How do I know if they're lying?

Most of the time, you can’t tell someone is lying by watching them, and neither can anyone else. Lies tend to come out through information that doesn’t fit (a detail that contradicts another detail, a story that shifts) gathered over days or weeks, so the useful question isn’t “did they look guilty?” but “does this still add up?”

Can you see a lie on someone’s face?

Not reliably. In a large 2006 review, Charles Bond and Bella DePaulo pooled 206 sets of results covering 24,483 people judging whether others were lying. On average, people were right 54% of the time, barely above a coin flip. They correctly spotted 47% of lies and correctly accepted 61% of truths.1 That second number matters: most of us lean toward believing people, which is usually a good thing for getting through life with other humans.

So if you’ve been studying someone’s eyes across the dinner table, you’re doing what nearly everyone does. When researchers asked people in 58 countries how they could tell someone was lying, the most common answer worldwide was that liars avoid eye contact.2 It’s a near-universal belief. It just doesn’t hold up well.

Why don’t the classic tells work?

Because the differences between liars and truth-tellers are small, and they overlap with how honest people act when they’re nervous.

Bella DePaulo and colleagues combined 1,338 measurements of 158 possible signs of lying. Liars, on average, gave somewhat less information, told less convincing stories and seemed a bit more tense.3 But “on average” is doing a lot of work there. Plenty of truthful people are tense when they feel accused, and plenty of practiced liars are relaxed.

Here’s the part most people find surprising. The problem isn’t that we’re watching the wrong things. In a 2019 review of the field, Aldert Vrij, Maria Hartwig and Pär Anders Granhag concluded that the nonverbal signs of lying found so far are faint and unreliable.4 Training yourself to watch hands or blink rates won’t fix that, because there’s not much signal there to find.

Picture Jess. Her partner, Tom, comes home late, says he was stuck at a work thing, and fumbles for his phone while he talks. Jess’s stomach drops. But the fumbling, the half-second pause, the glance at the floor: all of it fits a tired person who got grilled the moment he walked in. Jess’s alarm is real. The evidence for it, so far, isn’t.

So how do lies actually come out?

Mostly through facts, and mostly later. Researchers who ask people how they discovered real lies in their own lives keep finding the same thing: the lie came out through physical evidence, something a third person said, or a confession, not through a nervous face in the moment.5 The same research found that people still believe they catch lies through behavior, which is part of why we keep trying. Deception researcher Timothy Levine argues that this is where the real skill lies: in content and context, meaning what you know about the facts and the situation, rather than in someone’s demeanor.6

That changes what’s worth paying attention to. One moment, however suspicious it feels, carries very little information. A pattern over time carries a lot:

• Stories that drift. The trip was for work, then it was with friends, then it was “a bit of both.” One version can be a misunderstanding. Three versions is a pattern.
• Details that collide. What they said versus what someone else saw, a receipt, a date that doesn’t work.
• Gaps that stay gaps. Ordinary vagueness (“I don’t remember, it was a while ago”) is normal. A gap that gets defended, re-explained, then changed is different.

None of these prove anything about intent. People misremember, round off, and avoid awkward topics for reasons that have nothing to do with betrayal. But they give you something concrete to ask about, which a pause never will. Asking also helps more than staring: in a 2017 meta-analysis, interview approaches that ask for more detail or pose unexpected questions led to more accurate lie and truth judgments than standard questioning.7 You’re not running an interrogation, but the lesson carries over. A curious, specific question beats a long hard look.

What should you separate before you say anything?

When suspicion kicks in, four different things tend to fuse into one hot feeling. Pull them apart:

1. What happened. “He said he was at the office until nine. His coworker mentioned everyone left at six.”
2. The reading. “He was somewhere he didn’t want me to know about.”
3. The fear. “I’m being made a fool of, again.”
4. The next step. “Ask him about the gap between six and nine.”

Only the first one is information. The second is a guess, the third is about you and your history, and the fourth is the only part you control. If you skip to the reading, you’ll walk in with a verdict. If you start from what happened, you walk in with a question.

What to do this week

• Write down the mismatch, not the vibe. One or two sentences, facts only. If you can’t write a fact, you have a feeling, and it deserves attention, but not an accusation.
• Ask once, specifically, and calmly. “You said the trip was for work, and then you mentioned it was with friends. I’m confused. What was it?” Then listen to the whole answer.
• Watch the next few weeks, not the next five seconds. How they handle a fair question over time tells you more than how their face moves when you ask it.
• Don’t build a case in secret. Going through their phone or quietly testing them tends to make you more anxious and gives you less you can actually talk about.
• Notice if this is new or familiar. If every partner eventually feels like a suspect, that’s worth looking at for its own sake.

If suspicion is running most of your day, wrecking your sleep, or lasting for weeks with nothing new to explain it, a licensed mental-health professional can help you sort the fear from the facts.

What doesn’t the research show?

• Most lie-detection studies are lab studies, often with strangers and low stakes. Partners know each other’s habits, which might help a little, or might just make people more confident without making them more accurate.
• These are averages. They don’t say whether your partner is lying. They say that watching for nervous behavior is a poor way to find out.
• Discovery research relies on memory. People recalling how they caught a lie may remember the dramatic evidence more than the slow build of suspicion.
• None of this tells you their intent. A mismatch is a reason to ask, not proof of a betrayal.