SPLIT SAMPLE VALIDATION OF HUMAN RESOURCES PROGRAMS
Trying to convince your management team that a new human resources program really will do what it is intended to do?
I recently read about a scientific approach suggesting the use of split samples
– splitting your target population into two groups and training only one group. Then after an established period of time passes, you measure the difference in performance of the two groups. The expectation being that the trained group will show a notable improvement.
I remember a friend telling me about a program they implemented several years ago. I mentioned the topic and its famous leader Jack Stack a few blogs back. It’s called Open Book Management.
In this case my friend’s organization had about 120 field offices. Of those, twelve were selected to participant in the implementation of the Open Book Management program. If in six month’s time there was evidence that the program worked – that is, there were noticeable cost savings — and employees willingly participated, then other locations would be brought on board.
It did and they did.
Now for split sampling to work, there have to be certain critical steps taken into consideration. First, the subject group has to be random. Second, nothing about the split sample should be made public and nothing else should be changed. Finally, you need to be sure to use jobs where the output in quality, quantity and dollars are already measured on a regular basis. That’s why sales is generally considered a good place to start.
So the next time you want to present a program to your management team and need to be able to justify the cost versus the return on investment, you can consider how a split sample validation might be useful to prove your point.
Explore posts in the same categories: Company Culture, Leadership
