If you typed "should I get a divorce" into Google tonight, you're probably not in a crisis you can name in a sentence. You're somewhere harder. You've been living in some version of this question for weeks or months or years. You've thought about it in the car. You've thought about it after a fight, and then again three days later when there's no fight, just the quiet of being roommates. You've imagined leaving and you've imagined the next ten years of staying. Both versions feel impossible, and neither feels right.

This is the longest piece we've written on Emira because the question deserves it. Most articles you'll find tonight will fail you in one of three ways. The faith-based ones will tell you what scripture says, which is useful only if scripture is your decision-making framework. The therapy-affiliated ones will quietly upsell their program. The hot-take ones will tell you to leave or stay based mostly on the writer's own life. Almost nobody gives you a real, structured way to actually answer the question with the partner who's on the other side of it.

That's what this is. We won't tell you what to do. We're not in your relationship. What we will give you is a framework that does three things: separates the four very different situations the question hides (because the right next move depends entirely on which one you're in), names what most articles refuse to name about the lived shape of this question, and gives you a concrete way to find out what your relationship is actually pointing toward. The pieces of the framework are drawn from Gottman's research, attachment theory, the Doherty studies on divorce reconciliation, and our own work with couples at decision-point.

A note on language. We'll use "divorce" because that's the keyword you typed, but most of this applies just as cleanly to long-term unmarried committed partners weighing whether to step away. The legal mechanics differ; the human question is the same.

A note on what this guide is not. It is not a substitute for a therapist if you have one available. It is not a tool for a relationship that involves ongoing harm, if your relationship is unsafe, please call the National Domestic Violence Hotline (1-800-799-7233) or your local equivalent. And it is not a guarantee of either outcome.

If you'd rather start with a faster self-read, the 2-minute disconnection quiz will tell you which decision-band you're in. Use this guide as the deeper context.

The question hides four very different situations

The single biggest reason most "should I get a divorce" articles fail is that they treat the question as one thing. It's not. When people search this query, they're almost always in one of four situations, and the right next move is completely different for each.

Situation 1: Acute crisis. Something specific has happened, recently, that has put the relationship at a fork. An affair has just been discovered. A betrayal of trust has surfaced. A boundary has been crossed in a way that can't be ignored. The partnership feels cracked open. Decisions made in the first weeks after a crisis are statistically poor, the nervous system isn't yet in a state to evaluate. The right move here is almost never "decide now." It's "stabilize first, decide later."

Situation 2: Slow drift. The relationship hasn't broken; it has thinned. Conversation has reduced to logistics. Physical affection has become rare. You can name a year, possibly several, where the relationship has been quietly running on routine. Nothing specific is wrong, which is its own version of wrong. This is the most common situation behind the search query, and also the most recoverable.

Situation 3: One-sided wanting-out. One of you is much further along in the question than the other. One partner is actively imagining leaving, has been for a while; the other doesn't know yet, or knows and is hoping it passes, or is working hard to repair while the first partner is mostly already gone. Therapist Becky Whetstone calls this the Decider / Leaner-in dynamic, and it's the lived shape of most "should I get a divorce" Googling. It's also the situation almost no article addresses honestly.

Situation 4: Hard-line dealbreakers. Something is going on that, in most readers' moral framework, would qualify as a reason to leave: ongoing abuse, untreated severe addiction, repeated infidelity without remorse, fundamental safety risk. The question being asked here is rarely "should I leave" in the analytical sense. It's "do I have permission to leave," "can I afford to leave," or "what will it cost me if I do." The answer to the should-I question is almost always yes; the work is the practical execution and the support to do it.

The four situations call for fundamentally different next moves. We'll work through each.

Situation 1: Acute crisis (something specific just happened)

Don't decide right now. That's the whole framework for this situation, but it's worth unpacking.

When a discovery has just landed (an affair, a hidden financial life, a betrayal you can't unsee), the brain is doing two things simultaneously. The first is shock processing, which produces certainty in either direction that often doesn't last. The second is loss-aversion, which makes the cost of change feel disproportionately large compared to the cost of staying. Both of these distort the question. Decisions made in this state are unreliable.

The research from William Doherty's work on divorce ambivalence is useful here. In a 2011 study of divorcing couples in Minnesota, roughly 30% of partners actively in divorce proceedings reported real interest in attempting reconciliation. The decision and the wanting were not the same thing, even after legal proceedings had begun. The takeaway is not "stay no matter what." It's that the decision is not as clean as it feels in the immediate aftermath of a crisis.

The right work in the first 30 to 90 days after a crisis:

  • Get separation of any kind that allows your nervous system to settle. This may be physical separation, may be sleeping in different rooms, may be an explicit pause from major shared decision-making. The point is to reduce the volume of the immediate stimulus so processing can happen.
  • Get individual support. Not couples therapy yet. Individual therapy, a trusted friend, a peer support group. Both partners, separately. Couples-mode in the first weeks after a crisis often re-traumatizes rather than helps.
  • Set a date, weeks or months out, when you'll revisit the question with a clearer head. This is not procrastination. It's deliberate.
  • Hold the question open. Don't make irreversible moves (lawyer retainers, public announcements, asset division) before the deliberate revisit. Don't make reconciliation promises either.

After the crisis processing has happened, the question typically migrates into one of the other three situations: slow drift (often with new awareness from the crisis), one-sided wanting-out (the crisis revealed an asymmetry that was already there), or hard-line dealbreaker (the crisis was actually a pattern, not an event). Then the work in those sections applies.

If you're in this situation, the next step in this article is not to read further today. It's to get individual support and put the decision down for 30 days. Bookmark this. Come back when the static has cleared.

Situation 2: Slow drift (nothing's wrong, which is the wrong)

This is the most common situation, and it's where the framework actually does its work. You're here if: you can't point to a specific event that broke things; the texture of the relationship has thinned; you fight less than you used to but feel less close; sex is rare or absent; you describe yourselves to friends as "fine"; the question of whether you should leave returns periodically, gets put down, returns again; you both agree that things aren't right but neither of you can name what would fix it.

Three things to know about slow drift before any framework is useful.

First, drift is recoverable more often than you think. Couples in long-term drift consistently underestimate what's reachable. Research on the U-shaped marriage curve (the dip in satisfaction that hits most marriages between years 5 and 15, especially for couples with young kids) finds that satisfaction recovers in a substantial share of couples who stay. The story you might be telling yourself, that you've simply grown apart and there's nothing to come back to, is sometimes true and often isn't.

Second, drift is also recoverable only with specific work, not with time. Time doesn't heal drift. Couples who do nothing different remain in drift, often for many years, until the question of leaving forces a decision. The research is clear on this: the couples who recover from drift are the ones who name it and act on it; the couples who hope it passes don't recover.

Third, "trying" without specificity is the most common failure pattern. "We've tried" is the phrase that comes up most often in failed reconciliation. Probed, "trying" usually means: a few date nights, two or three uncomfortable conversations, one round of couples therapy that didn't click, then back to the same texture. That isn't trying in the meaningful sense. The trying that works is dated, named, witnessed, and sustained over months, not weeks.

A real decision framework for slow drift

Most articles give you a list of 15 questions to ask yourself, mixed in tone, no real shape. Here's a cleaner version: a 2x2 that produces four concrete next moves.

The two axes are the only ones that actually matter:

Axis 1: Is the underlying problem fixable? Some problems are fixable in the strict sense (we don't communicate well, we've stopped having sex, we've lost our friendship). Some are not (we want different things from life, one of us has fundamentally changed in a direction the other can't follow, the value-mismatch has revealed itself as load-bearing).

Axis 2: Is your partner willing to do the work? Not "would they grudgingly agree to therapy if you pushed hard." Willing in the meaningful sense: willing to look at their own role, willing to sustain effort over months, willing to consider that the way they currently relate may need to shift.

The four cells:

Fixable + willing. The relationship is recoverable, and both of you are reachable. The next move is structured reconnection work, sustained over months. Couples therapy if needed; a structured assessment to map what's actually going on; deliberate practice. Most couples in this cell who do the actual work, not the gestural version, end up in a different relationship within 6-12 months. Our reconnection guide is calibrated for this cell.

Fixable + unwilling. The problem is solvable in principle, but your partner won't do the work. This is the hardest cell, because it doesn't end on its own. The work here is not on the relationship; it's on the conversation about whether the relationship has both partners' investment. If after a deliberate, named, calm attempt to surface this, your partner remains unwilling, you have your answer. It may take months to get there honestly, but the answer is in the unwillingness, not in further effort.

Unfixable + willing. The relationship has a real underlying mismatch that effort won't fix (different life directions, value misalignment that didn't surface earlier, fundamentally incompatible needs), but both of you are good people who genuinely care about each other. This is sometimes the hardest cell emotionally, because nobody did anything wrong. The work here is often a different shape: a deliberate, kind, slow process of recognizing the mismatch and figuring out how to step away with as much grace as possible. Whetstone's work on managed separation and amicable divorce is the literature for this cell.

Unfixable + unwilling. The relationship has structural issues, and your partner is not engaged. This is the cell where staying does the most damage. The decision is usually already made in the body before the brain catches up; the work is gathering the practical resources (financial, emotional, logistical) to act on it.

The 2x2 doesn't tell you what to do. It tells you which conversation to have next. Most couples spend years trying to apply Cell 1 work (more dates, more reconnection, more therapy) to a Cell 2 or Cell 4 situation. The work doesn't take because it's the wrong work for the cell.

How to actually figure out which cell you're in

Most couples can't tell. The cells require honest answers from both partners about hard questions, and most "should I get a divorce" Googling happens by one partner alone, with no idea what the other actually thinks.

Three concrete moves to find out:

The structured assessment. Both partners independently answer the same set of research-backed questions about the relationship. This surfaces asymmetry that conversation hides. Couples often discover that they differ wildly on which problems they think are fixable, which they think are not, and what they're actually willing to do. Emira's couples assessment is one structured option for this. There are others (Gottman's relationship checkup, PREPARE/ENRICH, OurRelationship.com). The point is using something structured rather than asking each other open-ended questions in your kitchen, which produces guarded answers.

The dated, named conversation. Pick a day. Tell your partner: "I want to have a conversation about us. Not a fight. A real one. Can we sit down on Saturday morning and talk about where we actually are?" Then have it. The shape of that conversation: each of you names what you think is and isn't fixable; what each of you would actually need to feel reachable again; what you've each been thinking and not saying. This is a fundamentally different conversation than the one most couples have, and it requires both partners to actually engage. If your partner refuses or deflects, that itself is data: you're learning something about Axis 2.

A 90-day deliberate effort. If both of you are unsure but willing to find out, commit to 90 days of deliberate work. Pick three or four foundational practices (a weekly us-conversation, a daily real question, restored physical affection baseline, no-phones evenings). Do them, dated, witnessed, for 90 days. At the end, the question of whether the relationship is recoverable is much easier to answer because you have data. Did things shift? Did your partner stay engaged? Did the texture of the relationship change at all? 90 days of real work answers questions that 90 months of unstructured drift can't.

Our How to Know If You Should Break Up goes deeper on this 90-day frame and what specifically counts as a real attempt versus a gestural one.

Situation 3: One-sided wanting-out (the asymmetry question)

This is the situation almost nobody writes about honestly. The reason: it's uncomfortable, and most relationship writing is published either by the partner who's leaning out (in which case it's a self-justification) or by the partner who's leaning in (in which case it's a plea for the other partner to come back).

You're here if: one of you (probably you, since you're the one searching) is much further along in the question than the other. The other partner doesn't know how serious you're getting, or knows and is hoping it passes, or knows and is working hard to repair while you're already mostly gone. The relationship looks roughly normal from the outside; inside, you're carrying a private ledger they can't see.

Some hard truths about this situation:

The asymmetry is the actual problem, not the source of it. Most one-sided wanting-out situations have a backstory: years of bids you made that weren't met, conversations that didn't go anywhere, requests for change that didn't produce change. By the time you reach the asymmetry, you've often already done much of your grief privately, and your partner is just now getting handed the news. They will likely react like it's new information, and from their perspective it is.

Surface it, even though it's hard. The most common pattern in one-sided wanting-out is for the leaning-out partner to never name how serious it is, hope something shifts, find that nothing does (because their partner can't address what they don't know exists), and eventually leave with the leaning-in partner blindsided. This is worse for everyone than the surfaced version. If you are seriously considering leaving, your partner deserves to know that, even if it's painful, even if it changes the dynamic. Surfacing the asymmetry gives the relationship one real chance.

The honest shape of the conversation: "I want to tell you something I've been carrying, and I don't want it to be a fight. I have been seriously thinking about whether we should still be together. I'm not sure. I'm not telling you this as an ultimatum. I'm telling you because you deserve to know where I actually am, and because I think there's a version of us figuring this out together that's better than me carrying it alone or you finding out later. Can we talk about it, slowly, over the next few weeks?"

The reset window. After the asymmetry is surfaced, there's typically a 60-90 day window where the relationship has its best chance. Some couples in this window discover real change is possible (the partner who didn't know steps up in ways they hadn't been willing to before; the asymmetric partner discovers their wanting-out was largely about specific addressable things). Some discover it isn't (the partner can't or won't engage at the level required; the leaving partner's heart was further gone than they thought). Either outcome is information. The window is not a guarantee of staying or leaving; it's a deliberate effort to find out which one is actually true.

Therapy is more often valuable here than in any other situation. A skilled couples therapist can help asymmetry conversations not become destructive. If you can find one, try it.

If you're in this situation, our piece on When to Leave a Relationship walks through the reset window in more depth.

Situation 4: Hard-line dealbreakers

You're here if your relationship involves: ongoing physical, emotional, or psychological abuse; untreated severe addiction that is harming you or your children; repeated betrayal without remorse or change; substantial safety risk of any kind. If you're not sure whether what you're in qualifies, that's common; the line between "we have problems" and "this is harm" is often invisible from inside.

We aren't going to give you a list of signs of abuse, because that's not what this article is for, and there are better resources (the National Domestic Violence Hotline at 1-800-799-7233 or thehotline.org being the primary one). What we will say is: the "should I get a divorce" framework above does not apply to this situation in the same way. The 2x2 isn't the right tool. You don't need to figure out if the problem is fixable in a structured assessment. You need safety, support, and a plan.

Two concrete moves if you're here and unsure:

Talk to a domestic violence advocate, even once. They take calls from people who are not yet sure they're in an abusive situation. They will not push you to leave. They will help you think clearly, often for the first time in a long while, about what's happening and what your real options are. The hotline number above. It's free, confidential, and 24/7.

Work with someone who knows how to safety-plan. Leaving a harmful relationship has its own logistics that don't apply to leaving a non-harmful one. Financial separation, custody, housing, timing. A domestic violence advocate or a therapist who specializes in this work can help you build a plan. The plan does not commit you to using it. It commits you to having it.

If you're in this situation and reading further articles like this one is part of your processing, that's normal. Just know that the "should I" question, in a hard-line dealbreaker situation, is rarely the actual question. The actual question is usually "how" or "when" or "what will it take."

What every situation has in common: the things you need data on

Across all four situations, three categories of data make any decision better. Most of the SERP articles bury these or skip them entirely. They matter.

What divorce actually costs

The financial reality varies enormously, but real ranges that come up consistently in financial planning research:

  • A contested divorce in the US: $15,000-$30,000 per partner on average, with high-conflict cases reaching $50,000-$100,000+
  • Mediated divorce: typically $3,000-$8,000 total
  • DIY divorce in non-contested cases: $500-$2,000

Long-term financial impact: divorced women in particular tend to experience meaningful drops in household income that don't fully recover for years, especially in cases involving career interruption for childcare. The research on this (see the "marriage premium" literature) is clear and not optional to ignore. Divorced men tend to recover income faster but lose net worth substantially. None of this argues for or against divorce; it argues for budgeting honestly when making the decision.

What divorce actually does to children

The honest version: it depends on the marriage you're leaving and the divorce you do.

The Wallerstein/Hetherington/Amato debate is the relevant literature. Wallerstein's 25-year study famously found long-term negative effects on children. Hetherington's longitudinal work found most children of divorce adapt within 2-3 years and are functionally indistinguishable from peers in low-conflict intact marriages. Paul Amato's meta-analytic work threaded the needle: children of divorce do worse on average than children in low-conflict intact marriages, and better on average than children in high-conflict intact marriages.

The practical implication: if your marriage is high-conflict and the conflict is visible to your children, divorce often improves their long-term outcomes. If your marriage is low-conflict but you're personally unhappy, the picture is more complicated. This is useful data, not a verdict.

What the regret picture actually looks like

There's a regret data point that gets cited often: roughly 30-50% of divorced people, in some studies, report wishing they'd tried harder. The number is real but worth contextualizing. The flip-side data is also real: a substantial majority of divorced people, particularly those from high-conflict marriages, report no regret and improved well-being post-divorce. Linda Waite's The Case for Marriage cherry-picks the regret side; less popularly cited research (Amato, Hawkins) finds the regret picture is bimodal, heavy regret for some, none for most.

Daniel Gilbert's work on "future happiness bias" is the right framing here: humans are systematically bad at predicting how we'll feel about a major life decision, in either direction. That's not an argument against deciding. It's an argument for deciding with as much information about your specific situation as possible, rather than relying on generic regret statistics.

What this article won't do (the honest section)

A few things that other articles do, that we won't:

We won't tell you what God thinks. If your decision-making framework is religious, you have better resources for that than this article. If it isn't, you don't need to be talked out of leaving on theological grounds.

We won't tell you that 90% of marriages can be saved if you just try harder. That number isn't real. The number that is real (the 70% effectiveness rate of couples therapy from the Journal of Marital and Family Therapy) applies to couples who are actually engaged in therapy with both partners committed. It doesn't apply universally to all marriages.

We won't moralize against divorce. Some marriages should end. Pretending otherwise causes real harm to people who stay in situations they shouldn't.

We won't moralize against staying. Some marriages that look from the outside like they should end can recover into something durable and good. Pretending otherwise causes real harm too.

We won't pretend our paid product replaces this decision. Emira is a structured couples assessment, not a substitute for therapy or a real conversation. We mention it where it's actually useful (in Situation 2, where a shared structured read of where the relationship stands is genuinely valuable), and don't pretend it solves the question.

Where to go from here

Based on the four situations:

If you're in acute crisis: Stop reading articles for now. Get individual support. Set a 30-90 day window before any major decisions.

If you're in slow drift: The 2x2 above is the actual framework. The next moves are the structured assessment, the dated conversation, and the 90-day deliberate effort. Our reconnection pillar covers the reconnection work in depth; How to Know If You Should Break Up covers the decision frame; Signs Your Marriage Is Over covers when the data points the other direction.

If you're in one-sided wanting-out: Surface the asymmetry. Use the script above. Try a 60-90 day reset window, ideally with a therapist. When to Leave a Relationship walks through this.

If you're in hard-line dealbreaker territory: Call the hotline. Talk to someone who knows how to safety-plan. The framework in this article is not your tool.

If you don't know which situation you're in: The 2-minute disconnection quiz is a fast self-read, calibrated against the patterns we see most often.

FAQ

What's the success rate of marriage counseling?

Approximately 70%, per the Journal of Marital and Family Therapy, but with important caveats. The 70% applies to couples in which both partners are engaged in the work. It applies to specific evidence-based methods (Emotionally Focused Therapy and Behavioral Couples Therapy have the strongest evidence). It does not apply to one-sided therapy, to couples who attended one or two sessions and stopped, or to couples in hard-line dealbreaker situations. We cover this in depth in Does Marriage Counseling Work?.

Should I separate before deciding?

Sometimes useful, sometimes harmful. Useful in acute-crisis situations and in cases where ongoing daily contact is preventing either of you from thinking clearly. Harmful when it's used to avoid the actual conversation or when it functions as a soft exit (one partner uses separation as a stepping stone toward leaving while the other is still trying to repair). If you separate, name the purpose and the timeline explicitly; an undefined separation usually drifts into a divorce.

Is it worse for the kids if we stay or if we leave?

It depends on what staying looks like. The research consensus: high-conflict intact marriages produce worse outcomes for children than divorce. Low-conflict intact marriages, even unhappy ones, often produce better outcomes than divorce. The texture of the marriage your children are watching matters more than the legal status.

What if my partner won't go to therapy or do the work?

That itself is meaningful data on Axis 2 of the 2x2. Try once, calmly, with a specific ask ("I've found a therapist who specializes in this; I'd like us to go four times before deciding whether it's helpful. Will you?"). If the answer is a flat no, repeated, after a real attempt, that's information. Continuing to carry the relationship alone, without your partner's engagement, doesn't fix the underlying issue. Our reconnection guide covers what to do when only one of you wants to do the work.

How long should I try before deciding?

A deliberate, named, 90-day effort is usually long enough to know whether the relationship is responsive. If after 90 days of real work (not gestures), nothing has shifted in either of you, and your partner has been actively engaged the whole time, you have strong data that the underlying issue is not in the recoverable cell. This isn't a magic timeline; it's a structured way to gather information.

Should I tell my partner I'm thinking about divorce?

Almost always yes, with care. The most damaging pattern is for one partner to seriously consider leaving for months or years without telling the other, and then leave abruptly. Surfacing the asymmetry, calmly, gives the relationship its real chance and gives your partner real information. The script in Situation 3 above is one way to do it.

Can a marriage recover from infidelity?

Yes, but the work is specific and usually requires a therapist who specializes in affair recovery (the gold-standard approach is Sue Johnson's EFT for affair healing). Couples who do that work and stay are often surprised by how durable the recovery becomes. Couples who try to "move past it" without doing the structured work usually find the unaddressed material resurfaces over the following years.

What if I just want out and there's no big reason?

Sometimes there isn't a big reason. That doesn't mean the wanting-out isn't real or valid. The work in this case is usually around being honest with yourself and your partner about what's actually happening, which is often that you've changed in directions the relationship can't accommodate. This is closer to Situation 3 (one-sided wanting-out) and the framework in that section applies.

A final note

Most of the people searching this question are not in clear-cut situations. They're in some version of "I don't know, and I've been not-knowing for a long time, and I'm looking for something to help me decide." If that's you, the most important move you can make is moving from not-knowing to knowing. The structured assessment, the dated conversation, the 90-day deliberate effort: these are tools for shifting from not-knowing to knowing. The decision often becomes much clearer once you have actual data on which cell of the 2x2 you're in.

Some marriages end. Some marriages recover into something better than they were. The right question isn't "should I leave" in the abstract. It's "what is my relationship actually pointing toward, and how do I find out."

Read next, calibrated to your situation: