Last week I cheerfully invited all my readers to ignore me whenever all I had was correlation. This week, I’ll ask you kindly to always hear me out, even when all I have is correlation.
Yes, I’m kaleidoscopic that way. Still, all the same, I’ll ask you to ‘hear’ me out.
You see, statistical models are by necessity simplifications.
There is always more analytics possible. The question always is “What may we do, with the time we have, to create a better decision now, today, than we were able to make yesterday?”
Yes, you should always feel free to question a model. Especially one as simple as merely correlation(s). Nevertheless, it is only common politeness to have a better model yourself to offer in substitute. (It’s not a formal rule of analytics, but I rather wish it were).
I think for a long time in academia (at least our student success side), we’ve had a bit of a dodgy habit of simply using some very basic analytics. This understandably gets me overexcited when I see decision makers and colleagues not using the best we have on offer for assessment. It is one of the reasons (besides their adorableness) that I introduced the box-and-whisker plots. We can do and have better!
This is why, in some of my present and upcoming workshops in the real world, I’m going to invite my colleagues from all over campus to bring some of their data, and chat with me about what sort of dashboards or semi-regular insights you might wish to have and see.
Perhaps we can sort out a plan to get you your data and have a good model for it. These days, you’d be surprised how easy it is to graph and chart and explore data. In the past, the toughest part of analytics was actually doing the analytics. These days, I estimate 80% of my time is spent munging data. A fun word that generally means fiddling in Excel with messy data for hours or days that aren’t in the proper columns and rows format yet to be easily accessible to a program like SPSS, SAS, or R. My point is, if you have some system where you’d like deeper insight, it can be done! It might not even be as painful as you’d think, and I can definitely help you get your data collection process streamlined so it will go smoothly. I hope to see you in some of those workshops!