Start here, because most of what you’ve been told about this is wrong.
Not spun. Not exaggerated. Wrong — traced back to a source that doesn’t exist, or a study that got retracted, or a number somebody made up in 1977 that everybody has been repeating ever since.
This matters more than it sounds like it should. You are going to make one of the largest decisions of your life, possibly with a child watching. If the map in your head is drawn from a book that cited nothing, you will navigate accordingly.
So before anything else, here’s what the research actually says about the things everybody knows.
“Wait a year before you date again.”
There’s a whole family of these. Wait one year. Wait one year for every five you were married. Wait three years and your next marriage is twenty-two percent more likely to last.
That last one shows up everywhere. It traces to a commercial blog post. It is not in any research publication, from any institution, anywhere.
The actual study is Wolfinger, 2007, following 1,171 remarried adults. He tested exactly this question — does remarrying quickly predict a shakier second marriage? The answer was no. No relationship at all between how fast someone remarried and how stable that marriage turned out.
Now hold on, because there’s a real finding hiding next to the fake one, and they get confused constantly.
Courtship length does predict stability. Couples who dated one to two years before getting engaged had roughly twenty percent lower divorce odds than couples who dated less than a year. Three years or more, and it drops further — somewhere between thirty-nine and fifty percent lower.
Read those two findings side by side and you get the actual rule. The clock that matters starts when you meet this person. It does not start at your divorce.
You do not owe anyone a waiting period. You owe the next relationship enough time to actually see who you’re with.
But — and this is the part the waiting-period advice was clumsily reaching for — if you have a child, there is a real constraint here. It just isn’t a calendar. We’ll get to it in Chapter Eight, and it’s the most important rule in this guide.
“Sixty to seventy-three percent of second marriages fail.”
Also untraceable. Nobody can produce the study.
The Bureau of Labor Statistics has been following the same Americans since 1979. By age fifty-five, 39.1% of second marriages had ended.
That’s not nothing. But it’s not a coin flip weighted against you, and it’s not the seventy-three percent that gets quoted at divorced men by people who mean well.
“You have to love yourself before someone else can love you.”
This one is on ten million graphics. It has never been tested until recently, and when it was, it didn’t hold.
A 2025 study of 460 people looked at whether self-love actually functions as a prerequisite for romantic love. What they found was partial: self-care and self-acceptance did relate to passion, intimacy, and commitment. Other dimensions of self-love didn’t relate at all. The researchers noted plainly that there had been no scientific evidence for the popular claim before their study.
Here’s the defensible version. Low self-esteem is a genuine risk factor. People who think poorly of themselves tend to underestimate how much their partners love them, seek less reassurance than they need, and sometimes select partners who confirm what they already believe about themselves.
That’s a risk to manage. It is not a gate you have to pass through before you’re allowed to love somebody.
Plenty of people who don’t much like themselves love other people extremely well. If you’ve been waiting to feel fixed before you’re allowed to try, you’ve been waiting on a rule nobody ever established.
“Work on yourself first and the right person will show up.”
This is the bigger version of the same idea, and the research complicates it in a way that’s worth sitting with.
In 2020, a team led by Samantha Joel pooled 43 separate datasets covering 11,196 couples and used machine learning to find what actually predicts whether someone is happy in their relationship.
Your partner’s individual traits — their personality, their self-reported qualities, who they are on paper — explained about five percent of your satisfaction.
The things that happened between you — commitment, appreciation, sexual satisfaction, how conflict went — explained up to forty-five percent.
Sit with that. Who your partner is, as a person, is a small part of whether you’ll be happy with them. What the two of you do together is most of it.
So “become your best self and the right person will appear” has the mechanism backwards. Becoming a better person, in the abstract, doesn’t do much. Becoming someone who behaves differently in a relationship does — and those are trainable, specific things: how you regulate when you’re flooded, whether your partner feels understood, whether you express appreciation, whether you repair after conflict.
There’s a related finding worth knowing. Individual therapy is not established as improving relationship outcomes on its own. A 2025 review found couple and family interventions outperform individual therapy specifically for relationship distress. Individual work absolutely helps when it’s treating something specific — depression, trauma, regulation. But “I did years of therapy” is not the same as “I’m ready,” and nobody should be graded that way.
You don’t have to arrive finished. You have to arrive willing to do specific things.
“You can tell when someone is lying.”
You can’t.
This is the largest analysis on the question: 206 studies, 24,483 people judging whether someone was lying. Average accuracy was 54 percent. A coin flip is 50.
People caught 47% of lies and correctly identified 61% of truths, which means the whole edge came from a bias toward believing people, not from actually detecting anything.
Trained professionals — police, investigators — were not meaningfully better.
And the tells everybody knows? Gaze aversion, fidgeting, touching the face, looking up and to the left? A separate analysis found nonverbal behavior is barely related to whether someone is telling the truth. Confident liars look honest. Nervous honest people look shifty.
This is the single most important finding in this entire guide, and it’s why the system you’re about to read is built the way it is.
If you cannot read people — and you cannot — then reading people cannot be your method. Watching what they do over time has to be your method instead.
What actually improves accuracy isn’t better instincts. It’s checking stories against each other, checking claims against facts, and watching behavior across months and situations. That’s not intuition. That’s record-keeping.
“Opposites attract.” Also: “You need someone just like you.”
Both wrong, in an interesting way.
Three nationally representative samples — Australia, Britain, Germany, 23,250 people total — tested whether personality similarity between partners predicts relationship satisfaction.
Similarity explained less than half of one percent.
Not a small effect. Effectively nothing.
What did predict satisfaction was your partner’s personality itself. Specifically their agreeableness, their conscientiousness, and their emotional stability. Having a partner with those qualities predicted your happiness. Matching them didn’t.
And complementarity — the opposites-attract theory — has no support either. It’s not that difference is good and sameness is bad. It’s that temperament matching is measuring the wrong thing entirely.
Some things do need to match, and Chapter Four covers them. But they’re not personality. They’re the concrete ones: whether you both want children and on what timeline, whether your religious practice looks the same in an actual week, whether you drink the same amount, how you each handle money.
Those matter. Whether you’re both introverts does not.
“The algorithm found your match.”
In 2012, five of the most respected relationship researchers alive published a sixty-four-page review of the online dating industry in Psychological Science in the Public Interest.
Their conclusion: there is no compelling evidence that any online dating matching algorithm actually works.
Not “the evidence is mixed.” Not “more research is needed.” No compelling evidence, at all, that matching algorithms do what they claim.
The reason is structural, not technical. Compatibility isn’t a property of two people that exists before they meet. It emerges from how they actually interact — and no formula operating on two profiles can compute something that doesn’t exist yet.
The apps are useful for one thing, and it’s a real thing: access. They put you in front of people you’d never otherwise meet. That’s genuinely valuable.
The matching is marketing.
Worth knowing alongside this: about 68% of couples were friends before they were romantic. Two-thirds. Meanwhile roughly three-quarters of relationship research studies strangers meeting cold, because that’s easier to run in a lab. The dating-market frame everyone has absorbed describes a minority pathway.
“Most couples divorce after losing a child.”
The number you’ve heard is eighty or ninety percent.
It’s false, and its origin is documented. It appeared in a 1977 book called The Bereaved Parent. The book cited no source. Nobody has ever been able to find one. It became fact by repetition.
The actual numbers: a large survey by The Compassionate Friends found about 16% of bereaved parents divorced, and fewer than half of those cited their child’s death as a factor. An earlier survey of 14,852 bereaved parents found 12% divorced, with 72% still married to the same person.
A 1998 review of the literature found no evidence of elevated divorce among bereaved parents at all.
I include this one because it illustrates something important about how these numbers work. A statistic that terrible, repeated for fifty years, told grieving parents their marriage was probably doomed. It came from nowhere. Somebody wrote it down and everyone believed it.
“Men leave their wives when the wife gets sick.”
There was a study. It made headlines in 2015. It said divorce risk rose when wives fell ill but not when husbands did.
It was retracted.
Researchers at Bowling Green couldn’t replicate it. The original author found the error: one line of code had miscoded marriages lost to attrition as divorces. The reported divorce rate of 32% was actually about 6%.
The corrected analysis found no general gender effect. There was one narrow exception — heart problems specifically — and that’s it.
Most marriages survive serious illness. The story you’ve heard about men abandoning sick wives came from a coding error that got fixed, after the headlines had already gone around the world.
“It gets better after the kids leave.”
This is the U-curve — the idea that marital happiness is high at the start, dips during the child-raising years, and recovers in the empty nest.
It’s an artifact of bad study design.
When researchers followed the same couples across five waves over seventeen years, instead of comparing different couples at different stages, the U disappeared. What they found was decline at all durations, and no late-life recovery.
The reason cross-sectional studies show a U is survivor bias. Unhappy marriages end. So when you survey couples married thirty years, you’re surveying the ones that made it — and they look happier than the twenty-year group because the miserable twenty-year couples are still in the sample and the miserable thirty-year couples are divorced.
Why include something this bleak in a guide about choosing a partner?
Because you should know what you’re actually signing up for. The average marriage does not bend back upward on its own. What holds one together is not the natural arc. It’s dedication — an actual decision, made repeatedly — plus the belief that this is built rather than found, plus the ability to repair. Chapter Nine is entirely about this.
Going in knowing that is better than going in expecting a rebound that isn’t coming.
“Black fathers are absent.”
This one has a different character than the others. It isn’t a bad statistic that spread. It’s a story that persisted while the data said the opposite.
The CDC’s National Survey of Family Growth asked fathers about daily caregiving. Among fathers living with their children:
- 70% of Black fathers bathed, dressed, diapered, or helped their child use the toilet every day — compared to 60% of white fathers and 45% of Hispanic fathers.
- 41% of Black fathers helped with homework daily — compared to 28% of white and 29% of Hispanic fathers.
Black fathers were the most involved group measured, on the measures that describe what a father actually does on an ordinary day.
Among fathers not living with their children, Black fathers were also more involved than nonresident fathers of other races on many measures. And research on fragile families found something directly relevant to this guide: Black fathers’ involvement declines less when a relationship ends and less when they repartner. The paternal role holds outside the couple.
If you are a Black father reading this while carrying around a story about what men like you do — put it down. It’s not true, it was never true, and the numbers have said so for over a decade.
What this means for how you read the rest
Nine myths. There are forty-five in the full list, and it’s on its own page.
But notice what they have in common. Almost every one of them offers a shortcut — a rule, a timeline, a tell, an algorithm, a personality match. Something that lets you skip the slow work of finding out who someone actually is.
The research keeps pointing the same direction. There is no shortcut. You cannot read people, you cannot match on paper, and you cannot know in three dates.
What you can do is watch behavior across time and situations, ask about what someone has actually done rather than what they say they would do, and keep an honest record of the difference between what you were told and what you saw.
That’s the system. The rest of this guide is how to run it.
Next: Chapter Two — Do You Actually Want This? The question almost nobody asks first.