beware of averages

If we added up all the men, and all the women on our planet, we’d find that, on average, the typical adult human being had exactly one breast and one testicle. Yet how many people actually fit that description?

Statistics are used the create meaning. Yet, very often they create the opposite. In a world for of numbers, statistics and analytics (yes, even the Google kind) we are better off deciding what we want to find out, than we are looking at the available statistics and asking what they mean.

Startup Blog says: First decide what you want to find out, then devise a way to measure it.

1. It’s not really your point, but average without standard deviation is often meaningless.

2. Even with a standard deviation averages are meaningless, especially if the standard deviation is significant. Then it means that the average is totally pointless. As are many statistical measures.

“There are three kinds of lies: lies, damned lies, and statistics.” – Charles Wentworth Dilke

Steve.

3. Here’s an example: a running back has a run of 60 yards, 2 yards, 6 yards and 4 yards. Four carries for 72 yards is 18 yards per carry. Which is completely misleading.

So what’s the answer? Do you drill down into certain segments, or timeframes? Obviously, you have to at some point get a simple, high-level view of what’s going on. If it’s not averages, what is it?

@bdunc1