The Best Way to Use the Kirkpatrick Model

May 26, 2017

The Kirkpatrick Model is a staple of learning and development teams to measure results, but it can easily be misused.

The Kirkpatrick Model – which was developed by Donald Kirkpatrick in 1955 for his Ph.D. dissertation – is one of the most commonly used methods to evaluate the effectiveness of learning solutions.

And yet, despite being very common, it’s also very easy to misuse.

The model is composed of four levels: reaction, learning, behavior and results. Your instinct is probably to follow them linearly, starting with how you want people reacting to the training you give them and ending with the result.

But that’s the exact wrong way to approach it, according to LinkedIn Learning Instructor Jeff Toister. A better method is the reverse.

“Start at level four and identify the results you want to achieve, then work backward to level three to think about what participants need to do on the job to achieve those results, and so on,” Toister said in his course, Measuring Learning Effectiveness. “This will make it easier to connect the training to organizational goals.”

A brief overview of the Kirkpatrick Model

Before we go any further, let’s delve into the fundamentals of the Kirkpatrick Model. As mentioned, it’s one of the most common ways learning and development (L&D) professionals and others measure the effectiveness of learning solutions. The model is composed of four levels, which get progressively more important.

Those four levels are:

1.    Reaction: How employees react to the training they receive. A common way to measure this is a survey after a training – what did the participant think of the training they received? Did they find it useful?

2.    Learning: Here’s what the employee actually learned from the training. Common ways to measure this are post-tests or hands-on assignments that demonstrate the person learned a new skill.

3.    Behavior: Next is to discover if participants actually used those new skills in their day-to-day jobs, i.e. incorporated it into their behavior. Common ways to measure this are in-field inspections or evaluations from participants’ managers.

4.    Results: Finally, what did this changed behavior result in? For example, if you had a management training, have your managers improved? Have employee surveys shown an improvement in manager effectiveness, thereby reducing retention?

The best way to use the Kirkpatrick Model

Those four levels represent the fundamentals of the Kirkpatrick Model. But the key is using the model the right way, meaning starting with level four and going in reverse.

Say you want to improve managers at your company. Perhaps your key metric is engagement – you believe that better managers will improve the engagement level of your employees.

That’s the fourth level of the Kirkpatrick Model, the result. From that, you can begin to work backwards on what you expect at each previous level.

So, continuing the example, you believe your managers will improve if they have more career conversations with their employees – a behavior, level three. So you create a learning solution around having that career conversation, which should result in them learning a new skill (level two).

Then think about how that might change their immediate response to the training, level one. It might not be so important that they thought the training was fun (although that never hurts).

Instead, it’s likely more important that they understand the importance of career conversations, and are bought-in on using them.

Is the Kirkpatrick Model the best way to measure learning effectiveness?

This is debatable. There are alternatives to the Kirkpatrick Model Toister covered in his course – the Phillips ROI Methodology (which is an extension of the Kirkpatrick Model) and the Brinkerhoff’s Success Case Method, along with countless others.

There are strengths and weaknesses to all these models. But here’s the reality: it’s not so much what model you choose, but instead how well you execute it.

Most of the time, the Kirkpatrick Model will work fine. It comes down to executing it correctly, and that boils down to having a clear idea of the result you want to achieve and then working backwards on how to achieve it.

The takeaway

Earlier this year, we surveyed more than 500 L&D professionals. One of the biggest challenges they face is showing the ROI of their learning solutions.

The model can help with that. By using it correctly – meaning, thinking through it backwards – you can show more ROI with your learning solutions. And that will give you a much higher chance of getting funding and appreciation for more learning solutions moving forward.