The Industry

2.6 Million Reasons to Keep Yelling About “Bias”

Why the president and the right have latched onto a flimsy study about Google manipulating votes.

Robert Epstein, the former editor in chief of Psychology Today, responds to Sen. Cruz during a hearing in July.
Epstein claimed that bias in Google’s search engine is “one of the most powerful forms of influence ever discovered in the behavioral sciences.” C-SPAN

On Monday, when President Donald Trump tweeted about a report showing that Google had “manipulated from 2.6 million to 16 million votes” in favor of Hillary Clinton, it was obvious what had put the idea in his head. That morning, Fox News’ sister network Fox Business had discussed the July Senate testimony of a psychologist named Robert Epstein, who said that “Google’s search algorithm likely impacted undecided voters in a way that gave at least 2.6 million votes to Hillary Clinton.” Though the clip of Epstein’s testimony was a month old, Trump’s tweets inspired conservative media outlets to jump on the story—and the rest of the media to point out just how bogus Epstein’s claims are.

Epstein is a senior research psychologist at the American Institute for Behavioral Research, but other than a technologist who doesn’t appear to help with the studies, Epstein is the organization’s only staff member, according to the AIBR website. Epstein is the co-author of a study published by his employer, only the summary of which is available, that claims that biases in Google’s search results may have tilted millions of undecided voters to vote for Clinton in the 2016 election. That was the study Epstein described in his Senate committee testimony, where he didn’t mention that his huge claim is based on monitoring the search results of just 21 undecided voters out of 95 voters for a 2017 white paper. In his submitted testimony, Epstein did provide seven pages of citations—but all of them are papers or op-eds he wrote or co-wrote himself. Only one of them—a 2015 paper in the Proceedings of the National Academy of Sciences, about how biased results produced by search engines could have the ability to sway undecided voters—was peer-reviewed. Even that study didn’t demonstrate that this has actually happened.

It’s easy to see why Trump—and conservatives like Sen. Ted Cruz, who was effusive during Epstein’s testimony—have seized on these claims. Trump has repeatedly criticized Google and the other Silicon Valley giants as being biased against conservatives, and he’s long made spurious claims that he won the popular vote in 2016. With both charges, no proof exists. But Epstein’s conclusions offer conservatives a way to say both of those things are true—and conveniently, they come from a Harvard-educated former editor in chief of Psychology Today who says he supported Hillary Clinton.

It doesn’t matter that Epstein’s claims are so easily dismantled. Cries over social media bias have been a right-wing refrain since an overbaked controversy over Facebook’s Trending News section demonstrated the company will bend over backward if enough conservatives get upset loudly enough. After all, Facebook, Google, Twitter, and their peers want their products to be used by everyone, and they certainly don’t want to tick off politicians of either party. What’s telling is that even when the major technology companies indulge conservative complaints, there’s still no proof that they’re systematically injecting political bias into their products. (What the platforms have actually started doing is take measures to ban figures who peddle in conspiracy theories and hate speech, and who incite harassment and violence.)

Take a Facebook-commissioned study that the company released on Tuesday. Faced with complaints that it was biased against conservatives, Facebook asked former Republican Sen. Jon Kyl to conduct an audit of the platform. The result focused only on complaints conservatives had against Facebook—and there are many—but didn’t even endeavor to do any kind of comprehensive study into what kind of information Facebook serves, promotes, or removes. In announcing the report, the company conceded that its moderation policies have been “applied unevenly in the past”—it didn’t say how—and announced one policy change that has nothing to do with the content users post to the platform. (It will now start allowing ads containing images of people with medical tubes, which will likely allow anti-abortion groups to show more graphic imagery.) But, once again, no actual evidence of systematic political bias in Facebook’s products emerged.

The Epstein meme typifies this dynamic. A pop psychologist and cable-news regular, he regularly publishes op-eds in conservative outlets like the Epoch Times and the Daily Caller, as well as the occasional piece in Politico or USA Today. He once appeared on The O’Reilly Factor to explain why he thinks teens should be treated as adults (in a book on that topic, he argued in favor of corporal punishment), and on a Fox News morning show in 2008, he argued in favor of teens getting married, pointing out that “Mary was 12 or 13 when she had Jesus, was she not?” More recently he’s focused on Google, Facebook, and why he believes they can throw elections. As he explained to Tucker Carlson this March, “Google can take a 50/50 split among undecided voters and change it to a 90/10 split with no one knowing they have been manipulated.” In his testimony to the Senate, Epstein called the effect of search engine manipulation “one of the most powerful forms of influence ever discovered in the behavioral sciences.”

Epstein may prefer Clinton to Trump, but the thing he seems to really dislike is Google. An article in the New York Times from early 2012 points to a tiff the psychologist had with the company after his website was hacked. Google directed visitors not to go to his page until the malicious code was removed—and kept the warning up even after Epstein tried, apparently unsuccessfully, to clean up his security and begged Google to remove the label. Epstein threatened to sue the company for not removing the warning, explaining to the Times that he felt like he was yelling at a brick wall. Later that year, he published a series of articles in the Huffington Post about why Google should be regulated. For the next few years, he began to publish more regularly about how easily Google could throw an election, largely citing himself. Starting in 2016, he become a regular on Breitbart discussing the Google topic.

Among academics, Epstein is a lonely voice on Google’s election-throwing powers. “When Dr. Epstein says the effects are ‘huge’ and ‘more powerful’ than anything he has ever seen, I respectfully suggest that he needs to read the political science literature before making that claim,” Katherine Haenschen, a communications professor at Virginia Tech University who studies internet targeting on voter turnout, told Mother Jones this week. “Large-scale digital mobilization has basically failed to deliver sizable effects in terms of persuasion or turnout.” Never mind the fact that in Epstein’s study, it’s not clear what search terms were used by his participants, or what the “biased” search results were. In his research, Epstein graded search engines for bias, determining that mainstream news outlets like the New York Times dominated over conservative sources like Breitbart in Google’s results. Epstein doesn’t explain the context in which the searches were conducted—which is important to know, since the whole point of Google Search is that it personalizes results based on prior searches and the user’s location. Someone with a recent search history about guns in Tennessee will likely see different search results than someone with a recent search history about women’s health care in New York City. And a good study would take care to somehow sanitize or disclose each participant’s search environment before reaching any conclusions.

Despite the severe limitations of Epstein’s research, Trump latched on, exaggerating the already unbelievable findings. Even Epstein didn’t claim that Google directly “manipulated” any votes, and his high number was 10.4 million votes, not 16 million.

Although Trump has reportedly recently weighed signing an executive order that would have federal agencies police how social media companies moderate user content, it’s hard to believe that some sort of real crackdown is in the works—or that if one was, it could survive a court challenge. It’s more likely that conservatives keep talking about bias on social platforms for two reasons. One, the topic really fires up their extremely online supporters—like the group of conspiracy theorists, Infowars regulars, and racist meme-makers that the White House hosted recently at a “social media summit.”

The second reason is social media really does work in conservatives’ favor, and they don’t want the companies to do anything that might change that. Month after month this year, the Trump campaign has outspent all the Democratic candidates combined on Facebook ads. Even after the company tweaked its news feed algorithm to deprioritize news and politics, conservatives across the board are thriving on Facebook. Fox News is one of the biggest publishers on Facebook, far outranking CNN in terms of both followers and engagement. Not coincidentally, all of this discussion of bias obscures many of the real problems with social media—like their ability to amplify emotional and divisive content and their vulnerability to misinformation.

Now, if tech companies do try to do more to protect the integrity of elections—say, by ferreting out pages that are boosting untruths about, for example, immigrants—they can expect to be accused of censoring conservatives. With or without Epstein’s claims of manipulation, that’s where we’d be. But this latest meme gives conservatives 2.6 million to 16 million reasons to yell “bias” even louder.