Everybody Lies

When was the last time you told a little lie? Maybe you said you were fine when you weren’t, or you clicked “yes” on a survey just to get it over with. Seth Stephens-Davidowitz’s Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us About Who We Really Are begins with a simple but mind-bending premise—everybody lies. We lie to our friends, our lovers, our families, and even ourselves. But while we hide our real feelings and behaviors in public, we reveal our truth to Google.

Stephens-Davidowitz argues that Big Data—especially anonymous online data like search histories—offers the most honest window into the human mind ever created. By analyzing billions of Google searches, social media clicks, porn site data, and other digital traces, he contends that social science has entered a new era—an era in which we can finally see what people actually do and think, not just what they say they do or think.

The Hidden Side of Human Nature

At the heart of Stephens-Davidowitz’s thesis is the idea that people use digital spaces as confessional booths. When we search “Am I depressed?” or “Is my husband gay?”, we are revealing secrets we may not even admit to ourselves. This trove of data—the words people type when no one is watching—functions as what he calls a Digital Truth Serum. It allows us to measure prejudice, desire, health fears, or loneliness with unprecedented honesty. For instance, in his early work studying racism, Stephens-Davidowitz found that regions with the highest searches for racial slurs corresponded to areas where Barack Obama underperformed in elections. These online behaviors painted a much darker picture of American racism than traditional surveys ever revealed.

From Surveys to Search Boxes

Social science has long relied on surveys and interviews, where people answer questions like “Are you happy?” or “Do you watch pornography?” The problem, Stephens-Davidowitz insists, is that people lie. Surveys are filtered through social desirability, moral self-image, and fear of judgment. In contrast, millions of Google searches made in private reveal our real curiosities—our hidden desires, secret fears, and uncomfortable truths. The gap between survey data and search data highlights just how complex truth-telling is. For example, while surveys say only 8% of women watch pornography, searches for pornographic content among women suggest that the real number is far higher.

The Four Powers of Big Data

Stephens-Davidowitz organizes his argument around four revolutionary powers of Big Data:

• New Types of Data: Big Data allows scientists to study topics they couldn’t before—racism through search terms, sexuality through porn data, or even health symptoms through queries like “back pain and yellowing skin.”
• Honest Data: People lie to surveys but tell Google the truth. This honesty enables researchers to uncover hidden anxieties, prejudices, and desires at scale.
• Zooming In: Big Data makes it possible to analyze small, specific populations—men aged 30 who search for relationship advice in Ohio—or even moment-by-moment behavior, such as when people search for “football scores” or “feel lonely.”
• Cause and Effect Through Experimentation: The digital age enables millions of randomized experiments (A/B tests) that let companies, governments, and scientists directly test what influences behavior.

Why It Matters

Why should you care that Google knows you better than your therapist? Because Stephens-Davidowitz suggests Big Data can move social science from speculation to real science. Data can help discover early signs of pancreatic cancer, predict crime risk, expose hidden discrimination, and measure genuine happiness. It can also correct false narratives. The author shows, for example, that violent movies actually seem to decrease crime rates—because potential aggressors are sitting in dark theaters rather than bars. And the most elite schools, like Harvard or Stuyvesant High School, don’t necessarily cause success; natural ability matters more than prestige.

Seeing Ourselves Clearly

Ultimately, Everybody Lies argues that understanding the data we quietly generate every day can transform how we see society—and how we see ourselves. The book bridges economics, psychology, technology, and ethics to teach you how Big Data can both illuminate and complicate truth. Stephens-Davidowitz’s big claim is that we now have a “microscope for human nature.” The data might not always be comforting, but it’s real. And by confronting the real world—through the numbers we leave behind—we can at last begin to understand why we do what we do when we think no one’s watching.

We all trust our gut. You might feel confident in your instincts about people, relationships, or decisions—yet Seth Stephens-Davidowitz opens his argument in Everybody Lies by showing how easily our intuitive judgments are fooled. Data, he argues, doesn’t just confirm what we believe; it regularly proves us wrong. Human intuition often misleads us, and Big Data provides a corrective lens that helps us judge reality more accurately.

Grandma as Data Scientist

Stephens-Davidowitz begins with a charming metaphor: his grandmother giving relationship advice at Thanksgiving dinner. Based on her lifelong observations, she predicts what kind of woman would make him happy. She’s using decades of social data, accumulated intuitively from experience. Grandma’s wisdom, he writes, is good data science—it looks for patterns and correlations between personality traits and outcomes. But even Grandma can be wrong, because her sample size is tiny and biased by her own experiences. Just like survey data, everyday “common sense” is limited and anecdotal.

From Human Instinct to Big Data

Intuition works only when we’ve seen enough examples. To truly separate perception from truth, we need massive datasets. The Columbia University–Microsoft study on pancreatic cancer illustrates this vividly. By analyzing millions of Bing searches, researchers discovered that specific search patterns—like “back pain” followed by “yellow skin”—predicted early pancreatic cancer. That insight wasn’t a product of intuition, but of data volume large enough to detect faint, life-saving signals that no individual doctor could identify from experience.

When Gut Feelings Fail

Stephens-Davidowitz reinforces his point with examples of erroneous intuition. People assume winter blues are minor, yet Google searches for “depression” show climate to be one of the dominant factors—warm regions have 40% fewer depression-related searches than cold ones. Likewise, many intuitions about relationships are dead wrong. His grandmother believed couples with shared friend groups last longer. But Facebook’s massive relationship data found the opposite: couples with separate social circles are more likely to stay together. More data, not instinct, reveals what’s true.

The Limits of Experience

The problem is our minds are biased toward dramatic and memorable stories. We mistake vividness for truth—a psychological bias documented by Daniel Kahneman’s “availability heuristic” (see Thinking, Fast and Slow). We fear tornadoes more than asthma because tornado deaths make the news. Stephens-Davidowitz uses countless examples to show how the dramatic skews our judgment: experiences of failure, prejudice, or fear distort how we estimate reality. Big Data corrects this bias by aggregating billions of dull but representative facts instead of relying on striking anecdotes.

Why Data Beats Intuition

In the end, intuition is powerful but not precise. “Grandma might be wise,” Stephens-Davidowitz concludes, “but Google is wiser.” Big Data allows us to see patterns invisible to human intuition and filter out our emotional distortions. For you, this means accepting that gut feelings may help start hypotheses—but the numbers finish the truth. From health to happiness, from relationships to social prejudice, the data continuously reminds us that what feels right is often wrong—and what feels irrelevant may shape your world the most.

Stephens-Davidowitz presents four transformative “powers” that make Big Data revolutionary. Each power redefines what counts as truth and how we research our behavior.

1. Reimagining What Counts as Data

Big Data isn’t just bigger—it’s different. Traditional datasets captured surveys or official reports; now, everything you click, type, or buy is a data point. The author’s story of Jeff Seder predicting the success of American Pharoah by measuring the horse’s heart size shows how unconventional data, used creatively, can revolutionize a field. Similarly, Google’s early success came from recording a new type of information—link structure, rather than keyword frequency—which transformed how we find knowledge online.

2. Honest Data: The Digital Truth Serum

People hide their true selves from friends and surveys but reveal them online. Search engines thus provide a unique moral mirror. Racism, sexual insecurities, fears, and prejudice—all appear in what people type privately. Stephens-Davidowitz calls Google “the most honest dataset on the human psyche.” This honesty allows researchers to measure phenomena that were once invisible, from anxiety patterns to racial bias, giving leaders and policy-makers a clearer picture of society’s hidden problems.

3. Zooming In on Small Groups

The size of digital data lets scientists magnify tiny populations. Stephens-Davidowitz’s analysis of Google users shows how overwhelmed parents in certain states search more about “child neglect,” or how men in specific age brackets search for “am I gay?” In one study, Harvard economist Raj Chetty used IRS data to track the income mobility of tens of millions of people—revealing that opportunity in America differs drastically depending on the city you grow up in. This granular “zoom” transforms vague averages into precise, local realities.

4. The Ability to Test Causality

Big Data allows scientists and companies to run millions of natural or randomized experiments online. The rise of A/B testing—randomly showing groups two different versions of a webpage or ad—means we can directly measure cause and effect. Google does tens of thousands of these experiments every year; Facebook runs over a thousand a day. This power removes guesswork from human behavior, letting researchers understand what actually changes minds rather than what merely correlates with change.

Bringing It Together

Together, these four Big Data powers redefine social science. You could think of them as the modern equivalents of the telescope and microscope—tools that reveal realities once hidden. They let us measure the private, the local, and the causal in ways that were impossible before. The data isn’t perfect, and Stephens-Davidowitz later explores its limits, but its potential is undeniable: Big Data provides the clearest map yet of the messy, secret world inside human life.

One of Stephens-Davidowitz’s most unsettling insights is the difference between what people say and what they search. Online behavior, he demonstrates, exposes secrets that make surveys look polite and naïve. This is what he calls the Digital Truth Serum—a dataset so honest it can reveal things humanity would rather conceal.

The Lies We Tell

Surveys, Stephens-Davidowitz explains, fail because people perform for others. They exaggerate virtuous behaviors (voting, charity) and downplay shameful ones (porn use, prejudice). He cites a University of Maryland study showing alumni lied about their grades and donations. Google searches bypass such performance; no one lies to a search bar about whether they’re “depressed” or “regret having kids.” Comparing survey data to search data exposes this hidden world—where confessions pour out in keystrokes.

Sexual Secrets and Social Prejudice

Using anonymous porn and search data, Stephens-Davidowitz reveals people’s private desires. Roughly 5% of American men search for gay porn; yet surveys claim only 2–3% identify as gay. People in the most homophobic states make more such searches, suggesting the closet remains enormous. Racism too appears in shocking clarity—regions with the most searches for racial slurs or “nigger jokes” correlate with areas of wage inequality and lower support for Barack Obama. Online behavior demolishes the illusion that prejudice has vanished.

Personal Vulnerability

Searches also reveal private suffering: people confess fears of abuse (“my mom hit me”) or despair (“I am sad”). During the Great Recession, child abuse reports appeared to decline—but searches by children showed an increase, meaning cases were going unreported. Likewise, in states that restricted abortion access, online searches for “how to self-abort” or “coat hanger abortion” spiked. These patterns prove that Google data can act like a real-time social fire alarm, warning society where pain hides behind silence.

The Power and Responsibility of Knowing

The author insists that acknowledging these painful truths isn’t pessimism—it’s utility. Data can help target resources and interventions, from mental health campaigns to child protection services. But it also demands empathy. Seeing the world through search data reminds you that behind every anonymous keystroke is a person vulnerable, lonely, or afraid. The digital truth serum may sting, but its honesty can heal if we’re willing to listen.

Stephens-Davidowitz calls his third Big Data power “zooming in”—the ability to see fine-grained detail at the level of neighborhoods, minutes, or even individual keystrokes. Big Data lets you explore not only what the average person does, but what specific groups, contexts, or times reveal about human life.

Cities, Health, and Inequality

Economist Raj Chetty’s use of IRS data illustrates this perfectly. By analyzing billions of tax records, his team showed that the “land of opportunity” in the U.S. varies dramatically by geography. A child born poor in San Jose has nearly triple the chance of becoming rich compared to one born in Charlotte, North Carolina. The zoom lens exposes America as multiple societies—some enabling upward mobility, others locking generations into poverty.

Life Expectancy and Contagious Behavior

Chetty found more surprises when he examined millions of death records. Rich people live about fifteen years longer than poor people, but local culture matters too. Where poor people live near rich people, they live longer—possibly because health habits like exercise and diet are contagious. The social environment, not just income, influences longevity. Big Data transformed mortality from an abstract statistic into a behavioral map.

From Yearbooks to Satellites

Zooming in applies beyond people. Using photos of yearbooks from 1900s to 2000s, computer scientists tracked how Americans’ facial expressions changed—from serious to smiling, mirroring shifts in culture and commerce (helped along by Kodak’s advertising). Likewise, economists studying satellite night-light images discovered they could measure economic growth in developing nations by changes in nighttime brightness. Data at any scale reveals subtleties of progress, emotion, and inequality.

Why “Zoom” Changes Understanding

Traditional surveys flatten reality; zooming in restores texture. Stephens-Davidowitz insists that understanding comes not from averages, but from detail—the When, Where, and Who of data. This technique teaches you that truth lives in variation. Whether you’re testing classroom success, local health, or even fame (as he does through Wikipedia data on notable births), the zoomed-in view reveals patterns that broad strokes will always miss.

What if every click you make is part of an experiment? Stephens-Davidowitz argues that the digital era has transformed everyday life into one vast laboratory. Online platforms run thousands of random tests daily, letting us learn not just what’s correlated but what’s causal.

A/B Testing: Rapid Causality

This technique—called A/B testing—randomly shows two versions of content to different users and measures which performs better. When Google engineers tested twenty versus ten search results, Facebook tested hundreds of versions of its buttons, and Obama’s campaign tried various home page photos, they weren’t guessing; they were measuring behavior scientifically. The winning photo of Obama’s smiling family increased sign-ups by 40%, proving that data-driven design can change real-world outcomes.

Natural Experiments: When Life Randomizes

Not all experiments are intentional. Stephens-Davidowitz shows how economists uncover cause and effect from the randomness of life—like the near-miss of an assassination or the result of a football game. When the New England Patriots barely beat the Ravens, advertisers unknowingly ran more Super Bowl ads in Boston than Baltimore. Comparing subsequent movie ticket sales showed a clear causal effect of advertising. Random life events thus become “Nature’s experiments,” revealing truths that scripted trials couldn’t.

From Medicine to Schools

Stephens-Davidowitz highlights psychologist Esther Duflo’s experiments in Indian schools, where paying teachers small daily bonuses cut absenteeism in half. He compares this to how digital testing is democratizing experimentation—you could test teaching methods, buying habits, even sleep improvement programs. Jawbone used experimental nudges (“Commit to sleep by 11 p.m.” reminders) to help users add twenty minutes of rest. Data turns behavior change into a measurable science.

The Ethics of Constant Experimentation

But when everything’s a lab, who gives consent? Stephens-Davidowitz raises ethical concerns about manipulation and addiction. Companies that test endlessly—optimizing color, timing, and tone—can design irresistibly addictive apps. As Tristan Harris warns, “There are thousands of people on the other side of the screen whose job is to break down your self-regulation.” For Stephens-Davidowitz, experimentation holds promise and danger: it’s a great teacher, but we must decide what kind of lessons we’re willing to learn.

Stephens-Davidowitz helps you see not only the power of Big Data but its pitfalls. He warns that data’s seductions—the numbers, the correlations, the scale—can mislead or harm if used carelessly or unethically.

The Curse of Dimensionality

One major flaw of Big Data is what statisticians call the curse of dimensionality: if you test enough variables, some will show spurious correlations. Stephens-Davidowitz illustrates this using a hilarious “Coin 391” metaphor: flip a thousand coins daily for two years, and one will seem magically correlated with stock market movements—pure illusion. Many marketers and financial quants make the same mistake, confusing random noise for meaningful patterns. The author’s failed attempts (with Lawrence Summers) to predict markets using Google searches prove that even brilliant minds can fall for data mirages.

When Measuring Distorts Meaning

Another trap is overemphasis on measurable outcomes. Stephens-Davidowitz tells the story of Yale professor Zoë Chance, who became obsessed with her pedometer scores—chasing steps instead of well-being. He argues that schools, companies, and governments do the same: when you measure what’s easy (test scores, clicks, steps), you risk losing sight of what matters (learning, happiness, meaning). Facebook may optimize for engagement, but engagement isn’t the same as connection. In baseball, teams once ignored defense because it was hard to measure—a mistake that cost the data-driven Oakland A’s up to ten wins per season.

Balancing Big and Small Data

The antidote, Stephens-Davidowitz says, is humility and hybrid measurement. Pair mass data with human judgment—small surveys, direct observation, interviews. Facebook mixes analytics with user feedback; the Bill & Melinda Gates Foundation combines test results with teacher evaluations. Big Data alone can’t capture empathy, creativity, or nuance. You must complement the numbers with stories.

Ethical Use of Insight

Finally, there’s the human side. Data insights can empower governments or corporations to predict behavior with uncomfortable precision—creditworthiness, political leanings, or risk of crime. Stephens-Davidowitz cautions that analytics must not become surveillance. Ethical use means transparency, consent, and awareness of data’s limits. Numbers enlighten, but they can also coerce. The challenge of Big Data, he concludes, is staying human while quantifying humanity.

In the conclusion, Stephens-Davidowitz declares that social science is finally becoming real science. Big Data allows us to ask—and answer—questions Karl Popper and past thinkers considered unfalsifiable: What causes happiness? What predicts prejudice? How do choices ripple through society? Through mass data and scaling, humanity is turning its mysteries into testable theories.

From Freud to Algorithms

Early psychology, like Freud’s dream interpretations, dealt in unprovable ideas. Now, search data lets scientists falsify them. Freud’s claim that people dream of phallic symbols? Wrong—Big Data proved bananas appear in dreams only as often as they’re eaten. Conversely, his intuition about sexual fixation and early experiences? Confirmed: incest fantasies and parental attractions appear frequently in porn searches. In this way, Big Data provides experimental evidence for the human psyche—no couch required.

Science at Scale

Stephens-Davidowitz envisions “science at scale,” where small insights multiply through massive datasets. The mapping of John Snow’s cholera cases becomes a model for modern digital epidemiology—tracking diseases through search behaviors. Similarly, A/B testing, once limited to companies like Google, can now improve education, sleep, and mental health. When students, patients, and users engage online, their collective outcomes become evidence for what works.

The Future Data Scientist

In a personal note, Stephens-Davidowitz situates his work as a successor to Steven Levitt’s Freakonomics. Levitt used creative small datasets to study quirky truths; Big Data lets us answer grander ones. The next Freud, Marx, Kinsey, or Salk, he predicts, will be a data scientist. Curiosity and creativity now combine with algorithms to decode humankind at scale.

Facing the Truth

When Stephens-Davidowitz finally stops writing—realizing most readers don’t finish every book chapter—his decision itself is an experiment based on data. His closing reminder is lighthearted but profound: learning from data doesn’t just mean collecting it; it means acting on it. If we follow wherever data leads—honestly, ethically—it can make social science real, policy effective, and introspection meaningful. In short, Big Data makes us see that knowing humanity’s truths is uncomfortable but necessary—and that’s what makes it science.