“I can prove anything by statistics except the truth” – George Canning.
Do you remember your statistic classes at university? Hopefully yes, because statistics are a fantastic tool to describe the world we live in. This is because many things in our universe are way too complex to describe them in detail. Just think of the room you are sitting in right now. The temperature you feel is mainly due to the individual movement of an unimaginable amount of separate molecules. It is simply not possible to calculate the movement of all of them. Because of that physicists developed what we know today as statistical thermodynamics. They don’t look at all particles individually any longer, but instead look at a statistical ensemble of particles. This use of statistics makes things much, much easier and still is incredibly accurate.
But despite its usefulness in many parts of our life, most people don’t especially like statistics. Many regard statistics as a very abstract and complex mathematical topic beyond the understanding of any sane being. However we’ll do as always in this blog and investigate things, no matter what (no need to worry: no formulas involved). But before we’ll do this, we’ll ask a very important question: why is understanding statistics doing you any good in your daily life? Where is the benefit in it?
That may sound very difficult to answer but in fact the answer is extremely easy. Let me explain: As a matter of fact all of our brains are hardwired to ignore statistics. Instead everybody of us is firmly believing in cause and effect. That’s an evolutionary thing which allowed humans to survive as a species and therefore it is incredibly hard to avoid. Imagine you are living in the Stone Age and two people of your tribe suddenly die after trying out a new fancy looking berry. In this case it certainly is a good idea to think of cause and effect (the berries are toxic) and not to think of statistics (don’t worry, might have been just a coincidence). Evolution has it that all of us always look for the cause of something and really never think of statistics (or luck) alone. Knowing this allows you to avoid all the many mistakes others make due to their hardwired belief in cause and effect. You’ll hopefully see what I mean after you’ve read this entry…
Having said that, we’ll do as promised and have a look at statistics. For today we have only one topic. It is called regression to the mean and was discovered by Franics Galton who was experimenting with sweet peas in the 19th century. He observed that extremely large parent peas have offspring who are not bigger but smaller than their parents. And on the other hand very small parent peas usually have bigger offspring. The reason for this is that extremely large peas usually do not only have genes that promote size, but have also experienced very favorable conditions. They receive the perfect amount of sun, nutrients and water. They had just the right distance to the next plant and were not disturbed by diseases or pest. In other words they had the genes and were extremely lucky. Their offspring will also have good genes but it is unlikely that they will be as lucky as their parents. As a result they will still be large, but smaller than their parents. See Galtons original measurement below:
Diameter of Parent and Offspring Generation (in 1/100 inch)
| Parents | 15 | 16 | 17 | 18 | 19 | 20 | 21 |
|---|---|---|---|---|---|---|---|
| Offspring (Average diameter) | 15..4 | 15.7 | 16.0 | 16.3 | 16.6 | 17.0 | 17.3 |
This effect is called regression to the mean. Every time you see a really exceptional result you will immediately attribute this to some cause. However in reality exceptional results are usually achieved because of a good portion of luck and because of that the next results will be not nearly as good. Let’s look at a prominent example to illustrate this effect: the Sports Illustrated cover jinx. This urban legend states that individuals or teams appearing on the cover of the Sports Illustrated magazine will perform much worse afterwards. In fact there are quite a few examples demonstrating this effect. But it’s not black magic or a curse that is the reason for that. Instead it’s regression to the mean. An individual or a team needs to perform really extremely well to make it to the cover page. It is very likely that there was also a bit of luck involved. And it is also very likely that this luck won’t last forever. This means after they appeared on the cover page they perform again according to their skill. Which might be still well above average, but compared to the outstanding (and lucky) performance getting them to the cover page this looks like a downturn.
The Sport Illustrated cover jinx is one of those rare cases where you can actually learn something from the life of a sport superstar. Let’s look at an example probably closer to your daily life: imagine you recently changed your sales organization. The first month afterwards was mediocre but all of a sudden sales explode. I guess most of us would be convinced that the reorganization was a complete success and sales will continue to rise steeply. Unfortunately that’s in most cases simply wrong! The right answer is that this is probably only a statistical effect and has nothing to do with your reorganization. Indeed you’ll probably see much more average sales next month again. This will usually make you wonder about the cause again (There is no cause! It was only statistics!). This example applies to many situations you might encounter and it is also universal for all KPIs you measure from productivity to quality.
The message is always the same: beware of a simple cause and effect explanation if you see exceptional numbers. This is just your brain trying to trick you! In most cases there is absolutely no cause, what you see is only due to good or bad luck. Just try to accept this and do not put too much weight or meaning in such results.
Before we finish for today I would like to make you aware of one more thing where the regression to the mean effect usually plays a huge role: your incentive system. Incentives should encourage and motivate people. However they always also have a huge potential to create frustration. Especially if you are awarding mainly the top performers who had a really exceptional performance record. At first sight this might sound reasonable, but as we just learned: exceptional performance is often due to luck and not (only) due to exceptional skill or attitude. Even worse it is very likely that the results of somebody with an exceptional performance in one evaluation period will be not nearly as good the next time. And this certainly creates frustration!
Does this sound familiar? Let us know what you think!
Take Care!
Business Psychology 