I Think vs. I Know

Data is everywhere in business. It makes no difference whether your company makes widgets or sells a service. You would think that it should be obvious, then, that using data to make business decisions is the right thing to do. It turns out, even with the best intentions at heart, our brains are still programmed to make decisions based on illogical bias. Have you ever been sitting in a meeting, looking at a set of numbers and heard somebody say something like, “What’s wrong with Team A? Their numbers are way down this week.”? I’m going to just assume you have. Did the speaker analyze the data visually or statistically before they commented? Here’s an example for you. The following number sets are production numbers from two teams at a widget production factory over the course of a week:

Team A: 5, 5, 5, 6, 4, 5, 7, 6, 5, 5, 5, 5, 6, 7, 6, 6, 5, 6, 6, 6, 7, 7, 5, 5, 6, 6, 6, 6, 5
Team B: 4, 5, 4, 4, 4, 5, 5, 9, 4, 4, 5, 5, 9, 8, 8, 3, 9, 9, 4, 3, 4, 5, 8, 4, 3, 3, 9, 3, 3

Which team had the higher average production rate? Is there enough evidence to say there is even a valid statistical difference between the two groups? ..

..Did you run the numbers or did you eyeball them and start making inferences? I’d be willing to bet that even though I started this post by implying inherent danger in assuming when it comes to data, your brain probably took over and made some kind of guess as to which was the top group. The interesting thing about that is that it’s natural. Perhaps even more interesting is that I bet, despite my best efforts to bait you into thinking Team B came out on top, some readers will probably still “guess” that A is the top team. My belief is that your guess has a lot to do with what you’ve learned is most valuable. People who guessed Team A most likely consciously or subconsciously value consistency over gaudy numbers. On the flipside, Team B choosers were drawn to the numerous 9’s and 8’s in that data set.

The truth here is that we can say, with a lot of statistical confidence, that Team A’s production rate was just under half a widget higher per interval than Team B’s. The extreme high intervals were outweighed by the extreme lows. The point that I’ve hopefully made is that this would never be obvious unless you take the time to calculate the averages and test the significance of the results.

I believe as HR evolves as a business function, it is going to be increasingly vital that we become the mathematical conscience of our leaders. I realize this is a departure from the conventional thinking but if we’re already being relied upon to be a “Business Partner”, who better to save our managers from their natural bias than the person who is already trusted as the organizational conscience?
This will of course require us as a group to be better prepared with basic math and statistics skills but everything you need to know you can probably learn from an online course from your local community college. Sure there is some value in becoming enough of a stats master to know when it would be appropriate to use a rank correlation coefficient test (I fairly certain I do not), but if you can test a simple hypothesis and understand how to properly use multiple regression, you’ve already made yourself more valuable to your organization.

I got really into statistics during an internship early in my career and haven’t stopped running numbers since. I went on to go through a five week Six Sigma course (basically a method of statistical process control through project planning) and there’s a basic tenet of the philosophy that I’ll never forget: “If you can’t measure it, you can’t understand it; if you can’t understand it, you can’t control it.” The further I go and the more I learn, the more obvious this idea becomes. I struggle on a daily basis to understand the data at my current place of business so that, hopefully, I can help my managers control their results. Our industry, like most, values reliability and continuous improvement. Neither is possible without first understanding your results.

I’ll close my statistics diatribe by mentioning that if you’re at all interested in not-so-obvious links between baseball and business, and you haven’t already read it, pick up Moneyball. The book is a treatise on the danger of statistics misuse. Michael Lewis is a talented writer and understands a lot more about microeconomics than I think he lets on. Moneyball will make any baseball or statistics nerd’s jaw drop when they see that baseball executives have been making multi-million dollar business decisions for over 100 years based on completely bogus statistics. For me it was a lesson in how dangerously easy it can be to convince yourself that you completely understand a thing when a lot of people who don’t understand math very well reach a consensus. I leave you with a quote from Bill James, the original “sabrematrician” that may very well be known someday for changing the way Major League Baseball is run:

“I do not start with the numbers any more than a mechanic starts with a monkey wrench. I start with the game, with the things that I see there and the things that people say there. And I ask: Is it true? Can you validate it? Can you measure it? How does it fit with the rest of the machinery?..Why doesn’t anybody say, in the face of this contention or that one, ‘Prove it’?”<sup>1

1</sup> James, Bill, and Bill James. The New Bill James Historical Baseball Abstract. New York: Free, 2001. Print.