Yup, praise. By now you probably thought I was incapable of that aside from a little crumb dropped here or there in the course of tilting at windmills. Tilting is what we do here at AHC and–there’s really no way around this–critical thinking involves criticism. It’s not critical thinking when everyone agrees. There’s a word for that practice in the English language: cheerleading. But let me get back to praise before I get caught up with another windmill.
In scouring sources for the definitive scoop on Pearson’s r I stumbled across the CoolData blog. Right at the start of the post on the subject the writer says “You can’t have a serious blog post related to statistics without tossing in the name of a dead white guy.” I knew right away this was going to be fun.
The truth is, even though I apply math I always have this sneaking suspicion that I’ve done something wrong. Partly I attribute that to spending a couple of years in a pre-engineering program where I was surrounded by living Bowmar brains with TI-XXs attached to their hips. And partly I attribute that to the way my brain is balanced. Of the 7 to 9 intelligences, my dominant one (surprise) is verbal so everything needs to be made verbal–even math.
There’s no better way for me to grasp a mathematical concept than to learn the story behind it. Math profs and teachers are usually negligent in this department so it’s a seek and find process. And what you find when you seek is, indeed, a lot of dead white guys with (especially among Edwardian Englishmen) belief systems that make modern folk like us shudder. (You also find a lot of dissolute French gamblers, which makes it an interesting trip.) So it’s refreshing to find anyone else who knows some of the back-stories of these giants and it’s always good to remember that heroes are human and have flaws.
This blog is a great reference. In the post on r the writer talks about ordering his regression inputs based on ranking the r-values. I’m as lazy as the next man and SPSS just makes it too easy to throw all your variables in so it can sort out the best order for you. But this is an easy step to add especially if, like me, you go to the trouble of running bivariates before you move into the regression.
Somewhere I saw promotional copy (Microsoft?) about breaking down the high priesthood of data analysis and methods for everyman. That may or may not be a good idea. Reading a practical, thoughtful, informed writer on how to actually apply statistics to business problems, though, is a great idea. If you need a place to start, start here.