« Beating the Program: HR Simulation Ain't no Thang | Main | Ethical My A** »
November 05, 2005
HR Again - Linear Relationships will not Help You in a 3D World
My previous post explained the relative benefits of linear correlations to the ultimate decisions made in our HR simulations. There is a severe disadvantage, however. We are missing multi-dimensional [so to speak] relationships. Thus, while we can see positive correlations between two variables, we miss the "ghost" variables. In such cases, non-correlated graphs, many of which had been discarded due to lack of significance, may indeed have correlation when three or more variables are considered. Currently, the graphs are only reinforcing what the team and I have already guessed with our instincts. Yes, increasing the quality budget will increase quality. Ok, that was obvious, and we can move on...right?
The most effective graphs however, do more than simply reinforce expectations. I want to look at a graph and say "wait a minute, this variable was what was impacting our stagnation in morale...Great! let's crack open a bottle of...Martinelli's Sparkling Apple Cider...and celebrate!" In the coming quarters I want to find those hidden variables affecting relationships. I sincerely hope it is just a matter of putting more variables on the same graph and more deeply analyzing the relationships. Otherwise it is time I review my high school math and find the proper graphs to expose multiple relationships within multiple variables.
Posted by jxw115 at November 5, 2005 08:56 PM
Trackback Pings
TrackBack URL for this entry:
http://blog.case.edu/johnson.wei/mt-tb.cgi/3842
Comments
I agree with want you are saying. There are definately many factors that are not given to us in the HR simulation. To continue with your idea further, I feel that my thought process of the incidents have been completely wrong. I feel that I have over-analysed the problems. Instead of looking at the problem as what is the best way to fix this right now, I have been thinking what is the best decision for the company long term. After reflection, I see that there is no long term consequences or benefits of the incident. It is either you pick the best or the worst choice. There are no side effects or small obstacles that get in the way. I got discouraged when I thought about this. Now when I look at these incidents, all I think is, "What does the computer want," and not "What is best for my company and its future?"
Posted by: Robert Bucek at November 6, 2005 02:07 AM