By Jon Dougherty
A new study by a researcher who, for months, has been looking into how social media platforms can influence election outcomes, found that search biases introduced by Google into its algorithms last year likely cost at least three Republicans their congressional seats.
The study was conducted by Dr. Robert Epstein, a San Diego-based Harvard Ph.D. who founded the Cambridge Center for Behavioral Studies. Epstein is the Senior Research Psychologist at the American Institute for Behavioral Research and Technology (AIBRT), a University of California-San Diego visiting scholar, and served as editor-in-chief of Psychology Today.
In addition, he supported Hillary Clinton (like Google did) in 2016.
The scientist and AIBRT analyzed Google searches that were linked to a trio of highly competitive congressional races in southern California that were won by Democrats. In doing so, the team discovered that Google’s “clear democrat bias” likely flipped those seats away from GOP candidates.
The study noted that at least 35,455 undecided voters within those three California districts were probably persuaded to vote for the Democratic candidate because of Google’s biased search results.
Epstein says that in the days leading up to the 2018 midterms, he was able to preserve “more than 47,000 election-related searches on Google, Bing, and Yahoo, along with the nearly 400,000 web pages to which the search results linked.”
Analysis of this data showed a clear pro-Democrat bias in election-related Google search results as compared to competing search engines. Users performing Google searches related to the three congressional races the study focused on were significantly more likely to see pro-Democrat stories and links at the top of their results.
As Epstein’s previous studies have shown, this can have a huge impact on the decisions of undecided voters, who often assume that their search results are unbiased. Epstein has called this the Search Engine Manipulation Effect (SEME).
Apparently tone-deaf to the fact that a Google search reveals additional evidence indicating bias, Democrats and officials with the tech behemoth are disputing Epstein’s findings.
Besides, this isn’t Epstein’s first foray into this type of research. Studies he has conducted in the past have revealed that Google has demonstrated it is capable of changing a 50-50 split among undecided voters into a 90-10 split without anyone realizing they had been manipulated, and without making it apparent to outside observers, based on a set of search suggestions that can be implanted into algorithms.
And Epstein made another startling revelation in an interview with Fox News’ Tucker Carlson earlier this year: A single shift in Google search results on Election Day 2018 very likely shifted anywhere from 800,000 to 4.6 million votes to Democratic candidates.
He noted that “individual employees at Google have the power to manually remove material, remove content that we see, and they do so often with political ends.”
He also made a prediction during that interview.
“I guarantee you that this past election was affected,” Epstein said, noting then that he had research data he would be releasing — this month — substantiating the claim. He said Google’s posted reminder to “go vote” on Election Day in November shifted perhaps millions of votes to Democrats.
Just this week Carlson warned that POTUS Trump is in danger of having his 2020 reelection stolen from him because of what Google and other tech behemoths are doing to influence outcomes for Democrats.
- Follow Jon Dougherty on Twitter at @JonDougherty10