January 31, 2003
Fuzzy Math II
In the last two months, Bush's approval rating has declined from 66 - 62 percent (ABC News...); from 63 to 57 percent (Zogby...); from 64 to 60 percent (Gallup); from 60 to 55 percent (Newsweek...) ... well, you get the picture. Some polls show a steeper decline - NBC News' poll shows a slip from 62 to 54. But if you average it all out, the drop is probably around 3 - 4 percent.... [Krugman's] column headline? "Credibility Problems." Yep. He got that right.
But the average of -4, -6, -4, -5, and -8 isn't "3 - 4." It's 5.4.
Can't we have a rule that somebody be able to average five numbers before they are allowed to post on the internet? :-)
Posted by DeLong at January 31, 2003 09:29 AM
How can Andrew Sullivan get an average below each of the numbers he averages without noticing it???
That was a rhetorical question right, JP?
Do remind me, who wrote: Ours is not to reason why, Ours is to read and believe ?
In all fairness, one should point out that Sullivan deferred to a Wharton professor of statistics regarding the various ways poll numbers can be read when determining whether they constitute a "plunge" or not.
Also in fairness one might note -- as the Economist just did -- that Bush's approval numbers have plunged all the way to where Reagan's were when he carried 49 states.
Now that doesn't seem to be quite the impression that Professor K. was trying to give: "Bush's approval ratings have plunged to the level of historic landslide victories". So perhaps he didn't know that (or did but was being disingenuous, or was whistling in the dark about it).
But if mere internet posters should be tested on math literacy, maybe national political writers should be tested on political literacy (or aptitute towards disingenuousness, or whistling)?
Alas no humor for Jim Glass, when it comes to that evil evil man Paul Krugman.
On the other hand, Jim classifies Sullivan as a "mere internet poster" so perhaps he does have a sense of humor after all.
>>At the end of 2002, his numbers stopped declining, which gives the appearance of a plunge at the end of January, 2003, since those numbers are in line with the overall trend. <<
Hahahahaha. Is that what they teach 'em at Wharton these days? Note for the time being that a mountaineer could make the same proud boast after descending 10,000 feet from the summit of Everest then falling down a crevasse.
Can't we have a rule that somebody be able to average five numbers before they are allowed to post on the internet?
No. One of the great benefits of the internet is that anyone can post material with having it vetted first.
Of course, that includes criticism. If someone makes simple math mistakes, others can ridicule the errors. :-)