Google users get different results when they search for the same terms, even when logged out of their Google account and in private browsing mode, according to a study by DuckDuckGo, a Google rival based in Paoli, Pa.
Google said the study was flawed.
DuckDuckGo said most of its 87 survey participants received unique lists of links when searching Google for "gun control," "immigration," and "vaccination." For example, 76 people across the United States saw 62 different sets of search results for "gun control," despite using Google's private browsing mode, called "Incognito." The users submitted the identical queries at the same time, and the survey didn't count different local websites as a variation in results, DuckDuckGo said.
One person who searched for "gun control" saw a link to the National Rifle Association at the top of the results page, with Wikipedia listed later. Another's top result was Wikipedia without any link to the NRA on the first results page. A third got the NRA link, but nothing from Wikipedia. DuckDuckGo did not provide information about the participants' political leanings.
Even if results generate the same list of links, the order they occur in can have a major influence on which one a user clicks, according to DuckDuckGo. It said a given link can get twice as many clicks as the link after it.
DuckDuckGo is a nationally-known privacy-centric internet search site that says it doesn't store or share user data. CEO Gabriel Weinberg is a frequent critic of internet privacy practices.
Sponsored Video Stories from LifeZette
The survey used a tiny sample size and was not scientific. DuckDuckGo solicited volunteers on Twitter and participants submitted screenshots of their search results. But Weinberg said the "intense variation" of search results suggests results may be personalized to a user or device. He said DuckDuckGo chose political topics to show the level of influence Google could have on voters.
"We don't ascribe any intention around influencing political topics. It's more a natural consequence of these algorithms," Weinberg said. "That said, it's a bit reckless."
Google said DuckDuckGo's suggestions that the search results were personalized is "simply not true."
"This study's methodology and conclusions are flawed since they are based on the assumption that any difference in search results are based on personalization," a Google spokesperson said in a statement. "In fact, there are a number of factors that can lead to slight differences, including time and location, which this study doesn't appear to have controlled for effectively."