Google has previously been accused of "purging Conservative news sites." There are other accusations that Google has manipulated search to harm Rick Santorum (more here). Al Gore has been a Senior Advisor to Google. Well, I decided to do a test by searching on my own name. Here is what I get on my name for Google and Bing for "best match" news search.
The first screen shot shows what happens with Bing and the second with Google. Notice how under Google the lead searches are for a couple Media Matters pieces and other attacks on me. But for Bing the first search findings are for Fox News; for John Lott, the sports writer who writes about the Toronto Blue Jays; and then outdoor life.com and the Daily Caller and Breitbart. The Media Matters links come in at number 9. Click on the screen shots to make them bigger.
I then tried doing the search using the "most recent" option for both searches. With Bing, You get a lot stories from Outdoorlife and Shooting Sports News to Fox News and the Daily Caller. For Google, two of the first three hits are attacks on me. Three of the first six are attacks. And Media Matters shows up in two of the first six hits.
So why does Google put so much more weight on pieces in Media Matters? Interestingly, there were other "Opposing Views" columns that were favorable to me, but Google only seems to pick up the critical columns.
This bias in Google searches is something that I have seen over the years, but it isn't just important to me, it is Google's way of impacting the political debate and giving it a decidedly leftward tilt.
One final point, Bing seems to find a lot of stories that Google doesn't find.
A computer science professor that I know who specializes in these search questions wrote me:
this is the kind of thing we have seen for years, and seems to be a classic example in point. Google's secret sauce is flavored with larger doses of some sites than others, which is to say, their computation of page rank places more credibility on sites which emphasized one rather than the other set of sources. There are lots of benign reasons why this could be true in any one point example (and I can explain them in detail if you like), but to see this as a pattern over time tells us something about their algorithms. . . .