Yesterday, I wrote about Eli Pariser’s intriguing new book, which argues that social media like Facebook and search engines like Google are creating filter bubbles. In turn, he maintains these filter bubbles result in increased political polarization among Americans.
Pariser has argued that Google uses 57 signals to personalize someone’s search results.
Today, I tried a brief experiment to test Pariser’s theory.
I asked my liberal mother-in-law in Massachusetts to Google the word, “war.” I did the same.
Below are my results:
Here are my baby-boomer, liberal, mother-in-law’s results:
At first glance, both searches seem nearly identical. There certainly does not appear to be any political bias. After all, the White Aryan Resistance showed up in the fourth spot for both our search results. This website is hardly one a liberal would embrace, and certainly not one I would either.
However, upon closer inspection, the ordering of our top three results differs. My first result is the standard Wikipedia entry on War. My mother-in-law’s is for the band War.
Why is this the case? Well, the band War was formed in 1969 and is popular with the baby boomer generation. I never heard of them until I did this little experiment.
The bottom line is that there are minor differences between my search experience, and my mother-in-law’s. However, the difference was presumably based more on age than on political affiliation.
However, I believe that my results would have reflected more differences if I were living in Texas instead of the very liberal San Francisco Bay Area.
If you try the same experiment and get different results, please email a screenshot to shazlett (at) mba2006 (dot) hbs (dot) edu and I will include it in a second post with more samples.