Bob Pockrass of Sporting News presented the results of a reader poll of NASCAR announcers. By way of disclaimers, this wasn’t a scientific survey – it was a sample of people who feel strongly enough about NASCAR announcers to answer an on-line poll. My analysis is of that poll data and doesn’t reflect my opinions about announcers and announcing.
The results were interesting, but I was having trouble getting my (admittedly warped) brain around the results. They were presented in a bar chart with the number of ‘best liked’ and ‘least like’ votes. I’ve reproduced it here, but make sure you look at the original article so they get their clicks!
The order is ‘most liked’ on the left, with Joy, Bestwick, McReynolds and Darrell Waltrip at the top. But boy, there’s no correlation with the bars on the right, is there? Ole’ DW also has the most ‘disliked’ votes of anyone. I started thinking about how you might present this data in a different format to make it clearer.
Don’t let anyone tell you the most important things you learn in math and science classes are definitions and algebra. It’s reading and interpreting data, charts and graphs.
Please don’t think this is a knock of any type against Bob – the man was covering races and writing multiple articles per day while I was sitting around watching movers load the truck. I’m a data geek. I can’t help myself around data. Surely there’s a twelve step program somewhere for people like me, right?
I realizes this probably needed a two-dimensional plot, with the horizontal axis being “liked” and the vertical axis being “disliked”. It took me a couple tries. I flipped the “liked” axis so that small numbers of likes were in the same quadrant as large numbers of dislikes, since those two ideas are similar.
I also set the axes so that they crossed at the mean value of each variable. Above the horizontal line, the number of dislikes was more than the average of the entire poll. To the right of the vertical axis, the number of likes was more than the average of the entire poll.
A couple things became apparent immediately from re-plotting the data.
First, DW is in a class of his own. Yep, you already knew that, but here’s the proof. You either love him or you hate him in a way that no other announcer shares. I labeled that quadrant “Polarizing” because people have really, really strong feelings about him both ways. The dislikes, though, did outnumber the likes.
The universally ‘Liked’ announcers are in the lower left-hand quadrant – they have few ‘dislike’ votes and lots of ‘like’ votes. Joy, Bestwick, McReynolds and Jarrett fall into this quadrant, with Joy leading the pack having most ‘liked’ and least ‘disliked’ votes.
The ‘Disliked’ announcers are in the upper right-hand quadrant – they have predominantly ‘dislike’ votes and few ‘like’ votes.
Then we have the lower right-hand corner, which I tentatively labeled ‘Agnostic’, meaning that people didn’t feel very strongly about the announcers either way.
So there is it, for what it’s worth. Another way of looking at numbers.
And to think I used to hate USA Today for using bubble charts!