Example 15: Meta-Cognitive SATA

Economics - Policy Analysis

Scenario: Minimum Wage Policy Analysis

A government is considering raising the minimum wage by 50% (from $10/hour to $15/hour) to address income inequality. The proposal has strong political support but faces economic scrutiny. You are an economic policy analyst tasked with identifying likely economic effects and explaining the causal mechanisms connecting them. This is not about political opinion - it's about understanding economic cause-and-effect relationships regardless of your policy preferences.

Part 1: Which economic effects are LIKELY to occur? (Select ALL that apply)

Two-Part Assessment Structure

Part 1 (50%): Select all correct economic effects
Part 2 (50%): Explain the reasoning chain connecting your selections (minimum 3 cause-effect steps)

You cannot earn points for Part 2 unless you complete Part 1. Both parts required for full credit.

A. Some low-skill workers may face unemployment

B. Consumer prices in labor-intensive industries increase

C. All workers experience immediate income growth

D. Small businesses face higher labor costs

E. Inflation immediately spirals out of control

F. Workers above minimum wage may demand raises (wage compression)

G. GDP doubles within one year

H. All poverty is eliminated

Part 2: Explain Your Reasoning Chain

Explain the REASONING CHAIN connecting your selected answers. Show cause-and-effect relationships. Your explanation must include:

At least 3 cause-effect steps
Logical connections between economic concepts
Explanation of how each effect leads to the next
Recognition of tradeoffs and feedback loops

Sample Strong Response (Score: 95/100)

1. DIRECT EFFECT: Minimum wage increase from $10 to $15 → Higher labor costs for businesses (D)

2. BUSINESS RESPONSE: Higher labor costs → Businesses face 3 options:
  a. Raise prices to maintain margins (B)
  b. Reduce workforce to control costs (A)
  c. Accept lower profit margins

3. LABOR MARKET EFFECT: Some businesses choose option (b) → Unemployment increases for marginal workers whose productivity < $15/hour (A). This disproportionately affects low-skill workers, teenagers, and workers in areas with lower cost of living.

4. RIPPLE EFFECT: Workers earning slightly above new minimum wage ($16-$18/hour) see compression in wage hierarchy → Demand raises to maintain pay differentials (F). This creates upward wage pressure throughout the lower wage distribution.

5. FEEDBACK LOOP: Price increases (B) partially erode purchasing power gains for minimum wage workers, but net effect still positive for those who keep jobs. However, this creates distributional tradeoff: higher income for employed workers vs. unemployment risk for marginal workers.

KEY INSIGHT: The policy creates TRADEOFF between:
• Higher income for workers who keep jobs (BENEFIT)
• Unemployment risk for marginal workers (COST)
• Price increases affecting all consumers (COST)

Policy decision depends on weighing magnitudes of these effects and distributional preferences. Empirical evidence shows effects vary by region, industry, and size of wage increase.

Scoring Rubric (Two Components)

Component 1: Selection Accuracy (50%)

Correct Answers: A, B, D, F (4 out of 8 options)
Scoring: (Correct selections / 4) × 50%
Example: If you select all 4 correct = 50% score

Component 2: Reasoning Chain Quality (50%) - AI Assessed

Logical Connections (0-20 pts): Are cause-effect links valid?
Completeness (0-15 pts): Are all selections explained?
Economic Accuracy (0-20 pts): Correct application of principles?
Insight Depth (0-15 pts): Shows understanding of tradeoffs/complexity?
Coherence (0-10 pts): Reasoning flows logically?
Total: 80 points → Converted to 50% scale

AI Grading Process

LLM evaluates reasoning chain against rubric. Looks for: accurate economic mechanisms, logical flow, recognition of tradeoffs, avoidance of absolutes ("all", "never"), appropriate use of economic terminology, and demonstration of systems thinking.

SAMPLE STUDENT SCORES

Student A (Excellent): Part 1: 50% + Part 2: 47.5% (95/100 reasoning) = 97.5%

Student B (Good): Part 1: 37.5% (3/4) + Part 2: 40% (80/100 reasoning) = 77.5%

Student C (Weak): Part 1: 25% (2/4) + Part 2: 15% (30/100 reasoning) = 40%

Why Meta-Cognitive SATA Works

Prevents Lucky Guessing: Must explain WHY answers are connected, not just select them
Assesses Deep Understanding: Reveals causal reasoning and mental models
Metacognition: Students must articulate their thinking processes
Professional Realism: Mirrors real analysis work (justify recommendations)
Rich Diagnostics: Shows WHERE reasoning breaks down (which step in chain?)
Scalable Grading: AI can evaluate reasoning at scale with rubric
Higher-Order Thinking: Tests analysis, synthesis, evaluation (Bloom's taxonomy)
Transfer Assessment: Shows ability to apply concepts, not just recall them

Implementation Notes

Minimum Reasoning Length: 100 words (prevents "A causes B" type responses)
Maximum Reasoning Length: 500 words (prevents essays)
AI Grading Model: GPT-4 or Claude with detailed rubric
Human Review: Sample 10% of AI-graded responses for calibration
Inter-Rater Reliability: AI vs. human grading typically 0.85-0.90 correlation
Feedback: AI provides specific feedback on each rubric criterion
Time Allocation: Part 1 (2-3 min) + Part 2 (8-10 min) = 10-13 min total