I woke up in a cold sweat last night and realized we were in the midst of a perfect storm: the corona virus, political campaigns, baseball. And the problem: numbers. Or more precisely, how statistics are used in newspapers, television and the Internet.
As we’re looking at the spread of the corona virus, which is more important: absolute numbers, or percentages? Do we focus on current raw percentages, or percentages of increase and decrease?
Unfortunately, I also had a nightmare about news report, with graphics, describing the connections between numbers of virus cases as they relate to spouses of farmers living in odd-numbered zip codes.
The political campaigns are a God-send for pollsters and reporting poll results.
But I have to ask, what is the connection between a poll result today and the actual election? If the polls are correct, and are accurately reflecting what people believe, then do we need to have an election at all? Why not just take a sample at 10:30 tomorrow morning and let that be the election?
And if the polls do not accurately reflect what people think and do, then why bother with them? Oh, I know the answer: because politicians like to use them to show how well they are doing. Even if they are behind in the polls, well, that’s a good thing because politicians and pundits say it shows there is going to be a ground-swell of support in the next couple of days.
Then there is baseball, probably the most statistically-based human endeavor on the planet. There is the OPS: On-base plus slugging. The DICE – Defense-Independent Component ERA. And the ever-popular pNERD and tNERD, which are actual statistics used to measure the esthetic value of pitchers and the entire team.
I guess it won’t take long before someone tries to compare NERD values across World Series winners.
But, you have to say this about statistics: they are a great form of employment for statisticians.
I’m Larry Burriss.