On Sunday, I published a humorous correlation about improving economic data and increased Birther stories. It was a fun little story for the weekend.
A few people pointed out some issues right away. The “Birth Control” articles of late likely generated lots of hits on the word “Birth.” Contraception, not Birther related; So too did the many articles on President Obama releasing his birth certificate last year.
I ran these by Floyd Norris, who re-crunched the numbers — and once correcting for those, the correlation went away.
Floyd writes:
“I made a simple – but critical – mistake on my birther analysis. I failed to exclude articles on birth control, which of course are plentiful over the past few weeks.
Attached is a revised spreadsheet over the past several months. (birther). In the early months, there were virtually no birth control articles in the count, but they dominate it in 2012.
This, the thesis in the original message has not been proven.
I should have known better. The first Nexis search I ever saw, back in the 1980’s at Barron’s, was by a colleague who was preparing an article arguing the then new RICO law would be used against businesses more than against racketeers. To see what others had written, he asked Nexis for articles about RICO, and got a zillion – a problem since in those days Nexis charged by the article it found, not by the search. His next search was for RICO without Puerto.
(Thanks for Barry Ritholtz for calling this to my attention.)
Thanks goes to TBP readers and Twitter followers who brought it to my attention.
What's been said:
Discussions found on the web: