What the Research Actually Says About Attachment Styles and Relationships
Attachment style content online is mostly recycled without citation. Here's what a large national survey and a 132-study meta-analysis actually found.
Quick Answer
A large U.S. national survey (the National Comorbidity Survey Replication, over 5,600 adults) found 63.5% of adults classified as securely attached, 22.2% avoidant, 5.5% anxious, and 8.8% unclassified — a wider gap between avoidant and anxious attachment than popular dating content usually implies. Separately, a 2019 meta-analysis of 132 studies found that both attachment anxiety and avoidance correlate with lower relationship satisfaction, and that a person's own attachment insecurity predicts their satisfaction more strongly than their partner's does.
Why Attachment Style Statistics Are Hard to Trust Online
Attachment theory is one of the most-cited frameworks in dating content, and also one of the least carefully sourced. Search for "what percentage of people are securely attached" and most results either cite no source at all or point back to the same few numbers repeated across dozens of SEO-optimized listicles, often without any indication of what study they originally came from or how large the sample was.
That matters because the actual figures vary meaningfully depending on methodology — how attachment is measured, what population is sampled, and whether a fourth "disorganized" or "fearful-avoidant" category is included alongside the classic three. Citing a specific, credible study beats citing a rounded number nobody can trace.
The Actual Distribution of Attachment Styles
The NCS-R-based study found the following breakdown among U.S. adults: 63.5% secure, 22.2% avoidant, 5.5% anxious, and 8.8% unclassified. That's a notably wider gap between avoidant and anxious attachment than most popular dating content implies — a lot of casual writing on the topic treats anxious and avoidant attachment as roughly equally common, when this large-sample data suggests avoidant attachment is close to four times more prevalent than anxious attachment specifically.
This matters practically: it suggests that patterns like the ones described in dating someone avoidant are, per this data, considerably more common to encounter than anxious attachment specifically — which may explain why avoidant dynamics come up so often in dating discourse relative to their assumed prevalence.
Attachment Insecurity Actually Predicts Relationship Satisfaction
Beyond simple prevalence, the more practically useful question is whether attachment style actually predicts relationship outcomes, or whether it's a popular framework without much real evidence behind it. The 2019 meta-analysis — "Insecure attachment and relationship satisfaction: A meta-analysis of actor and partner associations," pooling 132 studies — found a clear answer: both attachment anxiety and attachment avoidance were negatively associated with relationship satisfaction.
That finding has a real practical implication that a lot of attachment-style dating advice misses: it's tempting to focus entirely on screening a partner for secure attachment, but the research suggests your own attachment security matters at least as much, if not more, for how satisfied you end up being. This lines up with the case made in secure attachment in dating for treating security as something to actively build in yourself, not just a trait to look for in someone else.
What This Means for How You Read Attachment Advice
None of this undermines attachment theory as a useful lens for understanding dating patterns — the meta-analytic evidence for a real link between attachment insecurity and relationship satisfaction is solid. But it's worth being skeptical of specific numbers that circulate without a source, and worth remembering that population-level prevalence data (who is more or less common) is a separate question from outcome data (what actually predicts satisfaction) — a lot of casual content blurs the two together in ways the underlying research doesn't support.
Sources:
- Associations between adult attachment style and mental health care utilization: Findings from a large-scale national survey — PubMed
- Insecure attachment and relationship satisfaction: A meta-analysis of actor and partner associations — Personal Relationships, 2019, ScienceDirect
Frequently Asked Questions
What percentage of adults are securely attached?
A large U.S. national survey (NCS-R, over 5,600 adults) found 63.5% classified as securely attached, 22.2% avoidant, 5.5% anxious, and 8.8% unclassified. Other studies report somewhat different splits depending on methodology, but secure attachment consistently comes out as the most common style, generally in the 55–65% range.
Does attachment style actually predict relationship satisfaction?
Yes, according to a 2019 meta-analysis of 132 studies published in Personal Relationships. Both attachment anxiety and attachment avoidance were found to correlate with lower relationship satisfaction, with a person's own attachment insecurity affecting their satisfaction more strongly than their partner's attachment insecurity does.
Is avoidant attachment more common than anxious attachment?
The largest national survey data suggests yes — the NCS-R study found avoidant attachment at 22.2% versus anxious attachment at 5.5%, a substantially wider gap than the roughly even split often assumed in popular dating content.
Does your partner's attachment style affect your relationship satisfaction, or just your own?
Both, but unevenly. The 2019 meta-analysis found that a person's own attachment insecurity (the 'actor effect') predicted their satisfaction more strongly than their partner's insecurity did (the 'partner effect') — meaning your own attachment patterns matter more to your experience of the relationship than your partner's do.
Where do most online attachment style statistics actually come from?
Most circulating figures trace back to a small number of academic studies, often cited without attribution, mixed with unsourced numbers from SEO content. Two of the more credible sources are the National Comorbidity Survey Replication (a large U.S. national sample) and peer-reviewed meta-analyses published in journals like Personal Relationships.
Related reading
What Pew Research Actually Found About Online Dating
Pew Research Center's 2023 survey of U.S. online daters is one of the few large, methodologically solid datasets on the subject. Here's what it actually shows.
How Long NYC Daters Actually Talk Before Meeting Up
Data on time-to-first-date shows most matches either meet within a week or drift into an endless talking stage that never converts. Here's what the numbers say.
What the Data Actually Says About First Message Response Rates
Most advice about opening lines is guesswork. Here's what response-rate data actually shows about what gets a reply — and why the whole premise may be the wrong question.
Ready to find your person?
Tenr is the invite-only 10-minute date app for high achievers.