before I forget to tell you...
From an article in the NYTimes.
Recently, I have started to notice pseudo stats. Numbers that seem to imply something profound, but are missing the data that matters. Take the quote above. On first read, the numbers above seem to indicate that ~ 99% of all incumbents win (or at least, that is how I read it). But, that’s not what it says. It says in one percent of elections, an incumbent lost.
Here’s what it doesn’t say: Of all the elections since 1982, how many of those elections involved an incumbent seeking re-election?
Without changing the numbers above, both of the following statements could be true:
A. Incumbents win 99% of the time.
B. Incumbents lose 100% of the time.
If all 2,958 elections involved an incumbent seeking re-election, and the incumbent lost in 39 of those elections, then A is true.
If only 39 of the 2,958 elections involved an incumbent seeking re-election, and all 39 lost, then B is true.
Don’t get me wrong, I am sure the win percentage for incumbents is shockingly high, but the stat above doesn’t get us there. And it scares me that if I had been reading a bit quicker, I’d be trotting out “99% of all incumbents win” at dinner parties when the conversation veered toward the horror show that is Albany.
