Political bias

While listening to the radio, the sound engineer made an error and started to play another story over the top of the one that was already playing.  Only four or five words escaped before the mistake was resolved.  But four words stuck with me long enough to do a curious comparison.

President Obama

The words were, "Democrats, as a rule. . ."

I never did learn what Democrats do as a rule.  So I can't give an opinion as to whether I agree, disagree or both with the content of the story.  But the question that sparked in my mind was whether I would find more instances of "Democrats, as a rule" or "Republicans, as a rule".  So once I returned home I executed both searches using the omnipresent, if not omniscient, Google.

One of the searches yielded about 58,300,000 hits.
The other search located 43,500,000 hits.

President Bush

Before I reveal the result I want you to consider which group you believe had the higher return.

I was quite surprised by the disparity.  Someone with better statistical skills than me will have to clarify because I don't know if I take the 15 million difference and divide it into the total of both searches or if I divide it into the larger or smaller number.  However, I don't believe there is a question that there is statistically significant difference between those numbers.

Statistics get a bad wrap.  And here's why.

Statistics, alone, tell us nothing.

I started to wonder about some possible causes for the disparity. My first thought is whether one group posts on-line more than the other.   Is it possible that while explaining its views,  one group either has access to more computers and is more active in posting?  I also considered whether it is possible that it is the other group that is doing most of the posting, but rather than explain its own view it spends more internet space explaining the basic and general rules of the other party from its point of view.

A completely different explanation would require an understanding of the algorithms employed by Google to achieve those results.  Does Google somehow favor one political group over the other?

What was abundantly clear to me is that I don't have any idea regarding the answer to any of these questions and it would require some significant work and metadata study to be able reach some reasonable explanations for this divergence.  I certainly invite anyone else who may read this and is so predisposed to follow up and do that research.

But back to the question at hand.  Which group did you select?
Would either result actually surprise you?  If it turns out different than what you originally believe, what explanation would you provide?


Well here are the results with the "about" qualification that Google likes to give.

Remember, one of the searches found about 58,300,000 hits.  The other search located 43,500,000 hits.

Democrats yielded "about 58,300,000 hits."
Republicans show up with "about 43,500,000 hits."

So now what?