How do I know if my partner is lying?

You usually can’t tell from how your partner looks, sounds or texts. When lies do come out, it’s mostly through facts that don’t fit together, and often well after the fact.1 So the most useful thing you can do tonight isn’t watching harder. It’s separating what you know from what you’re guessing, and asking about the gap.

Why can’t you just tell?

Picture Rachel, 34. She and Owen, 36, have lived together for a year and have a trip booked to see her parents in November. For three weeks Owen has been getting home around 8 instead of 6:30, which he says is the gym. Last Thursday he texted at 11:40 p.m.: “dinner ran long, phone almost dead, home soon.” On Sunday he mentioned that same dinner as a client thing, though on Thursday it had been his team. Rachel has been replaying his face from Sunday morning ever since.

Here’s what the research suggests about that replay. In a 2002 study by Hee Sun Park, Timothy Levine and colleagues, people described a time they had actually caught someone lying. Most of those lies weren’t spotted in the moment. They came out later, through someone else’s account, physical evidence, a confession or a detail that clashed with what the person already knew.1 A 2020 replication found that when people described real catches, they mentioned behavior less and context more than when they described how lies should be caught, though behavior still came up most in their answers.1

There’s also a reason you didn’t suspect anything until recently. Levine’s truth-default theory argues that people generally presume what others tell them is honest, and that this default works well most of the time.2 Believing your partner isn’t naivety. It’s how close relationships are supposed to run.

Do partners lie less than other people?

In ordinary life, mostly yes. In diary studies by Bella DePaulo and Deborah Kashy, 77 students and 70 community members logged every social interaction and every lie for a week. People told fewer lies per interaction to the people they felt closest to, and felt worse when they did.3 That doesn’t mean a partner can’t lie about something big. It means a lying-to-you default isn’t the starting assumption the research supports, and a few weird weeks are worth understanding before they’re worth prosecuting.

Then what does count as evidence?

In Rachel’s month, the evidence is short:

• Owen’s routine changed and has stayed changed for three weeks.
• He described the same dinner two ways.
• The trip has been moved twice and isn’t rebooked.

Everything else (the gym being a cover, the late text meaning something, the “rehearsed” calm) is reading. The reading might turn out right or wrong. Right now it’s built from fear, and fear is very good at stories.

It helps to keep four things apart:

• What happened: the routine, the two versions, the trip.
• The reading: he’s hiding someone.
• The fear: the life we’re building is not what I think it is.
• The next step: ask about the specific things, together, in person.

Should you check his phone?

A lot of people do. In a Pew Research Center survey, 34% of partnered adults said they had looked through a partner’s phone without them knowing, including 42% of women and 25% of men, while seven in ten Americans said that’s rarely or never acceptable.4 The problem isn’t only fairness. A phone gives you fragments without context, which tends to feed the reading rather than settle it, and now you’ve got something you’re hiding too.

If you’re the one being watched, the picture changes. A partner who demands your passwords, tracks your location without agreement or checks who you’ve been talking to is doing something worth taking seriously, whatever they suspect.

How do you actually ask?

Bring the evidence, not the theory. “You’ve been home at 8 most nights, the dinner came up as two different things, and we’ve moved the trip twice. I’m not accusing you of anything, but I’ve noticed, and I’d rather ask than keep guessing. What’s going on?”

Talking about it directly is what more secure people tend to do. In a study of 213 people who had discovered a partner’s lie, those with a secure attachment style were more likely to talk about the issue, people with an anxious style were more likely to avoid the topic, and people with an avoidant style were more likely to avoid the partner and end things.5 Avoiding the subject can feel safer at 11:40 p.m. It mostly keeps you guessing.

Then listen for two things. Does the explanation fit the facts you have? And how does he handle being asked? An answer that fills in the picture, even an uncomfortable one (stress at work, a health worry, cold feet about the trip), is information. So is anger or contempt for the question itself.

What should you do this week?

• Write the evidence list. Only things you could show someone else. Keep the guesses on a separate page.
• Pick a calm time. Not at midnight when the late text arrives, and not by text at all if you can help it.
• Ask about the shared plans first. The trip, the calendar, the lease. Those are yours too, and they’re the easiest place to start.
• Watch what happens over the next few weeks. Does the routine make sense once it’s explained? Do the stories stay consistent?
• Get support for yourself. A friend, a couples counselor, or, if the worry is constant and wearing you down, a licensed mental-health professional.

And if you ever feel afraid of your partner, or watched, controlled or threatened at home, that matters more than any question about the truth. Reach out for help.

What doesn’t the research show?

• The detection studies relied on people recalling lies they caught, so lies that were never caught are missing by definition.
• The diary studies tracked everyday lies, mostly small ones, not lies about affairs or major decisions.
• The attachment study asked people to look back on a lie they’d discovered, so it describes how they remember responding, not what works best.
• Truth-default theory is a framework, not a measurement of how often your particular partner tells the truth.
• The Pew figures describe what people report doing, not what they found or whether it helped.