Emotional Maturity in a Relationship: What the Evidence Supports

The phrase has no measurement behind it, and that is the first useful fact

Almost every article on this subject is a list of adjectives. Self-aware. Secure. Accountable. Good with feelings. Nobody reads self-aware and thinks not me, which is exactly why the format survives.

The deeper problem is that emotional maturity is not a construct in relationship science. There is no validated scale for it and no longitudinal study that measures it at one point and relationship outcomes at another. What exists is a set of measured behaviours and traits the phrase gets used to gesture at, and they do not all point the same way.

The closest measured trait is emotional stability, the low-neuroticism pole of the five-factor model. A meta-analysis of 19 samples and 3,848 participants found that four of the five factors correlated with the intimate partner's relationship satisfaction: low neuroticism, high agreeableness, high conscientiousness and high extraversion. Those associations did not differ between men and women, or between married and unmarried couples. Notice what that sentence actually says — the personality belongs to one person, the satisfaction being predicted belongs to the other.

The largest project on the question cuts across the trait framing entirely. Forty-three longitudinal datasets, 29 laboratories, 11,196 couples, analysed with random forests: the strongest predictors of relationship quality were all relationship-specific — perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, conflict. Relationship-specific variables accounted for up to 45% of the variance at baseline. Individual differences reached 21%, and added nothing at all once a person's own relationship-specific reports were already in the model.

So the honest version of this article is not a personality profile. It is a list of behaviours, each carrying whatever evidence stands behind it, including the places where that evidence is thinner than the genre admits. That standard is the one this site sets out in its editorial policy, and this is the topic where it bites hardest.

Before the list: maturity is the wrong frame for a relationship where someone is being controlled

Everything below describes ordinary variation between two people who are both free to leave, to disagree and to be wrong.

Monitoring a partner's phone, messages or location; controlling money or transport; isolating someone from friends and family; threats; making one person permanently responsible for managing the other's anger — none of those are low scores on a maturity scale. In England and Wales they are set out in the Home Office statutory guidance on controlling or coercive behaviour as a pattern of behaviour that makes a person subordinate and dependent. Reading them as immaturity, as something a better-phrased request would fix, is one of the reliable ways people stay in those situations far longer than they should.

If that is the situation, the next step is a domestic abuse service rather than a communication technique. In the UK, Refuge runs the National Domestic Abuse Helpline on 0808 2000 247. In the US, the National Domestic Violence Hotline is 1-800-799-7233. Elsewhere, Find A Helpline lists services by country.

This section sits at the top rather than the bottom because the people who most need it are the least likely to reach the end of a long article, and because everything after it assumes a relationship in which both people can actually act. The narrower case — a feeling that is ordinary attached to behaviours that are not — is worked through in the psychology of jealousy, which draws the same line in more detail.

The most repeated claim in this genre did not replicate

If you have read anything at all about emotional maturity in relationships, you have met this finding: the husband who accepts his wife's influence is the one whose marriage survives.

It comes from Gottman, Coan, Carrere and Swanson's 1998 study of 130 newlywed couples followed for six years. The paper tested seven process models — anger as a dangerous emotion, active listening, negative affect reciprocity, negative start-up by the wife, de-escalation, positive affect, physiological soothing of the male — and reported divorce and stability predicted with 83% accuracy and satisfaction with 80%. The husband's rejection of his wife's influence was among the variables separating the groups.

Two things happened afterwards that almost never get repeated alongside it.

First, a methodological objection. Heyman and Smith Slep argued in 2001 that divorce-prediction equations were being reported without crossvalidation — tested, that is, on the same data used to build them. They demonstrated the cost with 528 people from the 1985 National Family Violence Survey, split into halves. On the half the equation was built from: 90% overall accuracy, 92% sensitivity, 89% specificity. On the fresh half, sensitivity fell by 45% and specificity by 15%. Their conclusion was that results without crossvalidation should be read with extreme caution no matter how impressive they look.

Second, and more directly: somebody tested the models in a new sample. Kim, Capaldi and Crosby used a community sample of 85 married or cohabiting couples with at-risk backgrounds, measuring affective processes at around age 21 and relationship status and satisfaction about two and a half years later. The major findings failed to replicate. Men's rejection of their partners' influence, the absence of men's de-escalation of partner negative affect, and women's negative start-up did not predict relationship status. The same study found the affective processes looked different depending on whether the couple was discussing his chosen topic or hers.

None of that says being influenceable is unimportant. It says the confident version — this one behaviour predicts divorce — is doing work the evidence will not carry, in the same way the three-day rule and the 3-6-9 rule do with numbers nobody ever measured.

What being influenceable looks like when somebody actually measures it

One study gets much closer to the behaviour, and it is more useful precisely because it claims less.

Overall, Simpson and Struthers video-recorded 180 couples having two discussions in which one partner wanted the other to change something. Afterwards both people reviewed the recording and reported, at the end of every 30 seconds, how angry they had been during that stretch. Independent coders rated two things over the same intervals: how much the person being asked to change withdrew, and how much the person doing the asking softened — meaning they were sensitive to the other's need for autonomy and conveyed that the other person was valued.

The headline result is unsurprising: people high in attachment avoidance got angrier and withdrew more when they were the target of a request, and those reactions made the discussion less successful. The useful result is the within-person one. The same people's anger and withdrawal dropped in the intervals where their partner softened. Comparing across couples, avoidant people whose partners softened more felt less anger, and their discussions went better.

Two things follow that no list of adjectives can give you.

One: being influenceable is not a fixed property of the person being asked. It moves inside a single conversation, in half-minute increments, in response to how the request is arriving.

Two: the active ingredient is specific. Not warmth in general, and not softening the request into vagueness. Sensitivity to the other person's autonomy, plus conveying that they are valued, while still asking for the change. That is neither of the two things people normally do, which are pressing harder and dropping it.

The retrospective self-test still works and is still uncomfortable: name a position you held six months ago that changed because of something your partner said. Not a preference you gave way on to end an argument — a view you now hold differently. If nothing comes to mind, the traffic in that relationship is running one way, and the pattern in the next section is how it usually gets established.

Demand and withdraw: the pattern, its real size, and who does which half

The pattern the observational literature keeps finding is one person pressing for change while the other retreats, with the pressing escalating in response to the retreat.

The size of it is known. A meta-analysis pooled 74 studies and 14,255 participants and found a moderate association between demand/withdraw and poorer outcomes overall, around r = .36, rising to roughly r = .42 for relational outcomes specifically. That is one of the larger effects in this literature, and it is not specific to one direction of the pattern.

Who does which half is messier than folklore suggests, and the two best sources disagree in a way worth seeing. The meta-analysis found wife-demand/husband-withdraw more frequent overall, while its lead author was explicit that it was not more or less damaging than the reverse. A diary study that took the question into people's homes found something different again: across 116 US couples reporting conflicts as they happened, husband-demand/wife-withdraw and wife-demand/husband-withdraw occurred at equal frequency. Both were more likely when the topic was the relationship itself, and in both, the person who raised the issue tended to be the one doing the demanding.

That last detail is the one to keep. The role follows the topic and whoever opened it, not the gender of the person occupying it. Which means describing your partner as a withdrawer, or yourself as the one who always has to bring things up, is usually a description of which subjects each of you raises.

A more recent dyadic study of 263 couples over a year found attachment insecurities linked to lower satisfaction in both partners indirectly, through greater use of demand/withdraw — and attachment avoidance specifically through mutual withdrawal, where both people disengage and nothing gets raised at all.

If you want a read on which half you occupy, our conflict style test scores pressure and withdrawal separately, because a good number of people do both inside one argument, and it says plainly on the page what a result cannot tell you. The version of this loop that runs over a phone — faster, and with worse information — is in the texting habits that quietly push people away.

And then the finding that should make anyone writing this kind of article nervous

Here is where the confident version falls apart. It gets its own section because the genre never includes it.

Ross, Karney, Nguyen and Bradbury set out to reconcile two competing results: that demands met with withdrawal predict declining satisfaction, and that demands also give partners a way to voice concerns while withdrawal de-escalates conflict. Their argument was that withdrawal is damaging among couples who have the social, educational and economic capital to actually act on what is being asked, and constructive among couples who do not.

Study 1 followed 515 couples with an 18-month follow-up and replicated the harmful effect of observed demand/withdraw on wives' satisfaction among relatively affluent couples — while documenting benefits of the same pattern among relatively disadvantaged ones. Study 2 followed 431 couples across four waves at 9, 18 and 27 months, and socioeconomic risk moderated the association again, with higher withdrawal proving beneficial where socioeconomic risk was high.

The authors' own conclusion is the sentence to carry away: efforts to change couple communication without appreciating the larger social and economic contexts of those behaviours may be counterproductive.

Read that against the whole advice industry, this article included. The same observable behaviour appears to do different things depending on whether the couple can realistically deliver what is being asked. Someone who goes quiet when their partner asks for something they cannot provide is not necessarily avoiding the partner. They may be declining to have a fight about an outcome neither person controls.

What survives is narrower than the usual conclusion, and it is the actual skill. Withdrawal from a request you could meet is avoidance. Withdrawal from a request you genuinely cannot meet may be the least destructive move available. Telling those two apart honestly is a question about the request, not about your temperament.

Negative behaviour is not uniformly bad either, and that has been tested twice

The same contingency shows up in a second place, which is what makes it hard to dismiss as one odd result.

McNulty and Russell ran two longitudinal studies of newlyweds and asked how direct negative problem-solving behaviour — blaming, commanding, rejecting — related to changes in satisfaction. The answer depended on the severity of the problems the couple actually faced. Among couples facing only minor problems, those behaviours predicted steeper declines in the person's own satisfaction. Among couples facing more severe problems, the same behaviours predicted more stable satisfaction. Changes in the severity of the problems themselves mediated the effect: the blunt conversation was associated with the problem getting smaller.

One behaviour did not flip. Indirect negativity — the sighing, the pointed silence, the shortness that denies being shortness — was associated with stably lower satisfaction regardless of how serious the problems were. If any single behaviour in this article deserves to be called immature without qualification, it is that one, and the reason is that it puts the other person in the position of having to guess, after which guessing wrong is held against them.

A review by Overall and McNulty pulls the pattern together across the literature. Direct opposition is beneficial when serious problems need addressing and the partner is able to change, and harmful when the partner is not confident or secure enough to respond. Cooperative communication built on affection and validation can be harmful when serious problems need to change, and beneficial when problems are minor, cannot be changed, or when the partner's defensiveness is what is blocking progress.

So there is no communication behaviour that is mature in the abstract. There is a match or a mismatch between the behaviour, the size of the problem, and whether the other person can act. That is a considerably harder thing to sell than seven habits, and it is the closest thing to a real answer this literature has produced.

The finding that undercuts the premise of the whole genre

Everything so far assumes the causal arrow runs from communication to satisfaction — behave better, and the relationship improves. That assumption has now been tested directly, and it did not survive intact.

A team led by Matthew Johnson examined three longitudinal studies and asked the question at the within-couple level. Not: do couples who communicate worse report worse relationships? That much is thoroughly established and mostly uninformative. The question was: for a given couple, does communication getting worse predict their satisfaction getting worse later?

There were some lagged links from negative communication to later changes in satisfaction, and some running the other way. But the robust finding was concurrent: at times when a couple had less negative communication than usual for them, they were also more satisfied than usual for them. Positive communication was rarely associated with satisfaction at the within-person level at all.

The authors' framing is deliberately deflating. Within-person changes in negative communication primarily covary with satisfaction rather than predicting it.

That is not a reason to stop trying. It is a reason to be honest about what a behaviour change is likely to be. The way couples talk to each other and how they feel about each other move together, and pulling them apart to say which one is driving has proved much harder than three decades of advice writing assumed. It is entirely plausible that a good week produces gentler conversations rather than the other way round, and this design cannot rule that out.

It is the same problem that turns up whenever a behaviour gets read as a cause — the measurement issue behind what body language actually predicts is the identical shape, and so is the reason the numbers that circulate about modern dating rarely survive contact with their sources.

Regulating your own emotion: what was found, and for whom

One of the few studies to track emotion regulation and marital satisfaction over a long stretch produced a result more specific than the advice it gets used to support.

Bloch, Haase and Levenson used a 13-year, three-wave study of middle-aged (40 to 50) and older (60 to 70) long-married couples. During a conflict discussion at the first wave they measured downregulation of negative emotion — how quickly each spouse's signs of negative emotion subsided after a negative event, in reported experience, in visible behaviour and in physiological arousal. That is a much better measurement than asking people whether they are good at calming down.

What they found: greater downregulation of wives' negative experience and behaviour predicted greater marital satisfaction for both spouses at the time, and greater downregulation of wives' negative behaviour predicted increases in wives' satisfaction over the following years. Wives' constructive communication between the first two waves mediated the longitudinal effect.

The asymmetry is in the paper's subtitle — more than a wives' tale — and it is worth sitting with rather than smoothing over. The effects ran through the wives' regulation. There was no equivalent finding for husbands' downregulation. The authors do not claim men's regulation is irrelevant, and a study of long-married heterosexual couples in two specific age bands in one country is not a general law of relationships. But it is a caution against the symmetrical version of the advice, and a reminder that most of what gets said about calming down in an argument is untested folklore rather than a measured effect.

What generalises better is the unit being measured: how fast the arousal comes back down, not whether it went up. Going up is not the variable. Everyone's goes up. The observable difference is whether twenty minutes later the conversation is still being fought, which is the same variable running through the repair research in how to reconnect after a fight.

Saying what is happening in you, and what it costs to hold it in

The mature version of this in ordinary life is unglamorous: being able to say I am anxious and it is not about you, or that landed badly and I do not entirely know why yet.

What makes it mature rather than merely expressive is that nothing is attached to it. Naming a state so the other person has the information is a different act from naming it so they will do something about it. The second one is a request wearing a disclosure's clothes, and most people can feel the difference from the receiving end even when they cannot name it.

The cost of the alternative has been measured, in a study that looked at suppression during sacrifice from three angles. In the lab, dating couples discussed real sacrifices they had made; suppressing emotions carried costs for the partner doing the listening, not only for the person suppressing. In a 14-day daily experience study, within-person increases in suppression during daily sacrifice were followed by decreases in emotional wellbeing and relationship quality as reported by both members of the couple. And suppression predicted decreases in satisfaction and increases in thoughts about breaking up three months later. Authenticity mediated the effect: the cost was not the hidden feeling, it was the sense of not being oneself in the interaction.

That last point is what stops this being an argument for saying everything. Suppression in that research means holding down an emotional response while it is happening. It is not the same as deciding a passing irritation does not need airing.

There is also a measured distinction between the two examples at the top of this section and an ordinary account of a day. In the diary work that separated them, disclosing a feeling predicted how intimate an interaction felt, and disclosing a fact predicted nothing at all — those numbers, and what they ask of the person listening, are in how to get someone to open up.

Repair, and the part nobody mentions: accepting one

A repair attempt is anything that interrupts an escalation before it completes — a joke, a hand on the arm, a change of subject, an apology arriving before either person has finished being angry.

The part people underestimate is the receiving end. A repair only works if it is accepted, and refusing one is a choice that is usually invisible to the person making it. There is a specific move worth learning to notice in yourself: rejecting an olive branch because accepting it would end the argument before you have won it.

Maturity here is symmetrical and uncomfortable in both directions. Offering repair while you are still angry is hard. Accepting one while you are still right is harder. The mechanics of doing this after a real fight rather than a minor one are in what the repair research supports, including what to do in the hour afterwards.

This is also the point at which the obvious extension — forgive more — stops being supported, which is worth stating plainly because forgiveness is the single most over-recommended item in this genre.

James McNulty ran a seven-day diary study with newlywed spouses and found that partners were more likely to have engaged in a negative behaviour on the days after they had been forgiven than on days after they had not. In a separate longitudinal study of newlyweds over four years, spouses who reported being relatively more forgiving experienced psychological and physical aggression that stayed stable, while spouses who were relatively less forgiving saw declines in both.

The mechanism the author proposes is unsentimental: expressing forgiveness removes consequences that would otherwise discourage the behaviour. Read this carefully, because it is not an argument for holding grudges. It is an argument that repair and accountability are separate things, and that a relationship which supplies the first without ever supplying the second has a name and it is not maturity. It also marks the point where this section connects back to the second one: a pattern of aggression that does not decline over four years is not a communication problem.

Tolerating being the object of someone else's displeasure

This one predicts a great deal and gets discussed rarely, partly because it looks like nothing from the outside.

When someone cannot bear being the object of another person's displeasure, everything downstream distorts. They over-apologise to end the discomfort rather than to acknowledge anything. They concede positions they will resent conceding. They rush to resolution before the issue has been described, which guarantees it comes back.

Attachment research gives a reasonable account of where this comes from. For people high in attachment anxiety, a partner's withdrawal registers as threat, and the behaviour that follows is aimed at restoring closeness rather than at the substance of the disagreement. The dyadic version of this — what a partner can do about it — is set out in the buffering model developed by Simpson and Overall, where the partner's regulation of the insecure person's reaction is treated as part of the system rather than as their problem alone.

The mature version is the ability to sit in someone's annoyance for a day without either capitulating or retaliating. It is invisible, which is why it is almost never on lists like this one, and why the people who do it well rarely get credit for it.

Two warnings, because this is the section most easily misread. Tolerating displeasure is not tolerating contempt, and it is not tolerating being punished with silence for days — withdrawal used as a sanction is one of the behaviours named in the second section of this article, not a test of your composure. And the capacity is not a fixed trait: the same person handles it well in a relationship where they feel secure and badly in one where they do not.

Where the anxiety is coming from the inside rather than from anything the other person is doing, the difference between love and anxious attachment works through the distinction, and our attachment style test gives a rough reading in about two minutes.

Responding when something goes right, not only when something goes wrong

Almost all of the above is about conflict, because that is where the research money went. The best-supported everyday behaviour in this article is not about conflict at all.

Gable, Reis, Impett and Asher studied what happens when someone shares good news — a promotion, a small win, an idea they are excited about. What mattered was the response. Enthusiastic engagement with the good thing was associated with relationship wellbeing over and above how the person responded to bad news. Passive or deflating responses to good news were associated with worse outcomes, again independently of support during difficulty.

That finding has an unusual practical shape. Most advice about being a better partner concerns the hard moments, which are rare, unpleasant and hard to rehearse. This one concerns the ordinary ones, which are frequent and low-stakes, and where the failure mode is not cruelty but distraction — the that's nice delivered to a screen.

It also lines up with the largest dataset in the field. In the 11,196-couple project, appreciation was among the five strongest relationship-specific predictors of relationship quality, alongside perceived partner commitment and perceived partner satisfaction. Three of those five are about what one person believes the other feels — which means a considerable share of what predicts relationship quality is not what you do but whether it registered.

There is a measured reason to worry about whether it registers. In a 14-day diary study of 67 cohabiting couples producing 1,768 reports, on the days one partner said they had done something thoughtful, the other partner recorded it 51.2% of the time. Roughly half of what one person does is never seen by the person it was for. The version of this that holds up over years, rather than over a fortnight, is in the evidence that appreciation only counts once it arrives, and the version that explains why the same relationship can feel different at month three and month thirty is in what builds attraction over time.

How someone tells the story of their past, and how much that is worth

This is the item that costs nothing to observe, which is why it has been repeated so often and examined so rarely.

The folk version: listen to how someone describes previous partners, and listen for whether any account includes a part they played. A history in which every former partner was crazy, cold or dishonest and the narrator was uniformly reasonable is a story with a missing variable. The inverse — someone who takes all the blame for everything — is informative too, usually about self-punishment rather than accountability.

That is sensible, and it is an extrapolation. What has actually been measured is narrower and more interesting.

The Adult Attachment Interview does not ask about romantic partners at all. It asks about childhood relationships with caregivers, and what gets scored is not what happened but how coherently the person can talk about it. A recent preregistered analysis used transcripts from the NICHD Study of Early Child Care and Youth Development, coded for secure base script knowledge at around age 18, and followed 479 partnered participants to around age 30. Higher script knowledge at 18 predicted better dyadic adjustment in adulthood at r = 0.17, and the association held controlling for other AAI-derived measures as well as sociodemographic and cognitive covariates.

Two honest readings of that number. It is real, it is preregistered, and it survives controls — which is more than most claims in this genre manage. And r = 0.17 is small: it accounts for something like 3% of the variation in how well those relationships were going. The sample was also 82% White and non-Hispanic, which limits how far it travels.

So the way somebody narrates their past does carry information about their relationships a decade later. It carries much less than the confident version implies, and the thing being scored is coherence, not whether they were fair to their exes. Treat it as a weak prior you keep updating, not a verdict delivered on a second date — and bear in mind that the study which checked those narratives against 23 years of records found the darkest childhood stories were often not the ones with the documented difficult childhoods, which is covered in what is known about attachment styles changing. The related question of whether a person has finished with a previous relationship is a different measurement problem, and it is worked through in the signs somebody has actually moved on.

Having a life that is not only this

The self-expansion model describes close relationships as a process in which each person's sense of self grows to include the other. The literature review of this work is explicit that the inclusion is two-way: it is not a merger in which one self dissolves into the other.

In practice, this looks like having a life that continues to exist. Friendships that are not joint. Work that matters independently. Some interest that has nothing to do with the relationship and does not need the other person's approval to be worth doing.

The reason it belongs in an article about maturity rather than one about independence is what its absence does. When one person's entire emotional world runs through the relationship, ordinary friction stops being ordinary. A disagreement about a weekend becomes a referendum, and the other person finds themselves managing not a disagreement but a threat. Nothing about that requires either person to be badly behaved; it is a structural consequence of there being nowhere else for the weight to go.

This is also where the article has to be careful, because the failure mode of the idea resembles something much worse. A contracting world can be a drift that two people let happen. It can also be something one person has been doing to another on purpose, and the guidance referenced in the second section names exactly that: isolating someone from friends, family and sources of support. The test is whose choice it was, and whether it can be reversed without a fight.

What happens to attraction when both worlds contract to the shared one is covered in what builds attraction over time, and the mirror image — what separation does and does not do — in whether distance makes you like someone more.

Why be more mature is weak advice, stated with the actual numbers

It is worth being blunt about the size of what is on offer here, because the genre is built on implying it is enormous.

In the 11,196-couple machine learning project, relationship-specific self-reports predicted up to 45% of the variance in relationship quality at baseline and up to 18% at the end of each study. But change in relationship quality — whether a relationship got better or worse over the course of a study — was largely unpredictable from any combination of self-report variables the field currently uses. The strongest predictors in relationship science tell you where a relationship is. They tell you very little about where it is going.

The downstream effects are real and small too. A meta-analysis of 126 published articles spanning 50 years and more than 72,000 individuals put the association between marital quality and physical health between r = .07 and r = .21, with mortality risk at r = .11 and cardiovascular reactivity during marital conflict at r = -.13. The authors' own comparison is the useful one: those effects are similar in magnitude to associations between health behaviours such as diet and health outcomes. Some of the clinical outcomes showed signs of publication bias.

That is not nothing. It is also not the transformation the adjective lists promise, and knowing the difference is part of what this article is for.

Where that leaves a reader is with two honest moves. Watch what is actually happening in you around this particular person, rather than auditing them against a type — that is the argument in why smart people keep dating the wrong people. And if you want a structured read on where your own relationship currently sits, our relationship satisfaction test scores several dimensions separately and says on the page what it cannot tell you. Neither of those is a diagnosis.

Does couples therapy fix any of this? What the evidence supports as of 2026

The usual place these articles land is: consider couples therapy. It is decent advice, and the evidence behind it is weaker than most people assume, so it should be given with the caveat attached.

A 2026 systematic review replicated and updated the standard evaluation of empirically supported treatments for couple relationship distress, applying the Chambless criteria that the American Psychological Association uses to classify them. Across 40 eligible studies, the finding was that — in contrast to the earlier 1998 evaluation — no couple therapy model met the standards for strong empirical support. The most frequent limitations were insufficient sample size and a lack of independent replication.

Seven models demonstrated modest support: behavioural, cognitive-behavioural, emotionally focused, Gottman, integrative behavioural, insight-oriented and strategic couple therapies. The authors are careful about what that does and does not mean, and note that the Chambless criteria themselves have limitations, recommending the newer Tolin criteria, which synthesise a broader evidence base through meta-analysis.

What this does not say is that couples therapy does not work. Modest support is support, seven models have it, and the review is an argument for larger and better-replicated trials rather than against the practice. What it does say is that anyone citing a specific branded method as the evidence-based one is overstating the case. That includes Emotionally Focused Therapy, which is the framework the previous version of this article leaned on — via a link to its own institute rather than to a study, which was a real failure of sourcing. It is corrected here and logged in our corrections page.

The practical version: the choice of model probably matters less than whether both people show up and whether the therapist is any good, which is an unsatisfying conclusion that happens to be what the evidence currently supports.

What to actually watch for, over a fortnight

Pulling the evidence into something usable, with the confidence levels attached rather than sanded off.

Reasonably well supported. Whether indirect negativity is the default — the sighing, the pointed silence, the shortness that denies being shortness. That was the one behaviour in the contextual study that was bad regardless of how severe the problems were. Whether good news gets met with actual engagement, which has evidence behind it independent of how bad news gets handled. Whether requests arrive in a form that can be acted on rather than only defended against, which is the softening variable measured in half-minute intervals.

Worth watching, weakly supported. How someone narrates their past, which predicts adult relationship functioning at about r = 0.17 — real, small, and about coherence rather than fairness. Whether the relationship has one direction of traffic on questions that matter.

Not supported the way you have been told. That accepting influence predicts divorce; the original finding did not replicate in a new sample. That withdrawal is always destructive; it depends on whether the request can be met. That better communication causes better relationships; within couples, the two mostly move together.

Not a maturity question at all. Monitoring, financial control, isolation, threats, or a pattern of aggression that does not decline. That is the second section of this article, and it does not respond to any of the first.

One framing note to end on. Nearly everyone behaves more maturely in a relationship where they feel secure and less so in one where they do not, which means part of what looks like a person's character is a description of the situation they are standing in. That cuts both ways, and it is worth knowing before concluding anything permanent about anybody — including yourself. The early-relationship version of the same question, where there is less to go on and more temptation to over-read, is in the signs a new relationship has real potential.