmihalyi & Rathunde, 1993; Engeser & Rheinberg, 2008; Schweinle, Turner, & Meyer, 2008) 結果相似，但卻都與專業人士的心流研究結果(如：Csikszentmihalyi, 1990)並不相符。此外，事件階層中四個延伸性的心流條件對最佳經驗具有最高的預測力，優於過往相關研究重視的主要心流前提條件（技能與挑戰的平衡狀態)與三種教學策略(環境脈絡)。
研究二的對象和研究一相同，是同一組國小學高年級生。研究二分析課後活動中個人階層「自我決定動機」(self-determination motivation; Deci & Ryan, 1985) 對事件階層最佳經驗的預測力，並特別關注學生在四種課後活動（課後學習、主動休閒、被動休閒、維持性活動）中的心流條件(事件階層)是否為二者間的有效中介變項，執行跨個人與事件階層的中介模式統計程序(multi-level mediation model)。由於研究一發現個人特性能正向預測經驗品質，過去研究(Vallerand, 1997)指出個人特性中的內在動機對最佳經驗有顯著的預測力，因此研究二對比四種課後活動的中介模式，分析是否有某些活動的心流條件(環境脈絡)能完全或部分中介內在動機對最佳經驗的預測力。採取水流的比喻，內在動機對最佳經驗的預測力為水流，哪些活動的心流條件會使這一條水流完全阻斷/改道(完全中介)或能建立支流/削弱主流(部分中介)呢？結果顯示，當學生參與「積極休閒」(如打球、學樂器)時，心流條件完全中介了自我決定動機對最佳經驗的預測效果；「課後學習」(如補習、寫功課)和「消極休閒」(如看電視、打電動)兩個課後活動的心流條件只能部分中介自我決定動機與最佳經驗；而維持性活動裡的心流條件並非有效的中介變項。各活動心流條件的間接效果之強度依序為: 積極休閒、課後學習與消極休閒；維持活動的間接效果未達顯著，自我決定動機只能直接預測最佳經驗。這個結果說明了課後積極休閒之心流條件能完全阻斷原本內在動機對最佳經驗的直接預測力。自我決定動機越高的小學生，在課後如果投入積極性休閒，越能強烈感受到此活動具備優質的心流條件（目標清楚、自主性、活動重要性、專心），以至於自我決定動機完全要通過心流條件去預測最佳經驗，自我決定動機再也無法直接預測最佳經驗。相反的，自我決定動機高的小學生，在課後如果投入到維持性活動(例如吃晚飯)，感受的心流條件非常微弱，導致自我決定動機無法透過心流條件間接預測最佳經驗，此時自我決定動機只能直接預測最佳經驗。最後，自我決定動機高的小學生參與課後的消極休閒或學習活動時，感受到這兩項活動具有些微的心流條件，此動機不但局部通過心流條件去預測最佳經驗，同時也可以直接預測最佳經驗。
研究三的研究對象轉換為大學生，因為小學生的生活自理與學習型態較為侷限。採用以人為核心的統計程序—潛在群集分析(latent profile analysis; LPA) 企圖在271位來自臺灣北中南區五校、不同性質學院（人文社會與自然理工）的大學生中找出幾個次群體。LPA是統計混和模式(mixture model)中的一種，承認大群體中存在多個異質子群體的概率模式，以潛在階層的變項尋找分群，能有效避免測量誤差對分群演算的干擾，並提供BIC、AIC等指標以選取群體的合理個數。以事件階層的個人知覺經驗因素（最佳經驗、心流條件、負向情緒）與環境脈絡因素（知覺到活動類型的次數）為分組投入變項，並檢驗所形成的分組中是否有一群體比其他群體的大學生更具有正向心理特性與良好的時間使用（生活滿意、勇於接受挑戰，各類活動參與時間），目的是探索性地尋找具備自成目標傾向 (autotelic tendency) 的群體。LPA的結果區分出三群大學生，依其輪廓命名為：一般群集 (median group)、停留舒適圈群集 (staying in comfort-zone group) 與尋求有意義人生群集 (seeking for meaningful life group)。一般群集（人數為148人，54.6%）的大學生在一天事件之最佳經驗、心流條件、負向情緒與環境脈絡均屬中等，停留舒適圈群集（人數為81人，30%）的大學生又焦慮又疲憊，不想要高度的挑戰，不想要挑戰帶來的樂趣，對自己參與的活動有高度的控制感與清楚的目標就非常滿意，因此被命名為停留在舒適圈的一群。追求有意義人生群集的學生（人數為42人，15.4%）經常感受挑戰與技能平衡，聚焦且自主地追求樂趣、勇於接受挑戰，課堂學習活動與生產活動（例如：課外學習、打工、研究計畫等）的參與頻率最高。追求有意義人生的群集把每一節課堂(60分鐘)視為一個事件，能區分兩節課為兩個不同的有意義段落，而停留舒適圈的集群則把兩堂課的時間(100分鐘)視為同一事件，對各課堂的學習內容不太在意其意義的區別。無論停留舒適圈或追求有意義人生的群集，他們的生活滿意程度都很高，前者滿意停留於舒適圈，後者滿意於追求有意義的人生。三組中並沒有完美的自成目標人格組，但相較之下，追求有意義人生群集的特性與自成目標人格相似度較大。
綜合本研究結果，國小高年級學生下課時間的最佳經驗品質比上課時間為佳，以直接講述法教學的課堂為最差，呼應過去跨國評比研究(Mullis, Martin, Foy, & Arora, 2012)不斷發現我國學生學科成績不錯，但學習動機低落。活動式課程的最佳經驗是較好的，過去研究(如：Shernoff & Vandell, 2007) 也的確發現在參與體驗、實作中較容易產生忘我的經驗。國小學生學習的是較為基本的概念、技能，無論老師採用哪一種教學法，最重要的是讓學生感知到教學目標清楚、老師願意授予自主權、教學能引導專注力、讓學生了解課程重要性（以上均為心流條件），這樣就能引導出最佳經驗。小學生放學後的大部分時間還是要寫功課、學習，最佳經驗最好的不是在寫功課的時間，而是從事被動與主動休閒活動當下。大學生群體中有一小群稱為「期盼享受挑戰」，比較具有自成目標傾向，此一結果與 Asakawa (2004) 以日本大學生為受試的研究結果相似。
本研究採用DRM蒐集到事件階層的資料，也蒐集到個人階層的資料，以兩種階層性統計程序— HLM及MSEM同時分析兩個階層的資料，以變項為中心的統計可探求環境與個人的交互作用，以人為核心的統計取向，是基於既有的心理理論下尋求新的發現，可探求最佳經驗強弱的個別差異—自成目標傾向。啟動複雜的資料蒐集程序，優點在於可以蒐集到各課堂當下的經驗與動機，可以把多個同質課堂、同質環境的經驗與動機彙整加總(aggregate)，得到情境特定的動機(situation-specific motivation)，例如可以把所有用直接講述法教學的數學課之經驗與動機彙整成數學學習動機。這樣的資料可提供給老師詳細分析自己教學策略的成敗，提供的改進建議會比單一次問卷蒐集的資料更為豐富，老師有機會得到更多的證據尋找教學改進之道。本研究蒐集的兩份資料(小學生與大學生)都是以紙本問卷施測，資料輸入與資料檢誤耗工費時；未來如果要讓老師簡便地蒐集自己班上學生的情境式學習動機與學習成效，有必要發展數位版的DRM環境。美國學者建立對青少年發展的經驗取樣長期資料庫(Sloan Study of Youth and Social Development，1992-1997，主持人為 芝加哥大學的 Barbara Schneider教授)，良好的資料庫提供給學者申請使用，直到近十年仍有極多論文產出。作者歷經數年蒐集兩份密集式長期資料，誠摯地期盼我國有機會建立相似的資料庫，作為研究青少年發展的軟體建設。本論文最後提出未來研究的建議及教室教學與學習的實務建議。

Csikszentmihalyi (1975) identified flow, or optimal experience, “as a complex and highly structured state of deep involvement, absorption, and enjoyment”. Flow experience typically occurs when a person focuses on a set of clear goals that require appropriate skills. Many people describe such experience of effortless feeling as the best moments of their lives, saying, “It was like floating” or “I was carried by the flow.” The term “flow”, or “optimal experience,” was therefore proposed.
Flow typically occurs in the moments of activities and is highly influenced by external environments. Thus, previous researchers regard flow as a momentary experience, instead of a stable of personality (Seligman & Csikszentmihalyi, 2000). However, past studies have proved that some of individuals exactly often perceive the experience and further suggested that individual differences of flow existed. More specifically, the persons who have high frequency of flow have the personality of autotelic, which is associated with “a general curiosity and interest in life, persistence, low self-centeredness, and high creativity than others” (Nakamura & Csikszentmihalyi, 2002). However, most previous flow research has focused on adult experts’ sense of such effortless cognition/action in performing professional works or has examined teenagers’ optimal experience in daily life situations in western countries. Few but some studies have explored students’ optimal experience in high-school and elementary classroom settings (e.g., Schweinle, Meyer, & Turner, 2006; Schweinle et al., 2008). To our knowledge, there is also a lack of flow research conducted in Taiwan school environments.
This research aims to explore the flow in learning environments and attempts to expand the flow theory, modify measuring methods, and explore students’ optimal experience. In terms of relevant theories, the diverse operational definitions of flow have caused the fundamental problem of flow research. The present research defined optimal experience as a functional state assessed by students’ self-ratings of concentration, time distortion, satisfaction, and enjoyment in each specific class. The operational definition of flow used in the research is applicable not only to learning environments but also especially to elementary school students with younger ages. In addition, previous literature regarding research methods (e.g., Bradburn, Sudman, & Wansink, 2004) has criticized that traditional survey methods may have memory biases, where people may overestimate or underestimated their momentary experience. Optimal experience is the high engagement in an activity and it is necessary to ensure that students could continue to maintain such participation in their activities while collecting data. Therefore, a less disruptive data collecting method is more suitable for its momentary feature. The Day Reconstruction Method (DRM; Kahneman, Krueger, Schkade, Schwarz, & Stone, 2004b) has been developed recently and the current research consider that the use of the DRM in school settings is better than experience sampling method (ESM), which is well-known for its advantage to capture the moments when students were experiencing flow. The DRM is an improved survey method in order to explore how people plan their lives to spend their time and how they experience the various activities. Every morning, before participants engage in any activities, they are asked to recall the activities from the preceding day by generating a sequence of episodes in their diaries, e.g., “7: 30 to 8: 00 - waking up and cleaning” and “8:10 to 9:00 - participation of first-lesson in school.” After producing a diary consisting of contiguous episodes over the course of a day, the perceptual and contextual dimensions in each episode could be analyzed. Differing from traditional longitudinal data, which is gathered across longer time frame (e.g., cross-semester, cross-year) and/or by questionnaire, the use of the DRM includes some useful characteristics, such as gathering intensive longitudinal (Conner & Lehman, 2012), panel (repeat many times from the same one), and repeat-measured data.
The research consists of three studies and uses the DRM to collect intensive longitudinal data. The research recruited elementary school (N = 191) and college (N = 271) students to investigate students’ cognitive and affective experiences at school and out-of-school time. All the three studies capture the moments across school subjects taught with various instructional methods and across after-school daily activities. The research then uses the complementary approaches in order to clarify the relative contribution of interactions among students’ momentary experiences, contextual environments, and traits in specific activities so that both Hierarchical Linear Models (HLM; variable-centered approach) and Latent Profile Analysis (LPA; person-centered approach) are performed.
Study One has provided a relatively comprehensive analysis of optimal experience in across-course situations by conducting Hierarchical Linear Models (HLM). In three random days, 147 5th-graders answered questionnaires for each episode in the previous day; then, 2,288 episodes were collected. The episode data in school-day courses, where elementary school students perceived positive and negative subjective experience, were used in this study. These following viewpoints were considered in the analysis of HLM. Firstly, the perceptual factor of flow conditions was adopted in the episode situations. Another major contextual factor of the episode situations were instructional strategies. Finally, HLM was conducted to explore how the flow conditions and the instructional methods simultaneously impacted optimal experience while the individual level factors were controlled. The results showed that elementary school students perceived better quality of optimal experience during break time versus in activity, seat work, or lecture. Optimal experience varied much more across class episodes than among individuals. Surprisingly, optimal experience was higher when students perceived having high skill but experiencing low challenge. Four additional flow conditions were more effective than instructional methods and primary flow conditions proposed in previous research, in terms of the prediction of optimal experience.
The purpose of Study Two aims to use a multilevel mediation model to explore how motivation influenced flow conditions and optimal experience during after-school activities and further discussed the relationship between a trait-like motivation theory (i.e., self-determination theory) and flow theory. The current study, firstly, assumes that students’ after-school activities consist of four daily activities, which are learning, active leisure, passive leisure, and maintenance. Then, the current study uses the multilevel mediation model to explore whether students’ optimal experience (episode level rather than individual level) would be influenced by flow conditions (episode-level) and self-determination motivation (individual-level). Moreover, a total of 191 5th-graders reported 2,286 episodes by the DRM. The current study found that flow conditions in after-school active leisure significantly and completely mediated the effect of self-determination motivation on optimal experience. Students with higher self-determined motivation had better understanding about how to fully involve in active leisure, so that they could further experience effortless feeling as optimal experience of their activities. The results also indicated that students’ self-determination motivation had no significant indirect effect when students participated in maintenance.
The purpose of Study Three aims to use the latent profile analysis (LPA) to explore the college students with autotelic tendency. The main reason of using college students sample is that college students have more opportunities to plain and organize their daily schedules. This study differs from the others (variable-centered approach) because this is a person-centered study and focus on individual differences and consequences. The study uses the LPA to determine the optimal group numbers and cluster the variables of optimal experience and flow conditions, moods, and frequency in each daily activity in order to figure out homogeneous groups. Then, this study compares the indictors of autotelic tendency among the groups. A total of 271 college students reported 5,258 episodes by the DRM. The LPA results suggested three distinct profiles labeled median, staying in comfort-zone and seeking for meaningful-life. After comparing the indicators of autotelic tendency of each group, the author found that the students in the median group and staying in comfort-zone group provided similar results. However, the main difference between the median group and staying in comfort-zone group was that the staying in comfort-zone group students significantly had better life satisfaction than median group. In terms of the analysis of autotelic tendency of the seeking for meaningful-life group, the results indicated that students in this group had high life satisfaction and were willing to face and accept different kinds of challenges in their life. The evidences also showed that the students in the seeking for meaningful-life group could perceive and recognize the transfers among different subjects of courses and were also eager to participate in productive activities (e.g., participations of after-school learning, part-time work, and research project). According to the results, the author suggests that the students in the seeking for meaningful-life group conformed to the characteristics of autotelic tendency.