grad-cognitive-load
用認知負荷理論診斷學習者為何不堪負荷並重新設計教材
中文翻譯
--- name: "grad-cognitive-load" description: "運用認知負荷理論，在工作記憶的限制內管理內在、外在與增生負荷，最佳化教學設計。當使用者需要診斷學習者為什麼不堪負荷、為了更好的理解而重新設計訓練或文件、評估 UI/UX 資訊架構的認知負擔，或問到「為什麼這份教學讓人困惑」、「怎麼簡化複雜的說明」或「是什麼造成資訊過載」時使用。" metadata: category: "WP-29 心理學理論" tags: ["cognitive-load-theory", "Sweller", "working-memory", "instructional-design", "intrinsic-load", "extraneous-load", "germane-load"] --- # 認知負荷理論（CLT） ## 概述 認知負荷理論（Sweller, 1988）建立在人類認知的架構上：工作記憶的容量（7 +/- 2 個項目）和持續時間都嚴重受限，而長期記憶基本上是沒有上限的。有效的教學設計必須管理三種認知負荷：內在（任務複雜度）、外在（設計不良）和增生（schema 建構），讓總負荷不超過工作記憶的容量。 ## 何時使用 - 診斷學習者為什麼無法理解或記住教材 - 為了降低認知負擔，重新設計文件、教學或訓練計畫 - 評估介面設計、儀表板或資訊顯示是否過載 - 為複雜的學習教材排序，搭起 schema 習得的鷹架 ## 何時不要用 - 問題是動機而不是認知（學習者有能力處理，只是選擇不做） - 對專家受眾，他們已經有 schema，適用專業反轉效應 - 簡化會犧牲任務必要的擬真度（有些任務本質上就是無法簡化的複雜） ## 假設 ``` IRON LAW: Working memory capacity is FIXED and limited — instructional design must minimize extraneous load to maximize germane processing. Total load (intrinsic + extraneous + germane) must not exceed working memory capacity. ``` 關鍵假設： 1. 工作記憶處理新資訊；長期記憶儲存 schema，能繞過工作記憶的限制 2. 內在負荷由元素互動性決定，不改變任務就無法降低它 3. 外在負荷在設計者的掌控之下，永遠應該把它降到最低 ## 方法論 ### 步驟 1：分析元素互動性（內在負荷） 評估有多少資訊元素必須同時處理： - **低互動性**：元素可以獨立學習（單字表） - **高互動性**：元素必須整合起來才能理解（文法規則、電路設計） ### 步驟 2：找出外在負荷的來源 | 來源 | 說明 | 設計缺陷 | |--------|-------------|-------------| | 分散注意力（Split-attention） | 整合在空間/時間上分開的來源 | 文字離圖表很遠 | | 冗餘（Redundancy） | 以多種格式處理相同的資訊 | 旁白重複螢幕上的文字 | | 短暫資訊（Transient information） | 資訊在處理完成前就消失 | 快速的動畫沒有暫停 | | 專業反轉（Expertise reversal） | 對新手有幫助、對專家卻有妨礙的鷹架 | 強迫進階使用者走一步一步的流程 | | 誘人細節（Seductive details） | 有趣但無關的資訊 | 裝飾性的圖片、離題的故事 | ### 步驟 3：最佳化負荷的分配 管理總認知負荷的策略： - **範例（Worked examples）**：藉由展示已解出的問題，降低新手的內在負荷 - **漸隱（Fading）**：從範例逐步過渡到獨立解題 - **形式效應（Modality effect）**：使用雙通道（視覺 + 聽覺）來擴大有效的工作記憶容量 - **分段（Segmenting）**：把複雜的教材拆成由學習者自己掌握節奏的片段 - **預先訓練（Pre-training）**：在介紹互動之前，先教組成元素 - **消除冗餘**：移除各通道之間重複的資訊 ### 步驟 4：為增生負荷做設計 - 鼓勵自我解釋與精緻化 - 在練習題中使用變化，促進 schema 的抽象化 - 提供凸顯結構相似性的比較案例 - 把練習分散在不同時間（分散練習），以鞏固 schema ## 輸出格式 ```markdown ## Cognitive Load Analysis: [Context] ### Intrinsic Load Assessment - Element interactivity: [Low/Medium/High] - Key interacting elements: [list] - Learner expertise level: [Novice/Intermediate/Expert] ### Extraneous Load Audit | Source | Present? | Severity | Fix | |--------|----------|----------|-----| | Split-attention | [Yes/No] | [High/Med/Low] | [solution] | | Redundancy | [Yes/No] | [High/Med/Low] | [solution] | | Transient info | [Yes/No] | [High/Med/Low] | [solution] | | Seductive details | [Yes/No] | [High/Med/Low] | [solution] | ### Load Budget - Estimated total load: [Within/Exceeding capacity] - Extraneous reduction potential: [High/Medium/Low] ### Redesign Recommendations 1. [Primary extraneous load reduction] 2. [Segmenting or sequencing change] 3. [Germane load enhancement] ``` ## 常見陷阱 - 專業反轉效應意味著，對新手最理想的設計會積極地傷害專家，所以需要適應性或分層的設計 - 「7 +/- 2」是粗略的經驗法則，對新的互動元素來說，有效的工作記憶容量可能低到只有 3-4 個組塊 - 增生負荷在近期的文獻中有爭議，有些研究者把它歸到內在負荷的管理之下，而不是當成獨立的一類 - 降低外在負荷永遠是有益的；降低內在負荷則可能過度簡化，阻礙深度學習 - 形式效應只有在視覺和聽覺通道傳遞的是互補（不是冗餘）的資訊時才適用 - 認知負荷很難直接衡量，各種替代指標（表現、主觀評分、次要任務）各有其限制 ## 參考資料 - Sweller, J. (1988). Cognitive load during problem solving: effects on learning. *Cognitive Science*, 12(2), 257-285. - Sweller, J., Ayres, P. & Kalyuga, S. (2011). *Cognitive load theory*. Springer. - Paas, F., Renkl, A. & Sweller, J. (2003). Cognitive load theory and instructional design: recent developments. *Educational Psychologist*, 38(1), 1-4.
英文原文
--- name: "grad-cognitive-load" description: "Apply Cognitive Load Theory to optimize instructional design by managing intrinsic, extraneous, and germane load within working memory limits. Use this skill when the user needs to diagnose why learners are overwhelmed, redesign training or documentation for better comprehension, evaluate UI/UX information architecture for cognitive burden, or when they ask 'why is this tutorial confusing', 'how to simplify complex instructions', or 'what causes information overload'." metadata: category: "WP-29 心理學理論" tags: ["cognitive-load-theory", "Sweller", "working-memory", "instructional-design", "intrinsic-load", "extraneous-load", "germane-load"] --- # Cognitive Load Theory (CLT) ## Overview Cognitive Load Theory (Sweller, 1988) is grounded in the architecture of human cognition: working memory is severely limited in capacity (7 +/- 2 items) and duration, while long-term memory is essentially unlimited. Effective instructional design must manage three types of cognitive load — intrinsic (task complexity), extraneous (poor design), and germane (schema construction) — so that total load does not exceed working memory capacity. ## When to Use - Diagnosing why learners fail to comprehend or retain instructional material - Redesigning documentation, tutorials, or training programs for reduced cognitive burden - Evaluating interface design, dashboards, or information displays for overload - Sequencing complex learning material to scaffold schema acquisition ## When NOT to Use - When the problem is motivational rather than cognitive (learner can process but chooses not to) - For expert audiences where schemas already exist and the expertise reversal effect applies - When simplification would compromise essential task fidelity (some tasks are irreducibly complex) ## Assumptions ``` IRON LAW: Working memory capacity is FIXED and limited — instructional design must minimize extraneous load to maximize germane processing. Total load (intrinsic + extraneous + germane) must not exceed working memory capacity. ``` Key assumptions: 1. Working memory processes novel information; long-term memory stores schemas that bypass WM limits 2. Intrinsic load is determined by element interactivity — it cannot be reduced without changing the task 3. Extraneous load is under the designer's control and should always be minimized ## Methodology ### Step 1 — Analyze Element Interactivity (Intrinsic Load) Assess how many information elements must be processed simultaneously: - **Low interactivity**: elements can be learned independently (vocabulary lists) - **High interactivity**: elements must be integrated to be understood (grammar rules, circuit design) ### Step 2 — Identify Extraneous Load Sources | Source | Description | Design Flaw | |--------|-------------|-------------| | Split-attention | Integrating spatially/temporally separated sources | Text far from diagram | | Redundancy | Processing identical information in multiple formats | Narration duplicating on-screen text | | Transient information | Information disappears before processing completes | Fast animations without pause | | Expertise reversal | Scaffolding that helps novices but hinders experts | Forced step-by-step for advanced users | | Seductive details | Interesting but irrelevant information | Decorative images, tangential stories | ### Step 3 — Optimize Load Distribution Strategies to manage total cognitive load: - **Worked examples**: reduce intrinsic load for novices by showing solved problems - **Fading**: gradually transition from worked examples to independent problem-solving - **Modality effect**: use dual channels (visual + auditory) to expand effective WM capacity - **Segmenting**: break complex material into learner-paced segments - **Pre-training**: teach component elements before introducing interactions - **Eliminate redundancy**: remove duplicate information across channels ### Step 4 — Design for Germane Load - Encourage self-explanation and elaboration - Use variability in practice problems to promote schema abstraction - Provide comparison cases that highlight structural similarities - Space practice over time (distributed practice) for schema consolidation ## Output Format ```markdown ## Cognitive Load Analysis: [Context] ### Intrinsic Load Assessment - Element interactivity: [Low/Medium/High] - Key interacting elements: [list] - Learner expertise level: [Novice/Intermediate/Expert] ### Extraneous Load Audit | Source | Present? | Severity | Fix | |--------|----------|----------|-----| | Split-attention | [Yes/No] | [High/Med/Low] | [solution] | | Redundancy | [Yes/No] | [High/Med/Low] | [solution] | | Transient info | [Yes/No] | [High/Med/Low] | [solution] | | Seductive details | [Yes/No] | [High/Med/Low] | [solution] | ### Load Budget - Estimated total load: [Within/Exceeding capacity] - Extraneous reduction potential: [High/Medium/Low] ### Redesign Recommendations 1. [Primary extraneous load reduction] 2. [Segmenting or sequencing change] 3. [Germane load enhancement] ``` ## Gotchas - The expertise reversal effect means that designs optimal for novices actively harm experts — adaptive or layered design is necessary - "7 +/- 2" is a rough heuristic; effective WM capacity for novel interacting elements may be as low as 3-4 chunks - Germane load is debated in recent literature — some researchers subsume it under intrinsic load management rather than treating it as separate - Reducing extraneous load is always beneficial; reducing intrinsic load may oversimplify and prevent deep learning - Modality effect applies only when visual and auditory channels carry complementary (not redundant) information - Cognitive load is difficult to measure directly — proxy measures (performance, subjective ratings, secondary tasks) each have limitations ## References - Sweller, J. (1988). Cognitive load during problem solving: effects on learning. *Cognitive Science*, 12(2), 257-285. - Sweller, J., Ayres, P. & Kalyuga, S. (2011). *Cognitive load theory*. Springer. - Paas, F., Renkl, A. & Sweller, J. (2003). Cognitive load theory and instructional design: recent developments. *Educational Psychologist*, 38(1), 1-4.