A research design is a plan for conducting an evaluation study. Research design is a complex topic and it is vital to HRD evaluation. It specifies the expected results of an evaluation study, the methods of data collection, and how the data will be analyzed. When evaluating any HRD effort, the researcher will want to have confidence that any changes observed after a program are due to the intervention, and not to some other factor (like changes in the economy or reward structure).

Unfortunately, it still remains quite typical that, if outcomes are measured at all, they are only collected after the training program has been completed. This approach to evaluation doesn’t give any indication of trainees’ initial skill or knowledge level, that is, where they started, so one can’t be certain that the outcomes attained were due to the training. To have greater confidence that the outcomes observed were brought about by the training (and not some other extraneous factor), the following practices should be included in a research design.

Pretest and post-test—including both a pretest and a post-test allows the trainer to see what has changed after the training. A control group is a group of employees similar to those who receive training, but they don’t receive training. This group receives the same evaluation measures as the group that is trained, which allows for a comparison of scores. Combining these two points creates what can be called the “pretest- post-test with control group” research design. We view this as the minimum acceptable research design for most training and HRD evaluation efforts.

Recent efforts to combine data from different research studies via meta-analysis have also helped determine the impact (or effect size) of various training interventions. Choices concerning evaluation also force the trainer or researcher to make difficult ethical decisions.