A practical walkthrough of the four levels of the Kirkpatrick Model for evaluating training, with survey examples for measuring reaction, learning, behavior, and results.
Organizations invest enormous sums in training, yet many struggle to answer a basic question: did it work? The Kirkpatrick Model, first introduced by Donald Kirkpatrick in the 1950s and refined over the decades since, remains the most widely used framework for evaluating training effectiveness. It breaks evaluation into four levels that move from immediate reactions to lasting business impact. This guide walks through each level, shows how surveys fit at every stage, and explains how to use the model without falling into its common traps.
- Overview of the model
- Level 1: Reaction
- Level 2: Learning
- Level 3: Behavior
- Level 4: Results
- Applying the model in practice
- Frequently Asked Questions
Overview of the model
The Kirkpatrick Model evaluates training across four sequential levels. Level 1, Reaction, measures how participants felt about the training. Level 2, Learning, measures what they actually learned. Level 3, Behavior, measures whether they apply what they learned back on the job. Level 4, Results, measures the impact on business outcomes.
The levels build on one another. Positive reactions do not guarantee learning, learning does not guarantee changed behavior, and changed behavior does not automatically produce business results. Each level is harder and more expensive to measure than the one before, which is why so many programs stop at Level 1. The real value, however, lives in the higher levels.
A common refinement of the model is to plan in reverse. Rather than starting at Level 1 and hoping the chain holds, you begin by defining the Level 4 result you want, then work backward to the on-the-job behaviors that would produce it, the learning required to enable those behaviors, and finally the conditions that make the training itself land well. Designing backward keeps every level pointed at a real business need rather than at vanity metrics. It also helps you avoid the classic trap of celebrating glowing satisfaction scores while nothing actually changes in the workflow the training was meant to improve.
One more idea deserves emphasis before we dig into each level. Evaluation is not only about proving value after the fact. It is also a design discipline. Deciding up front how you will measure success forces you to articulate exactly what the training should change, which almost always makes the training itself sharper and more focused.
Level 1: Reaction
Level 1 captures participants' immediate response to the training. Did they find it relevant, engaging, and well-delivered? This is usually measured with a short post-session survey, sometimes called a smile sheet, completed right after the program ends.
Useful Level 1 questions go beyond satisfaction. Ask whether the content was relevant to the participant's role, whether they feel confident applying it, and whether they would recommend the training to a colleague. The confidence and relevance questions are especially predictive, because training that feels irrelevant rarely changes behavior no matter how enjoyable it was.
Modern practice distinguishes between several flavors of reaction. Satisfaction asks whether people enjoyed the session. Relevance asks whether the content connects to their actual work. Engagement asks whether they were actively involved rather than passively sitting through slides. Of these, relevance and intended application tend to forecast later behavior change far better than raw satisfaction, so weight your Level 1 questions toward them. A question such as "I plan to apply what I learned within the next month" captures intention, which is the first link in the chain that leads to real behavior change.
Treat Level 1 as a diagnostic rather than a scorecard to celebrate. If reaction scores are low, you have an early warning that the higher levels are unlikely to deliver, and you can investigate before investing in expensive follow-up measurement. If they are high, you have cleared the first hurdle but proven nothing about whether anything was actually learned or applied.
Keep Level 1 surveys brief and run them immediately while impressions are fresh. A simple survey template adapted for post-training feedback works well here. Remember that high satisfaction alone proves little. People can love a course and still learn nothing, so treat Level 1 as necessary but far from sufficient.
Level 2: Learning
Level 2 measures the actual gain in knowledge, skills, or confidence. The cleanest approach is a pre- and post-assessment: test participants before the training and again afterward, then compare. The difference isolates what the training added rather than what people already knew.
Assessments can take many forms, including quizzes, practical demonstrations, scenario-based questions, or self-rated confidence scales. For knowledge-heavy topics, objective quizzes work best. For skills, a demonstration or role-play scored against a rubric is more valid. The key principle is to measure against the specific learning objectives the training was designed to achieve.
Level 2 tells you whether learning happened in the room, but it cannot tell you whether that learning survives contact with daily work. For that, you need Level 3.
Level 3: Behavior
Level 3 is where many programs prove their worth or reveal their failure. It asks whether participants actually apply what they learned once they return to their jobs. Because behavior change takes time, Level 3 evaluation happens weeks or months after the training, not immediately.
Common methods include follow-up surveys to the learner and their manager, on-the-job observation, and reviewing performance metrics tied to the trained skill. A manager survey is particularly powerful because managers see whether new behaviors stick. Ask specific, observable questions: "Since the training, how often has this person applied the new technique?"
Level 3 also surfaces barriers. If people learned the material but are not using it, the problem may be a lack of time, tools, or managerial support rather than poor training. This insight is valuable, because it tells you the fix lies in the work environment, not the curriculum. Embedding short follow-up pulses into your broader employee engagement survey program is an efficient way to track behavior change over time.
Level 4: Results
Level 4 connects training to organizational outcomes such as productivity, quality, customer satisfaction, retention, or sales. This is the level executives care about most and the one hardest to measure, because many factors influence business results besides training.
The challenge is attribution. If sales rose after a sales-training program, was it the training, a new product, a seasonal trend, or all three? Strong Level 4 evaluation uses comparison groups, baselines, and trend data to build a credible, if rarely perfect, case. Rather than claiming the training caused a result outright, frame it as a contributing factor supported by evidence across all four levels.
Not every program needs Level 4 measurement. Reserve the heavy lifting for high-cost, high-stakes initiatives where demonstrating impact justifies the effort.
Applying the model in practice
A pragmatic approach is to apply the levels selectively. Run Level 1 for everything, since it is cheap. Add Level 2 for any program with clear learning objectives. Reserve the more demanding Levels 3 and 4 for your most important or expensive initiatives where proving impact matters.
Plan evaluation before the training begins, not after. Knowing how you will measure success clarifies what the training should achieve and what questions to ask at each level. Reusable survey templates keep your instruments consistent across programs so you can compare cohorts over time. Browse the HR survey templates library to standardize your post-training and follow-up surveys. Fast-scaling teams such as SaaS startups, which often onboard and upskill quickly, benefit from a lightweight but consistent evaluation habit that grows with them.
Frequently Asked Questions
Do I need to measure all four levels for every training? No. Level 1 is cheap and worth doing routinely. The higher levels cost more, so reserve them for important or expensive programs where proving impact is worth the effort.
When should I send a Level 3 behavior survey? Typically four to twelve weeks after the training, once participants have had real opportunities to apply what they learned. Sending it too early measures intention rather than actual behavior change.
How do I prove training ROI at Level 4? True ROI is difficult because many factors affect business outcomes. Use baselines, comparison groups, and trend data to build a credible case, and present training as a contributing factor rather than the sole cause.
Is the Kirkpatrick Model outdated? It has been refined over the years and remains widely used. Critics note it can imply a strict linear chain, so modern practice often emphasizes planning evaluation backward from desired results rather than treating the levels as a rigid sequence.
Measure your training impact with confidence. Build reaction, learning, and follow-up surveys in minutes. Create a free account or explore our survey templates to begin.