Nice one Jim. The Physics is Phun blog recently identified the silliness of the Washington Post with regards to listing voting results sorted by – yes, can you believe it – astrological sign.
What does this lead to? Undoubtedly… campaign ads:
Coming to your TV soon:
We will return to Transformers: The Next Generation after these messages.
***Fade to Black***
***Image of a cheerful politician appears.***
A deep voice begins: “Senator Blowhard claims to be for social security reform, lower taxes and cute, cuddly puppy dogs…”
***Image changes to Senator in a menacing pose***
“… but he’s really a Pisces and Pisces vote for nuclear waste and cat hunting.
Would you take his word for it?”
Perhaps this ‘service’ by the Washington Post is being provided to allow us to analyze how voting preference is completely NOT functionally related to astrology. It might be fun (in a data-geek sort-of way) to see if particular signs favor membership in Congress, or in a particular party.
SO since I’m pretty much a data-geek, I decided to dive into the pseudo-science:
I visited the Washington Post database and extracted the astronomical signs of the House members of the last 9 Congresses. There was no immediate way to do this, so I simply selected the vote with the widest margin for that House and sorted by sign. Since there are certainly incumbents who appear in multiple Congresses, the data is not really independent from year to year. Any bias that results, however, would tend to emphasize those sings that not only “predict” elect-ability, but perhaps the savvy to remain in office. Here are the results:
Graphical Results
Well, bad luck if you’re a Pisces I guess. Political office is just not for you. Cancer on the other hand pops about 30% higher than average with Pisces about the same amount low for the last 9 years. I didn’t do any Chi-Squared or anything to see if this is normal statistical variation, but the standard deviation seems to be about right for a uniformly random discrete variable with 12 possible values.
I was hoping to look at sign vs party, but the way the data is organized this would require some more advanced data mining… and I’m just not that interested.
