The Kirkpatrick Model

- Donald Kirkpatrick developed his four levels of evaluation model in the mid-1950s. Today, it's the most widely used model for measuring training. The four levels are level one, the participants' reaction to the training. Level two, what participants learned in the training. Level three, new on-the-job behaviors as a result of the training. And level four, the business results achieved because of the training. Let's take a closer look at each level, starting with level one. There's a good chance you've taken a survey after completing a training class. This is known as a level one evaluation. The idea is to gain insight into the training experience from the learner's perspective. Level one evaluations can help you improve training delivery and participant buy-in. It can also help you spot issues, such as a bug in an e-learning program. A level two evaluation measures the knowledge, skills, and attitudes developed as a result of the training. Here, we want to make sure that participants learned what they were supposed to have learned. Examples include post-tests, simulations, or hands-on assignments. Participants attending a class on running effective meetings might create a sample meeting agenda as part of a level two evaluation. Or if you trained a warehouse worker on how to package a shipment using the correct materials, you could have them package a sample shipment. A level three evaluation looks at whether participants apply the training back on the job. Examples include surveys, on the job observations, or quality inspections. A level three for a meeting management class might be a survey to ask participants what new skills they've used. A level three for training on packaging shipments might be a spot inspection of packages to make sure correct procedures are being followed. Finally, level four evaluations assess whether the training program goals were met. In other words, what business results have been achieved? For example, a business goal for a meeting management class might be to reduce the time people spend in meetings by 15%, or the goal for training on packaging shipments correctly might be to reduce damaged items by 30%. Now, there are a few considerations for using the Kirkpatrick model. First, the model works best when training outcomes are relatively isolated from other factors. So it might not be the best model to use for customer service training if customers are getting upset about a defective product that's outside of the participant's control. Another consideration is the degree of accuracy when you isolate the effects of training from other factors. For instance, participants might do well in a post-test because the training was great or because they already knew the answers. So should you do both a pre and a post-test to see what participants learned in training versus what they already knew? Pre-testing participants might eliminate the need for training altogether, if they can pass the test before attending the class. On the other hand, it takes extra work to isolate the impact of training by doing things like pre and post-testing. So it's not always worthwhile to invest in this extra step. Here's one last thought for using Kirkpatrick's four levels. 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. This will make it easier to connect the training to organizational goals.