WelderDestiny › E-Zine Back Issues › Issue #072
Wednesday, June 13, 2018 / Perth Australia / By Niekie Jooste
In this edition of "The WelderDestiny Compass":
When I was doing my engineering degree, statistics was one of those courses that I needed to complete to obtain my degree, but it was such a dry and boring subject that I despised it. Needless to say, I did not learn much from that course, other than how to pass the exam.
Later, when I had started working in the field of quality assurance, suddenly statistics took on another character. So much so that I completed a diploma in statistics. Even that diploma was not necessarily excitement from the start to finish, but at least I understood why it was important to grasp the concepts.
In the mean time I can assure you that I have forgotten a lot of the details of the statistics that I learned, but I have come to appreciate how to use statistical thinking to grasp the bigger picture of sometimes complex situations or systems. Statistical thinking allows us to sort the wheat from the chaff, and helps us to focus our attentions on areas that are important, while doing our best to not be distracted.
In the machine age, I believe the skill of statistical thinking will become more and more important as the world awash with information and data tries to distract our attention.
This is not the first time that we have discussed statistical thinking here in The WelderDestiny Compass, but I would like to make it a little more personal today. Now don't worry, we will not be going through equations and things. If you are interested, you can easily find a whole lot of information on-line that can teach you the mathematics part. No, today I want to try to just get you interested. I want you to appreciate the value of statistical thinking from the perspective of an entrepreneur.
If you would like to add your ideas to this week’s discussion, then please send me an e-mail with your ideas, (Send your e-mails to: firstname.lastname@example.org) or complete the comment form on the page below.
Now let's get stuck into this week’s topics...
We live in a world with a whole lot of variation.
Some people are shorter, and some taller. Some heavier and some lighter. Some smarter and some less so. Some richer and some poorer.
Some apples are sweeter than others, and some cars last longer than others. Some countries appear to have a stronger economy than others, while the health of people in some countries appear to be better than in other countries.
I am not rambling on about different places you can find variations for nothing. I am trying to communicate the concept of a "population" within statistics. A population is merely the bigger data set from which we are extracting our information to do comparisons on.
If we are comparing the typical salaries of people, we might decide to do so for all working aged people in a specific country like Australia, or we might decide to compare everyone in the entire world, or we may decide to compare only men in the 30 to 40 year old age group within a certain suburb of a city like New York.
Once we have done our measurements and found all kinds of variations within the data, what does this variation tell us? Actually, it is very difficult to decide what it tells us without finding a way to decide if it is "common cause" variation or "special cause" variation.
Common cause variation is variation that is inherent in the population we are measuring. In itself this tells us little more than that there is a variation, and it is of a certain amount. This is a variation that we cannot ascribe to anything in particular.
Special cause variation is a difference in certain of the members of the population due to them having something "special" that makes them "outliers" in the population.
So, when we hear a report in the newspaper about how shocking it is that half of all schools are below average, then we can ask ourselves whether this variation is common cause or special cause variation?
Given that we know that an average is established by adding all the measures and dividing be the number of measurements, then it is obvious that regardless of how good a schools system is, roughly half of schools will be below average, and half will be above average. In other words, this statistic is actually telling us nothing, other than that there is a "common cause" variation.
To establish which schools are true outliers, we need to do a lot more analysis of the data. After that analysis, we may conclude that there is no identifiable special cause. That means that if we want to improve the schools system, then we should not be trying to "fiddle" with individual schools. Rather we should be looking at the system as a whole and decide how to improve the whole system.
And even if we manage to improve the school system as a whole, there will still be around 50% of schools below average!
In our schools example, if we started fiddling with individual schools, rather than working on the whole system, we run the risk of introducing greater variability in the system that will make things worse rather than better.
As an entrepreneur, if you can understand if variation (a problem) is due to common cause or special cause reasons, you will have a much better probability of using that information to your own benefit.
While it may be a good idea for you, in your particular industry, to make a deeper study of how to do the mathematics to identify common or special cause variation, for most of us it is good enough to just understand the concept. In our schools example above, you did not need to do any statistical calculations to see that the sensational headlines were meaningless in themselves.
In our aforementioned schools example, the statistical thinking was focused on the origin of the data. In fact, most of the so called statistics that we see in the media revolves around the analysis of measured data. In other words, the statistics is focused on the origin of the data.
Another very important focus of statistical thinking is the consequence of actions.
If you are going to implement some kind of system or action to try to influence people, ask yourself what statistics are at play, and how this affects the way people will react to the action you are taking.
A typical system may be a pay for performance system. The typical key result areas (KRA) performance management systems. We may decide to rank people in 5 different bands. The middle band gets the average salary increase while the top two bands get above average increases and the bottom two bands get below average increases.
So, ask yourself. If this system works properly, what percentage of people will be in each of the bands? What would the consequence be in terms of employee performance?
Most people will in fact be in the middle band, because by definition, if you have a "normal distribution", most people are around the average value. Ever decreasing numbers will be further away from the middle band. This means that probably around 75% of your employees will be in the average to below average groups.
Seeing as nobody wants to be told that they are average, never mind below average, you have just designed a system that will demotivate 75% of your staff. This will decrease performance, rather than increase performance. And, this is regardless of how good they are at their jobs. Potentially your staff are so skilled that they are within the top 10% of people in their chosen fields, but still 75% will be judged average or below average, based on the population being measured.
Just not a good system for improving performance is it?
Now, if you had a means to identify which of the low performers were "true outliers", then you could be in a stronger position. In other words, if you could identify "special cause" variation, then you could focus on those people to try to see why they are low performers, and fix their problems.
If you cannot identify special cause variation, in the performance of the employees, then your only logical approach is to find a way to make everyone better. In other words, work on the system rather than fiddling with the individuals.
In this example, we see again that we do not need to be mathematical geniuses to harness the power of statistical thinking. We just need to stop and think logically. Not to mention, have a basic appreciation for human nature.
"There are three kinds of lies: lies, damned lies, and statistics." - British Prime Minister Benjamin Disraeli
We live in a world where it is difficult to know who to trust. Unfortunately, statistics are like political prisoners. With some torture, they will say anything you want them to say, regardless of how true or false it may be.
Having the ability to look at information through a lens of statistical thinking will help us to identify false information, or at the very least alert us to a situation where the information is not necessarily trustworthy. We may not know if it is true or not, but we will know to approach it with a healthy skepticism.
We live in a world awash in information. A lot of it is meaningless, because we are not able to know if the information was tampered with or not. A lot of it is plain misleading because so-called experts have distorted the information beyond recognition, or they made errors in the analysis of the data.
This barrage of information is only going to get worse. The only way of having even half a chance of making well informed decisions in such a world, will be to develop a statistical thinking mindset. In the future we will probably explore this topic further.
Yours in welding
WelderDestiny › E-Zine Back Issues › Issue #072
Do You Have Thoughts About This Week's E-Zine?
Now is your opportunity to contribute to the topics in this week's The WelderDestiny Compass. If you have thoughts or examples that you would like to share with other readers of the e-zine, then please contribute by entering the title of your contribution in the box below. Feel free to make a brief or more expansive contribution to our discussion...
Do you have an example of where statistics was reported in an incorrect or silly way? Have you got an example of where statistical thinking helped you make a good call? Please share your stories, opinions and insights regarding today's topic.