YOU KNOW WHO HUMANS DON’T LIE TO?
Why we tell Google our innermost thoughts and feelings — and that may be a problem for researchers
Humans are a dishonest species.
We’re not honest with our bosses, our friends, our family, our intimate partners, or with researchers who phone us for a survey.
But there’s one situation when we don’t lie. We don’t lie to Google. The search engine is simultaneously our therapist, our confessional and our most trusted friend. Data scientist and economist Seth Stephens-Davidowitz, formerly of Google, has pored over the unfathomably large body of data concerning when, where and what people search on Google around the world. In his book Everybody Lies, he reports insights about what people are really thinking and feeling. He’s learned Google users are kinkier than they admit, and more racist too, but most of all, people are desperately anxious and want to know they’re not alone.
You were laughed out of academic journals with this idea. Now everybody’s interested. What happened?
I was a graduate student in economics, and I became obsessed with all you can learn from what people search, particularly things people wouldn’t otherwise admit. People lie to surveys, but they tend to be really, really honest on Google. I was considered very strange. I couldn’t get the work published. I couldn’t get an academic job. I was looking for other ways to get the work out there, and now I’ve written this book.
Did you learn anything profound about human nature?
I think anxiety and insecurity is probably more widespread than we usually think. That would be a big takeaway. But I think also just people are complicated. The traditional methods of data collection have given us a very limited view of the human psyche.
Do you have a favourite finding?
Men make as many searches looking for how to give themselves oral sex as how to give a partner oral sex. My other favourite, which you would probably also put in the weird or one-off category — but I don’t think it is — is that the top search in India starting “My husband wants” is “My husband wants me to breastfeed him.” It points to this idea that there are facts about human nature we didn’t know. There seems to be a somewhat widespread interest in adult men being breastfed in India. It hadn’t been picked up by any of the usual data sources and it isn’t talked about. Well, why India? What causes this to be so much higher in India and nowhere else?
Is this (research technique) going to change the way we look at elections? (Searches containing racial slurs were strongly associated with areas that unexpectedly went for Trump, such as Michigan and western Pennsylvania).
Surveys are getting worse and worse. The response rate is now under 10 per cent. Internet data is getting better and better. In a couple of election cycles, we’re not going to be using surveys anymore. But (with election predictions), we’re just going to predict an event that’s going to happen in three weeks. The attention (on elections) is so enormous relative to its importance.
If elections aren’t interesting to you, what is?
I talk about child abuse in the book. That’s one area where the data is not good, because most child abuse cases aren’t reported. But it turns out a decent number of children, really sadly, make searches such as “my dad hits me” or “my mom beats me.” So we now maybe have the best data ever on when and where child abuse is rising. And I talk about racism and hatred. We can break it down minute by minute. We can see how people are responding to the words politicians use in speeches — is it calming an angry mob, or inflaming an angry mob? That is, I think, pretty revolutionary.
What are your findings on that front?
There is clearly a strategy that is much better at calming an angry mob. And it’s basically don’t lecture to them, but provoke their curiosity. Talk about Shaquille O’Neal being Muslim; don’t talk about how it’s someone’s responsibility not to hate Muslims.
What are you looking for in the data now?
I’m researching anxiety. I’ve become obsessed with it, because there are lot of things in the data that are really, really surprising. I’ll give you one example: When Trump was running for president, he was saying a lot of scary things. Pretty much all my friends and family members and liberal people said they’re terrified. Now, if you look at the data in parts of the United States that are really liberal, you don’t see an uptick in searches for panic attacks or anxiety or anything like that.
(Maybe) people don’t Google anxiety about Trump, even if they’re really anxious. I’d be really surprised by that. The second possibility is people have a fixed amount of anxiety — they would have been anxious about their jobs or their kids, but now they’re anxious about Trump. That would be a revolutionary change in how we think about anxiety. The third possibility is that people exaggerate how anxious they are about Trump because it’s politically correct, when they tend to actually be much more anxious about their own personal situation. But you don’t bother your friends with that. You sound like a good person if you’re anxious about Trump.
One of your oddest findings is that people often type confessions like “I’m drunk” into Google. What could they possibly be hoping to find?
It’s very strange. It’s a little bit like the confessional in Catholicism. It is a widespread use of Google to type complete sentences into the search engine. You (may) get message boards where people are feeling similarly, so you feel less lonely. If you type, “I hate my boss,” you might get message boards of people complaining about their bosses. If you type, “I’m sad,” you get message boards of people who are also sad and you realize, “Oh, I’m not alone.”