Some will read that last line and nod, others may be offended and some might claim something about the inadequate training of a few shouldn't cast a shadow on all.
So true, and experience informs me that that "all" group is quite large.
So let me offer a short refresher borrowed from physics.
A Beginner Mistake Everyone Makes
A common mistake freshman physics students make is called "confusion of units", and the truth is I've known university physics professors make this mistake.
Confusion of units goes like this: 2 apples + 2 oranges = ?
Mixing apples and oranges might lead to an interesting beverage but the apples and oranges don't mix, they don't blend, they aren't the same thing, you can't measure one by the other and therefore whatever conclusion you draw is irrelevant.
You've started with perfectly valid data - apples and oranges - and come up with a completely meaningless result for the purposes of analysis because the valid data isn't useful to the problem being solved.
Most data analysts, when told "Your data is valid and your conclusion is irrelevant" begin churning through all their known data in order to make their conclusion relevant. It's amazing.
They'll start mixing apples with bananas, oranges with cherries, et cetera et cetera, and when they've run through known fruit they'll go onto vegetables - apples with carrots, maybe? How about an orange with some lettuce? - then onto various meats, dairy products, pastas, anything and everything to make their conclusion relevant.
Is your stomach churning at the thought? Good. Now stop mixing things together that don't belong together. You'll hurt yourself if you don't.
Good Data Analysis Starts with the Conclusion
Take a moment and look at the conclusion you want to draw. Forget about the data right now, focus on the conclusion.
Now ask yourself, "What data is necessary to support that conclusion?" This leads to "What types of data support that conclusion?" and this leads to "Do I have any data of that type? onto "Do I have any useful data of that type?" to "Is that useful data valid?" and on and on and on.
Eventually you've got a trail from your conclusion to your data and are able to make such a strong case for what you're offering that only executives would ignore it.
Okay, I'm jesting about that executives part.
Anyway, do this often enough and you'll start laughing at the conclusions you're coming up with. "Where did I get that?", you'll ask yourself, and you'll catch yourself long before your boss or client stares at you asking, "How did you come up with that again?", and wonders if you have apples or oranges for brains.
You can follow me and my research on Twitter. I don't twit often but when I do, it's with gusto!
Have you read my latest book, Reading Virtual Minds Volume I: Science and History? It's a whoppin' good read.
Learn the latest regarding NextStage blog posts, conference sightings, whitepapers, tools, presentations and more via The NextStage RSS feed .
And if your tastes are more to fiction than science, give Tales Told 'Round Celestial Campfires a read. It's great fun and sure to make you wonder and think.