For those of you are not familiar with what the Kirkpatrick model is, it is a very respected model for evaluating training and instruction. There are 4 (sometimes 5, depending on who you ask) levels to account for:
- Level 1 – Reaction: satisfaction. Can be gauged by a survey.
- Level 2 – Learning: knowledge absorption. Can by measured with a quiz.
- Level 3 – Behavior: Change in training-related behaviors from before instruction vs. after instruction.
- Level 4 – Results: How your instruction/training impacts the business.
- *Level 5 – ROI: How your instruction/training impacts the bottom line.
(*Level 5 is not always included, and combined with Level 4.)
Once you get past level 2 (learning), I’ve found that you often have to get creative, and methodologies for measurement often depend on the learner and situation. Not only that, but often at level 3 and level 4 (results), it gets easier to confuse one for the other. Plus, let’s face it, it’s pretty complicated to measure at these levels. So complicated, in fact, that a lot of programs aren’t measured at this level.
I had to become creative and remove the focus from paper surveys to metrics to understand the ask. For example, in a client support context, measuring effectiveness of an onboarding program, I have compared time to work independently pre-onboarding program to the time to work independently for those who went through onboarding to measure at level 4 because that has a direct impact on the business via productivity. Probably the most well-known form of a level 3 evaluation is a performance review based on pre-defined goals.
Performing a full performance context analysis is key to measuring both at level 3 and level 4. This tells you how the training is intended to be used by the learner and characteristics of the performance environment.