Who knows if what Seth Stephens-Davidowicz told me was the truth? The data scientist with a PhD from Harvard just published a book called “Everybody Lies,” so that should have gotten my guard up.
But the book’s full title offers a little more shading. “Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us About Who We Really Are” suggests there’s some bona fide research here, not just accusation.
Stephens-Davidowicz, whose research is regularly found in The New York Times, worked as a data scientist at Google analyzing the search terms that you, me and everyone we know put into the great group brain. The results are somewhat horrifying. Facts don’t lie, and what they tell us is that everyone’s sex life is a bit of a disaster and there are racists everywhere. Okay, maybe not you, but more of your colleagues than you’d ever think.
The Times of Israel spoke with Stephens-Davidowicz about the unbelievable truths his work has uncovered, and what it means for our society at this particularly fraught period.
Do people get defensive about the title?
I think everybody lies. If you didn’t you’d be a horrible person. If someone says “how do I look?” or “is this meal delicious?” you frequently say “good” or “yes” even if you don’t mean it. But that’s not what this book is about, it’s about darker lies.
You refer to Google as “digital truth serum,” and before you wrote this you worked at Google as a data analyst.
I actually majored in philosophy, so pretty far from data science. I got my PhD in economics, which is a little more connected, but I was bored with the traditional economic questions. I was more interested in sociological and psychological questions, and how data can answer them. So much new and important information is available thanks to the Internet.
I used to joke with friends about how “Google knows so much about us.” Then I started playing around with Google Trends and right away I said “I want to study this for the next 10 years.”
From working in economics for so long I recognized how traditional information can have holes in it. There’s a lot of rounding and assumptions. Google’s search data seems a lot better than the existing data sources, and right away I wanted to focus on topics that people would be uncomfortable being open and talking about.
Not everything is dark, or even the majority. We ask for the weather or lyrics to a song. Google search, though, is biased to things we don’t feel comfortable talking openly about.
There’s definitely something a little darker there than what you think of with Google as a company. You think of the homepage, the Google Doodle, the nice, bright colors. What people actually search for can be a stark contrast.
You talk a lot about the racist searches. An example you give: Obama gave a remarkable speech about the dangers of Islamophobia and the mainstream press is hailing it as a shining moment, but the concurrent searches were the most rancid and deplorable jokes and stereotypes about Muslims. Do you think the Internet has churned up more racism than where our society would be at in our evolutionary trajectory if this technology did not exist?
I think we’ve merely learned that it exists. I don’t think we’ve made it any worse. Look at Trump’s rise. The strongest correlator with Trump’s support in the Republican primaries were racist Google searches. But Trump just tapped into something that was already there, or has never gone away. We’ve just never had a way to accurately measure it.
Do you worry about a feedback loop?
One danger is that once people realize how many people are racist it will become socially acceptable to be racist. You are starting to see people on the “alt-right” say “yes, I’m racist, there’s nothing wrong with that.”
One would hope somehow we’d find a way to make this new technology, you know, stop this.
It’s like the analysis of the Obama speech I did. You can say what causes racism or what lowers racism. With good data you could potentially turn this into more of a science. I think a lot of what we’ve been thinking about racism has been dead wrong. We’ve been using faulty data.
You are not a big fan of social media or the “Fakebook” phenomenon, right?
On social media people try to make themselves better, happier, richer than they are.
But people have to be aware that Facebook, Instagram and Twitter are, to a degree, a self-marketing scheme? Or has the bottom yet to drop out?
People are becoming more aware, sure. But the real danger is the data that suggests that social media makes people unhappy. The cultivated images people see of their friends. This could be dangerous for the longevity of social media companies. Perhaps Facebook will some day change their algorithm so it isn’t just photos of work colleagues in Hawaii.
What drives us to “hate-watch”? To masochistically seek out media that will drive us crazy?
[Sighs.] People are complicated. And this research has complicated my view of people. When you ask people questions they give you a coherent view of their own lives and their personalities, but when you see the Internet data you get a richer view. People tend to be different under different circumstances and have conflicting impulses.
There’s a lot in your book about sexual issues among married couples. The number one search query for married women is “is my husband gay?”
Yeah, it was higher than “is my husband cheating?” and compared to “is my husband depressed?” or “is my husband alcoholic?” it was 10 times more common. That shocked me, especially since, compared to the data I’ve seen, a woman is much more likely to be married to a man who is secretly depressed or alcoholic than gay.
Another thing that is revealing in the data is how many sexless marriages there are.
And your data is somewhat counter-intuitive to the old cliché of insatiable men – you say that there are more women searching about why their partners won’t sleep with them.
Particularly of younger women. A lot of “why won’t my boyfriend have sex with me?” One possibility is that pornography plays a role, that men are developing less of an interest in real sex because of porn.
Another thing that is revealing in the data is how many sexless marriages there are
Or perhaps an adjacent issue? That if a man is unhappy with an intimacy issue, he’s more likely to take action into his own hands, if you understand my meaning. Whereas women may be more prone to do a Google query.
Well, there are twice as many complaints of “my boyfriend won’t have sex with me” than “my girlfriend won’t have sex with me,” so I think there are more young men withholding sex based on this data.
Do you feel there is some healing you can do now that you’ve made this discovery?
Well, I publish in popular media, so I hope it trickles out there. I have a friend who is a therapist and I know the book has come up, which can make for some awkward conversations.
What is it about data that turns you on?
It’s because we’re so frequently misled, both by our own minds and by other people. We always assume people are exactly like us, or the people we hang out with. All my family and friends were shocked that Trump won, because my family and friends don’t know a single Trump supporter. Even a poll, which is so far from perfect, when it shows that 40% of voters like Trump, it’s like “oh, yeah.”
Were you yourself shocked by the election?
No. I predicted he would win from day one. I predicted it based on everything I saw on the Internet, all the “alt-right” sites and the racism.
‘When I did the search on racist jokes it was in the millions. It’s very widespread’
But what about the argument that, say, for every horrible, racist comment at the bottom of an article, there are a hundred silent people who don’t think that way? That there’s an overestimation? You know, one guy farts in a Church and it stinks the whole place up.
Yeah, but it’s not just one guy. I thought it was one guy. I thought these sites just appealed to a few people. When I did the search on racist jokes it was in the millions. It’s very widespread.
As a fact-oriented individual, the current administration and its love of “alternative facts” and “fake news,” it must be a difficult time for you.
Yes, because it isn’t hard to get the truth through reputable news outlets. We’re not living in an authoritarian regime, at least not yet. It’s just frustrating when you know he is lying, but the facts are so easy to find. What’s scary is what happens when someone is lying and we can’t access the facts.
It’s like when Scaramucci deleted his tweets, going in Orwell’s memory hole.
Even worse, there are people working to create that sound — and eventually look — exactly like a politician saying things they haven’t said.
That’s serious “1984″ business.
It’s all moving too fast for our old fashioned legal system, too.
Speaking of legalities, you show that a top search is for DIY abortions, which is terrifying.
I have been criticized that my areas of research skew left-wing. But when you get these data points, everyone agrees this is horrible, even someone who isn’t pro-choice.
What do you think about the phenomenon of “trolling”? Someone making a search for something they don’t believe, just to vent a demon out of them?
The thing is the data correlates so frequently with off-line behavior. The places that make the most racist searches are places that supported Trump more than the other Republican candidates, and it’s also where black people are paid less money. Time and again the correlation between areas with racist searches suggested poor outcomes for black people.
Okay, big question: How does this affect the Jews?
It’s not all bad news. Since the data set is so honest, you can find the times when you may just being paranoid. Anti-Semitism is a problem in the world and a problem in parts of the United States, but most of the places where anti-Semitism is high are where there are very few Jewish people. It’s among right-wingers in Idaho and Montana. So if you are worried that your neighbors are going home and Googling “kike jokes,” very few of them are. Anti-Semitism doesn’t have that much of a direct impact against Jewish people, but it is certainly part of this rising “alt-right” movement in certain places, especially the Pacific Northwest.
‘If you are worried that your neighbors are going home and Googling “kike jokes,” very few of them are’
The data will help us keep a watch on this. If Google Trends existed in the 1930s we’d say, “Okay, there’s clearly something happening in Germany and Europe that’s not good.” There was a lack of information back then. My grandparents in Hungary who had no clue what was going on. Data in general will make it harder to pull something off.
Did you learn anything positive in your research?
People are less judgmental about their partners’ bodies as you might think. People are very insecure about their own bodies and make all sorts of searches about it, but you don’t see many complaints about their partners and how to change them – so that’s hopeful.
Oh, that’s nice. Anything else, or that’s all ya got?
People who maybe think they are weird should feel less alone. Other people often keep the same issues and insecurities secret. People usually lie to you about that stuff.