Who's Afraid of Level 4 Evaluation? A Practical Approach

Article excerpt

There is no escaping it: Increasingly, trainers are having to account for training dollars spent. And they are having to do it in terms of business results and return on investment.

In 1959, Donald Kirkpatrick proposed a four-level model of criteria for evaluating training: learner reactions, learning, job application, and observable business results. But many organizations still don't evaluate training. Others base their evaluations only on trainees' reactions. Now, more sophisticated assessments are called for, including Kirkpatrick's Level 4 evaluation - observable business results.

A 1988 poll of about 300 leading organizations, conducted by the American Society for Training and Development, found that only 20 percent evaluated training in terms of its economic effect on the organization. In fact, many training professionals seem to think that measuring training results in terms of dollars and cents takes too much time, is too costly, and is susceptible to extraneous factors that may affect results.

A 1990 IBM study of six large corporations (including IBM itself) and several training consultants, including Kirkpatrick, found that even organizations that say they examine the economic impact of training don't do so directly; they rely on people's opinions. Asking trainees whether training has improved their performance or their organization's performance isn't the same as assessing performance directly.

Why avoid Level 4 evaluations?

One reason organizations may shy away from Level 4 evaluations is that collecting and interpreting the data is more difficult and time-consuming than surveying trainees. But a Level 4 evaluation, which should include a cost-benefit analysis, can provide data that are more thorough and more credible than information collected by surveying trainees.

In some instances, conducting a Level 4 evaluation can be fairly easy. When data are routinely collected - such as the number of hours worked, units produced, and defects - a Level 4 evaluation may simply be a matter of obtaining, organizing, and analyzing already available data.

But sometimes organizations try to conduct Level 4 evaluations when they're not appropriate. A Level 4 assessment may not be the proper evaluation method for training that doesn't affect observable outcomes - for example, training that aims to change only attitudes. Such training isn't likely to show changes in organizational output. If a Level 4 evaluation is used in such a case, its inappropriateness is sure to reinforce any negative opinions about the Level 4 approach.

Another example of the inappropriate use of Level 4 evaluations is that of a training manager who promises to conduct a Level 4 study of a particular program and then discovers that objective data aren't available or accurate. Simply put, the manager can't finish what he or she started, and the Level 4 approach again looks impractical. Last, some people are intimidated by Level 4 evaluations. Level 4 studies often are expensive; if they don't produce positive results, trainers believe they'll be held accountable to management. In other words, if the objective data show that the training failed, the trainer can't avoid culpability.

For example, if a training program aims to increase the number of units produced and reduce the number of defects, it's difficult for a trainer to deny the conclusion of a pretraining/post-training assessment that shows the goals weren't attained. If the data show that more units weren't produced, with fewer defects, then the inescapable interpretation is that the training failed. The credibility associated with objective data is one of the reasons that Level 4 studies should be carefully planned.

Clearly, there are situations in which Level 4 evaluations aren't the appropriate assessments to use. But if you are considering using a Level 4 evaluation, here are some guidelines. …