Ways to be heard
What is anxious attachment? What research says
What research says about anxious attachment: where the idea came from, why it's a scale rather than a type, whether it changes, and what may help it ease.
By the Ember team · Published November 22, 2026 · 6 min read
It's been three hours since you texted. You've reread your message twice, looking for what might have put them off. You know they're probably busy; the feeling doesn't care. Maybe your feed called this "anxious attachment," and you thought: that's me. Is it real? And are you stuck with it?
When researchers combined five long-running US studies, with ratings from age 13 to 72, attachment anxiety eased on average through middle age and later life.
The short answer is that anxious attachment is a real idea with nearly 40 years of research behind it, but it's a matter of degree, not a fixed type. Researchers call it attachment anxiety: how much you worry that someone close will reject or leave you. It's fairly steady, and it still moves, day to day and across decades.
What the research says so far
Cindy Hazan and Phillip Shaver, 1987. Their paper asked whether romantic love is an attachment process, like the bond between infants and parents, and carried three infant styles (secure, avoidant, anxious/ambivalent) over to adults. In two questionnaire studies, the styles appeared in adults in roughly the same proportions as in infants, and the three kinds of adults experienced love differently. They saw a person's style as shaped "in part" by childhood relationships with parents. What it doesn't show: that childhood settles it. The authors flagged measurement problems.
Later studies score attachment anxiety on a scale; one questionnaire item reads, "I'm afraid that this person may abandon me." A second scale, avoidance, measures discomfort with closeness. Low on both is called secure.
University of Illinois at Urbana-Champaign, 2015. R. Chris Fraley and colleagues tested whether adults fall into attachment types or differ by degree. In two samples, the differences between people looked "more consistent with a dimensional rather than a categorical model." What it doesn't show: where any one person sits on the scale.
University of Illinois at Urbana-Champaign, 2011. Fraley and colleagues followed adults in dating or married relationships: 203 answered daily for 30 days, and a second group weekly for 45 weeks. Security rose and fell around a steady center, fitting "a stable factor underlying temporary variations." What it doesn't show: change over years. Participants averaged 21 and 22.5 years old and were mostly secure; the authors note that rarer events, like a breakup or a new baby, may shift people more.
Michigan State University, 2019. William Chopik, Robin Edelstein and Kevin Grimm combined five long-running studies of 628 people born between 1920 and 1967. Attachment anxiety held fairly level through the teens and young adulthood, then declined through middle age and later life, more steeply for men. Being in a relationship went with lower anxiety. What it doesn't show: a big change, or one person's path. The changes were "relatively small," and no one was followed from 13 to 72. Study staff rated people from interviews. Most were white, and most grew up around Berkeley and Oakland.
Purdue University, 2014 and 2018. Ximena Arriaga and colleagues surveyed newly committed couples once a year for two years. At any one time, people who trusted their partner more were lower in anxiety. But what stood out in predicting anxiety easing over time was feeling that a partner encouraged their personal goals. A 2018 review by Arriaga and colleagues proposes why: reassurance, when needed, matters, but alone may not reach the doubt underneath; anxiety should ease most where people gain "greater personal confidence." What it doesn't show: cause. The couples study "does not afford causal inference," its authors write, and about half were university students. The review is a model to be tested.
What the research doesn't say
There's no type you're stuck with: the studies here measure anxiety as a scale, and in each one that followed people, it moved. A list of "signs" on your feed isn't a measure; these researchers used set questionnaires or staff ratings, and nothing here diagnoses anyone. Links aren't causes: being partnered went with lower anxiety, and feeling backed with anxiety easing, but neither study shows one caused the other. And most participants were white; in the shorter studies, average ages were 21 to 25.
Where it shows up
For some it's a text that hasn't come back: whether to text first or wait, what it means when someone suddenly stops replying, or whether no answer means no. For some it's a partner who says everything is fine while pulling away. For some it's being left again and again. And if the pull is toward someone who hurts you, that's something else: what research says about trauma bonds. And for the other side of the pattern, the partner who pulls back when things get close, what is avoidant attachment? looks at what research says.
What can ease it, slowly
- Naming the fear under the checking. The rereading may carry a plainer question: am I too much, are they leaving? Put into words, the question can be looked at, not only felt.
- Asking out loud instead of testing. Going quiet to see if they notice only shows whether they guessed. "I got worried when I didn't hear from you" is something a person can answer.
- Being backed, not only reassured. In the couples study, anxiety eased more over time for people who felt a partner encouraged their own goals. One-off moments that differ only a little from what you expect tend to get absorbed into old expectations, the 2018 review notes; ongoing backing over time can revise them.
- A before and after, from people who've felt it. Some once read every silence as goodbye. Where to find them.
How we built Ember around this
Ember is an AI. It can't tell you your attachment style, or what their silence means.
You can say what's going on, by text or voice, any hour. Ember listens and tries to understand what you mean. It won't tell you what to do, or whether to send the text. Behind the conversation is what real people in similar situations went through, what they tried, and what happened next, so Ember can tell you how it went for people who've been there.
Nothing you tell becomes a public post. No one can read your Ember unless you separately say yes. Ember is an AI; it isn't therapy or an emergency service.
When you need more than a place to think
An AI is not a counselor, a therapist, or an emergency service. If the fear of being left has turned into despair, or into thoughts of harming yourself, you don't have to be sure it's serious enough: you can call or text 988 any time to reach the 988 Suicide & Crisis Lifeline. If you are in immediate danger, call 911. Who to talk to when everyone you get close to leaves goes through more options.
Frequently asked questions
Is anxious attachment real?
It's a long-studied idea: researchers have studied adult love as an attachment bond since at least 1987. The evidence suggests people differ by degree in how much they worry about rejection, rather than falling into fixed types.
Can anxious attachment change?
It moves. In one study it rose and fell day to day around a steady center; in another, it declined on average through middle and later life, though only a little. In a couples study, it eased when people felt a partner backed their goals.
What is an anxious-avoidant relationship?
A pairing of one partner who worries about losing closeness and one who is uncomfortable with it. A 2018 review notes that mismatched closeness goals may strain a relationship over time. None of the studies here tested the pairing directly.
What is Ember?
Ember is an AI you can talk to about what's on your mind, any hour, by voice or text. It listens, remembers what you said, and knows how things went for people who stood where you're standing — what they tried, and what happened after. It won't tell you what to do.
Is Ember private, and does it cost anything?
Ember is private: nothing you tell becomes a public post, and no one can read your Ember unless you separately say yes. Ember is free while it’s in beta. Texting Ember is free, and calls come with 20 free minutes a month; in the US you can buy more, from $4.99 for 35 minutes, and they don’t expire. Ember is for adults 18+.
The fear of being left is real. It's also something that moves.
Talk it out on Ember
Pause or stop anytime · 18+ · Nothing you tell becomes a public post.