Dubious Statistics Watch: Thanksgiving Retail Sales

Bloomberg News seems to be running two stories – on the same subject,

and by the same authors – at the same time. The first has quite an apocalyptic

headline: "U.S.

Consumers Spent Average of 3.5% Less on Shopping". Which is scary since

we were given to understand that sales would rise by 4% or so over the Thanksgiving

weekend. But then comes the second story: "Holiday

Sales Increase; U.S. Shoppers Spend Less on Average" – which

is just plain weird. Apparently total sales on Black Friday were up by 8.3%,

and the 3.5% decline was in sales per person, which wasn’t at all clear

in the first story.

But what no one quite comes out and says is that the number of people shopping

on Black Friday rose by 12.1% this year over last year – which is what

would be necessary in order for those two numbers both to be true. The reason

they don’t come out and say it is that the second story quotes the National

Retail Federation as saying that the number of people shopping rose only 4.8%

this year.

Clearly, these three numbers are inconsistent with each other: you can’t have

total sales up 8.3%, sales per person down 3.5%, and total shoppers up 4.8%.

It’s mathematically impossible. But it’s bad form to point out that none of

these numbers are particularly reliable, and that at least one of them has to

be wrong. Instead, you simply report what you’re told, even if it makes your

story read like Lewis Carroll.

This entry was posted in statistics. Bookmark the permalink.