Why You Should Encourage Your Employees to Break Stuff (In Training)

Error Avoidance Training

In most training programs, we discourage employees from making errors. We start by presenting learners with a procedure and then encourage them to try it themselves without committing an error. If they do make a mistake, we punish them with negative feedback. The assumption is that if learners avoid errors in training, they’ll also be less likely to commit errors back on the job.

The legendary football coach Vince Lombardi summed up the Error Avoidance Training (EAT) philosophy when he said, “Practice doesn’t make perfect. Perfect practice makes perfect.” In other words, if you don’t want learners to make mistakes on the job, you should encourage them not to make mistakes in training, either.

You probably already know what EAT looks like in practice because it’s the default approach to training pretty much everywhere. Let’s look at an example anyway, just to make the point crystal clear:

Anika is training supervisors to enter performance data in a new system. First, she demonstrates exactly how to enter the data, explaining each step as she goes. Then, she gives each supervisor a manual that describes the entire procedure — step by step — just as she demonstrated it. Finally, she asks the supervisors to follow the procedure and enter data in the performance system exactly as described in the job aid. If they make a mistake, Anika says something like, “Sorry, that’s not right. Try that one again.”

Error Management Training

In Error Management Training (EMT), we provide employees with minimal guidance and encourage them to explore on their own. We don’t just allow learners to make errors, we encourage them to do so. The assumption behind EMT is that employees will learn more from making mistakes than they would from following step-by-step instructions.

Let’s go back to our software training example and see how we’d deliver the same material in an EMT format:

Rashon is training supervisors to enter performance data in a new system. First, he hands out a bare-bones job aid describing how to login to the system and a few basic tasks. Then, he asks them to login to a “sandbox” version of the software loaded with fake data and “play around.” When they make mistakes, he responds with, “Good! The more mistakes you make, the more you’ll learn.”

Although we’re using software training as an example, the EAT and EMT approaches can be applied to just about any training topic: soft skills, business acumen, technical training, you name it.

Which is Better?

The “holy grail” of learning and development is training transfer — the extent to which employees apply what they’ve learned in training back on the job. No matter how much knowledge or skill an employee acquires in training, it’s a wasted effort if they don’t change their behavior when they get back to work. So which approach maximizes training transfer, EAT or EMT?

Researchers have been trying to answer that question for a few decades. Their results have been mixed. In some studies, EMT blows EAT out of the water. In others, there doesn’t seem to be much of a difference at all.

In 2008, researchers Nina Keith and Michael Frese tried to make sense of all this conflicting data by conducting a meta-analysis of studies comparing EMT to other approaches. (A meta-analysis is a statistical technique for combining the results of multiple studies.) Keith and Frese analyzed 24 different studies with a combined sample size of 2,183 participants. Here’s what they found.

First, on average, EMT leads to better training outcomes than EAT. Encouraging learners to make mistakes in training leads to greater training transfer than discouraging mistakes. Rashon’s students — who played around in the sandbox and were encouraged to make mistakes — will probably apply more of what they’ve learned in class to the job than Anika’s students, who learned to follow a specific procedure and were given negative feedback for committing errors.

Second, the superiority of EMT only shows up in on-the-job performance, not performance during training. When we measure performance in the training itself, we can expect EAT to produce better results than EMT. (After all, we’re encouraging the EMT learners to make mistakes.) When we look at actual on-the-job performance, however, EMT learners outperform EAT learners. In other words, Anika’s students would probably score a little higher on an end-of-course exam than Rashon’s. But when it comes to actually entering performance scores in the new system back on the job, Rashon’s students will outperform Anika’s.

Third, the superiority of EMT was especially pronounced for new tasks that differed somewhat from the task taught in training. When employees acquire knowledge and skills in a training program, we expect them to be able to apply them to exact task addressed in training. Very often, however, employees apply what they learn in training to tasks other than those presented in training as well. Suppose, for example, that the user interface of the performance management software was upgraded six months after Anika and Rashon delivered their training. Anika’s students — who learned the exact procedure for entering performance data in the old system — would struggle to adapt to the new user interface. Rashon’s students, on the other hand, would better be able to apply what they learned in their class to the new interface.

This is one of the reasons researchers were getting mixed results. Some studies measured performance in training and others measured performance on-the-job. Some studies measured performance on the exact task addressed in training and other studies measured performance on different, but related tasks as well.

Keith and Frese identified another reason for the mixed results: different researchers had different ideas about what EMT really is. Some researchers believed, for example, that EMT meant simply allowing learners to make mistakes while others believed that EMT required instructors to encourage mistakes.

As part of their analysis, Keith and Frese identified the two “active ingredients” in EMT — the two conditions that must be met to get the full benefit of EMT. First, learners must be encouraged to carry out active exploration. In other words, they must be allowed to “play in the sandbox” instead of just following a script. Second, mistakes must be actively encouraged, not just tolerated. Learners should be positively reinforced for committing errors in training.

Why Aren’t We Using EMT All the Time?

If EMT is so great, why isn’t it the norm? The researchers didn’t address this question, but we can make some educated guesses by drawing on our HR experience. When Anika proposed her training plan, it was probably accepted by her leadership with little objection. Anika can create the training in a few days and deliver it in about an hour. It won’t cost much more than Anika’s time and the cost of printing the manuals. Evaluation will be cheap and easy, too — just a short quiz at the end the course. Most importantly, it’s exactly what everyone expects. Employees may not love it, but they’re not going to fight it, either.

Now imagine being a fly on the wall as Rashon pitches his course to senior management:

The idea behind this course is to encourage employees to play around with the system and make mistakes, so I’m going to need additional budget and I.T. support to set up a sandbox instance of the software. The class itself will probably run at least a few hours because employees will need time to explore at their own pace. Also, the only way we’re going to know if this course worked is to measure actual on-the-job behavior, so I’ll also need additional budget and time to make that happen. One last thing: you should prepare yourselves for massive pushback on this concept. It’s going to be a huge change for the trainees and their managers are going to question the wisdom of encouraging employees to make mistakes. Sound good?

What’s the Takeaway?

We have good evidence that, on average, EMT leads to better learning outcomes than EAT. In a perfect world with unlimited training budgets, an open culture, and supportive leadership, almost all our training would be EMT. In the real world, however, you’ll have to exercise your best judgement to determine when and where to introduce EMT, taking into consideration time, budget, effort, and cultural factors.

One potential starting point is to identify a new process or system that is likely to change substantially over time. In that case, stakeholders are likely to realize that standard procedural training won’t be sufficient and may be open to alternative approaches.

Source

Keith, N., & Frese, M. (2008). Effectiveness of Error Management Training: A Meta-Analysis. Journal of Applied Psychology.