Kirkpatrick's Four Levels of Training Evaluation: The Foundation of Learning Measurement

The Model That Started It All

In 1959, Donald Kirkpatrick introduced a deceptively simple question: How do we know if training actually worked? His answer—a four-level evaluation model—became the most widely used framework for measuring training effectiveness worldwide.

More than six decades later, the Kirkpatrick Model remains the foundation upon which most learning evaluation approaches are built. Understanding it is essential for anyone serious about measuring educational impact.

The Four Levels

Level 1: Reaction

What did participants think of the training? This level captures the immediate emotional and cognitive response to the learning experience.

Key questions:

• Did learners find the content relevant?
• Was the instructor/facilitator effective?
• Would they recommend it to colleagues?
• Did the format and delivery meet their needs?

Common methods: Post-course surveys, smile sheets, Net Promoter Score, focus groups

Important note: Positive reactions don't guarantee learning, but negative reactions often predict disengagement.

Level 2: Learning

Did participants acquire the intended knowledge, skills, or attitudes? This level measures actual learning outcomes, not just satisfaction.

Key questions:

• Can learners demonstrate new knowledge?
• Have skills improved measurably?
• Have attitudes or confidence shifted?

Common methods: Pre/post assessments, skills demonstrations, knowledge checks, simulations, case studies

Best practice: Always establish a baseline (pre-test) to demonstrate change, not just achievement.

Level 3: Behavior

Are participants applying what they learned on the job? This is often called the "transfer" level—where classroom learning meets workplace reality.

Key questions:

• Are learners using new skills in their work?
• How consistently are new behaviors applied?
• What barriers are preventing application?

Common methods: Observation, manager feedback, self-assessment, 360 reviews, performance data, work samples

Critical insight: Level 3 requires time—behavior change doesn't happen immediately. Plan to measure 30-90 days post-training.

Level 4: Results

Did the training impact organizational outcomes? This level connects learning to business or mission results.

Key questions:

• Did quality metrics improve?
• Were costs reduced or revenue increased?
• Did customer/patient satisfaction improve?
• Were strategic goals advanced?

Common methods: Business metrics analysis, quality indicators, financial data, customer satisfaction scores

The challenge: Isolating training's contribution from other variables affecting results.

The New World Kirkpatrick Model

In recent years, Jim and Wendy Kirkpatrick (Donald's son and daughter-in-law) updated the model with several key enhancements:

• Start with Level 4: Begin by defining desired results, then work backward to design training that achieves them
• Required drivers: Recognize that processes, tools, and reinforcement are needed to support behavior change
• Leading indicators: Identify early signals that predict eventual results
• Return on Expectations (ROE): Focus on stakeholder expectations rather than just ROI calculations

Pros of Kirkpatrick's Model

• Simplicity: Four levels are easy to understand and communicate
• Universal applicability: Works across industries, topics, and delivery methods
• Widespread adoption: Common language for L&D professionals worldwide
• Logical progression: Clear hierarchy from reaction to results
• Flexibility: Can be implemented partially based on resources and needs
• Stakeholder-friendly: Easy to explain to executives and sponsors

Cons and Criticisms

• Linear assumption: Implies positive reactions lead to learning, which leads to behavior—but research shows these correlations are weak
• Transfer gap: Doesn't explain how to bridge Level 2 (learning) to Level 3 (behavior)
• Time and cost: Levels 3 and 4 require significant resources to measure properly
• Attribution difficulty: Isolating training's impact at Level 4 is methodologically challenging
• Not healthcare-specific: Lacks nuances important for clinical education (e.g., competence vs. performance)
• Satisfaction overemphasis: Organizations often stop at Level 1, creating false confidence

Alternatives to Consider

Moore's Expanded Outcomes Framework

Specifically designed for CME and healthcare education with seven levels, including distinctions between declarative and procedural knowledge, and extending to patient and community health outcomes. Better suited for clinical education contexts.

Phillips ROI Model

Adds a fifth level focused on return on investment, providing explicit financial analysis. Useful when cost justification is paramount, though requires rigorous methodology to be credible.

CIPP Model (Context, Input, Process, Product)

Focuses on program improvement rather than just summative evaluation. Good for formative assessment and iterative program development.

Brinkerhoff's Success Case Method

Identifies and analyzes the most and least successful cases to understand what enables or prevents success. Highly practical for understanding barriers to transfer.

When Kirkpatrick Works Best

Kirkpatrick's model is ideal when:

• You need a simple, recognized framework to start measuring
• Stakeholders are new to learning evaluation concepts
• You're evaluating general corporate training programs
• Resources for evaluation are limited
• You need a common language across the organization

When to Consider Alternatives

Look beyond Kirkpatrick when:

• You're in healthcare/clinical education (consider Moore's)
• Financial ROI justification is critical (consider Phillips)
• You need to understand why transfer is or isn't happening
• Complex behavior change with multiple variables is involved
• Accreditation requires specific outcomes measurement

Practical Implementation Tips

1. Don't stop at Level 1: Satisfaction data is easy to collect but tells you little about actual impact
2. Use pre-tests: Without baseline data, you can't demonstrate learning gain at Level 2
3. Plan for Level 3 delays: Build follow-up assessments into your timeline—behavior change takes time
4. Involve managers: Supervisors are critical for both supporting behavior change and observing it
5. Define Level 4 metrics upfront: Know what "results" you're targeting before training begins
6. Accept imperfection: Perfect attribution at Level 4 is often impossible—focus on reasonable evidence

The Bottom Line

Kirkpatrick's Four Levels remain foundational because they ask the right questions in a logical sequence: Did they like it? Did they learn it? Do they use it? Did it matter?

While newer models address its limitations—particularly for healthcare education and behavior transfer—understanding Kirkpatrick is essential for anyone in learning and development. It provides the conceptual vocabulary and basic framework upon which more sophisticated evaluation approaches are built.

Start with Kirkpatrick. Then evolve your approach as your evaluation maturity—and your organization's appetite for evidence—grows.