Statistical sophistry? What on earth do I mean by statistical sophistry, other than repeating it for improved SEO purposes?
Well, one of the first things we should all learn about statistics is that you can pretty much use them to illustrate any point you like. People use them all the time, because they add a layer of credibility to an argument or case. We’ve all heard the phrase that 48% of all statistics are made up on the spot – feel free to insert your own stat as you read this – but the dangerous thing about statistics is that they can be created, skewed and twisted to serve any purpose. You only have to ask the global political establishment.
Then there’s the sophistry. They used to bandy the term about in Ancient Greece to draw the differences between genuine philosophers and thinkers and the sophists who argued for the sake of things, using trickery, guile and superficial nonsense to dupe their audiences. I originally typed ‘dope’ there my mistake; maybe the typo is more accurate.
The key to interpreting all statistics is to look behind the numbers. What do they really mean? How were they arrived at? What was the sample size? How rigorous was the analysis? How objective was the work, or was it done to justify a preconceived view? Often you can’t answer all these questions, but it still pays to look behind the numbers and peer into the ‘why is information being presented to me in this way?’ abyss.
Just because you use a stat, doesn’t mean it’s true. People who use statistics responsibly and clearly are edifying and educating us. People who use them to distract or obfuscate are not. It’s up to us to keep our wits about us to distinguish the true philosophers from the sophists.
I think that a lot of people misuse/misinterpret statistics without knowing it. I made a post on my blog a while back stating that statistics is based on partial information (half truths and such).
If I take a sample of let’s say for example 200 of patients’ heights (in ft.) and the sample average of the 200 comes out to say 6ft. How this result is interpreted and reported may differ (due to author bias and subjectivity). But to expand upon this and say that the average height is 6ft throughout the world is misleading and (potentially) dangerous. The sample size really matters here.
LikeLike
Yes, agreed, both the number and the make-up of the sample is key. Thanks for commenting.
LikeLike