2nd/10_Wiki/Topics/AI_and_ML/Cognitive-Evaluation-Theory.md

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

id	title	category	status	canonical_id	aliases	duplicate_of	source_trust_level	confidence_score	verification_status	tags	raw_sources	last_reinforced	github_commit	tech_stack
wiki-2026-0508-cognitive-evaluation-theory	Cognitive Evaluation Theory (Self-Determination)	10_Wiki/Topics	verified	self	self-determination theory SDT autonomy competence relatedness intrinsic motivation over-justification	none	A	0.88	applied	psychology motivation sdt intrinsic-motivation gamification game-design education productivity		2026-05-10	pending	language applicable_to psychology Game Design Education Workplace Product UX

Cognitive Evaluation Theory / SDT

매 한 줄

"매 reward 의 sometimes 의 passion 의 kill". Deci & Ryan 의 SDT (Self-Determination Theory) 의 sub-theory. 매 autonomy + competence + relatedness 의 3 의 intrinsic motivation 의 fuel. 매 modern gamification, 매 education, 매 product UX 의 base.

매 핵심

Self-Determination Theory 의 3 needs

1. Autonomy: 매 self-chosen.
2. Competence: 매 mastery feel.
3. Relatedness: 매 connection.

→ 매 3 의 satisfy = 매 intrinsic motivation.

Intrinsic vs Extrinsic

• Intrinsic: 매 enjoyment / interest 의 itself.
• Extrinsic: 매 reward / punishment.
• Internalization: 매 extrinsic 의 internalize 의 spectrum.

Over-justification effect (Lepper)

• 매 paid for activity 의 enjoy → 매 intrinsic 의 lose.
• 매 famous: 매 kid drawing experiment.
• 매 implication: 매 reward 의 careful design.

매 reward 의 effect

Type	Effect
Tangible + expected	매 intrinsic ↓
Verbal positive	매 intrinsic ↑
Unexpected	매 less harmful
Task-contingent	매 intrinsic ↓
Performance-contingent (informative)	매 mid
Choice-supportive	매 intrinsic ↑

매 application

Game design

• Autonomy: 매 quest choice, 매 customization.
• Competence: 매 progression, 매 skill curve.
• Relatedness: 매 multiplayer, 매 guild.

Education

• Autonomy: 매 project topic choice.
• Competence: 매 scaffolded difficulty.
• Relatedness: 매 peer collab.

Workplace

• Autonomy: 매 schedule, 매 method.
• Competence: 매 challenging task + feedback.
• Relatedness: 매 team belonging.

Product UX

• Autonomy: 매 customization, 매 control.
• Competence: 매 onboarding 의 mastery.
• Relatedness: 매 social feature.

매 dark side (anti-pattern)

• 매 tangible reward 의 dominate (cash, gift card).
• 매 surveillance / monitoring → autonomy ↓.
• 매 forced ranking → relatedness ↓.
• 매 over-difficult / under-difficult → competence ↓.

매 modern AI 의 의미

• AI assistant: 매 user 의 competence 의 reduce 의 risk.
• Recommendation: 매 autonomy 의 illusion vs reality.
• Gamification: 매 manipulation 의 risk (dark pattern).
• Productivity tracker: 매 surveillance.

💻 패턴 (응용)

Game progression (autonomy + competence)

class QuestSystem {
  // 매 autonomy: 매 player 의 choose
  getAvailableQuests(player) {
    return this.quests.filter(q => 
      q.unlock(player) && 
      !q.completed(player)
    );  // 매 multiple option
  }
  
  // 매 competence: 매 skill curve
  recommendNextQuest(player) {
    const available = this.getAvailableQuests(player);
    const skillLevel = player.estimatedSkill();
    
    // 매 zone of proximal development
    return available
      .filter(q => q.difficulty >= skillLevel - 1 && q.difficulty <= skillLevel + 2)
      .sort((a, b) => a.difficulty - b.difficulty)[0];
  }
}

Workplace autonomy (4-day workweek-style)

def autonomy_audit(team):
    return {
        'schedule_flexibility': measure(team, 'self-set start/end times'),
        'method_flexibility': measure(team, 'choose how to solve'),
        'topic_flexibility': measure(team, 'pick what to work on'),
        'tool_flexibility': measure(team, 'choose tools'),
        'location_flexibility': measure(team, 'remote / hybrid'),
    }

Educational scaffolding (competence)

def adaptive_difficulty(student, current_skill, performance):
    """매 zone of proximal development."""
    if performance > 0.85:
        return current_skill + 1  # 매 challenge ↑
    elif performance < 0.5:
        return current_skill - 1  # 매 ease
    return current_skill  # 매 stable

Avoid over-justification (verbal > tangible)

def reward_employee(employee, accomplishment):
    # 매 ❌ Tangible + expected (e.g., $100 for X)
    # 매 ✅ Verbal + specific
    feedback = f"""
    {employee.name}, your work on {accomplishment.project} was excellent.
    Specifically, your approach to {accomplishment.specific_thing} showed
    {accomplishment.competence_demonstrated}. This had {accomplishment.impact}.
    """
    
    # 매 unexpected appreciation 의 OK
    if random.random() < 0.1:
        send_appreciation_card(employee, feedback)
    
    return feedback

Recommendation system 의 autonomy preserve

def recommend_with_autonomy(user, items):
    """매 explainability + 매 user control."""
    recommendations = ml_model.recommend(user, items)
    
    return {
        'items': recommendations,
        'why': explain_each(recommendations, user),  # 매 transparency
        'controls': {
            'less_of_this': lambda item: user.feedback_negative(item),
            'more_of_this': lambda item: user.feedback_positive(item),
            'turn_off_personalization': lambda: user.toggle_personalization(False),
        },
    }

Mastery curve (competence)

def mastery_journey(skill):
    return [
        ('Novice', 'Hand-holding tutorial', 'high feedback'),
        ('Advanced Beginner', 'Guided practice', 'frequent feedback'),
        ('Competent', 'Independent task', 'less feedback'),
        ('Proficient', 'Complex challenge', 'occasional feedback'),
        ('Expert', 'Mastery + teaching', 'self-feedback'),
    ]

Anti-surveillance (autonomy)

# 매 ❌ Productivity tracker showing keystrokes
# 매 ✅ Self-tracker only the user sees

class SelfProductivityTracker:
    """매 user-only, opt-in."""
    def __init__(self, user_id):
        self.user_id = user_id
        self.private = True  # 매 not shared with manager
    
    def log_focus_session(self, duration):
        self.sessions.append({'duration': duration, 'date': now()})
    
    def insights(self):
        # 매 user 의 only
        return generate_insights(self.sessions)

🤔 결정 기준

상황	Approach
Game	Autonomy (quest choice) + Competence (curve) + Relatedness (multiplayer)
Education	Project choice + scaffolded + peer
Workplace	Schedule + method + team
Product onboard	Mastery feel + control
Reward	Verbal + unexpected > tangible expected
Productivity tool	Self-tracker > surveillance

기본값: 매 3 needs 의 audit + 매 reward design 의 careful.

🔗 Graph

• 부모: Psychology · Motivation · Behavioral-Economics
• 변형: Self-Determination-Theory · Over-Justification-Effect · Flow-State
• 응용: Gamification · Game-Design · Education-Design · UX
• Adjacent: Cognitive-Biases · Addiction-Neuroscience · Authenticity · Anthropomorphism
• 사상가: Edward-Deci · Richard-Ryan · Mark-Lepper

🤖 LLM 활용

언제: 매 motivation design. 매 game design. 매 education tool. 매 workplace policy. 매 product UX. 언제 X: 매 manipulation / dark pattern.

❌ 안티패턴

• Tangible expected reward 의 default: 매 intrinsic 의 kill.
• Cash for fun: 매 over-justification.
• Surveillance + autonomy 의 claim: 매 violation.
• Difficulty curve X: 매 competence ↓.
• No social feature (suitable game): 매 relatedness X.
• Dark pattern (FOMO, sunk cost): 매 ethics violation.

🧪 검증 / 중복

• Verified (Deci-Ryan SDT, Pink "Drive", Lepper over-justification).
• 신뢰도 A.
• Related: Cognitive-Biases · Addiction-Neuroscience · Game-Design · Authenticity · Bounded-Rationality.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — 3 needs + over-justification + 매 quest / autonomy / mastery code