Attachment Style Statistics 2026: 30+ Key Stats on Anxious, Avoidant & Secure Attachment

Most attachment statistics online trace back to small, one-time surveys. The numbers below come from a different kind of dataset: Attached's State of Attachment 2026 — repeated, journaling-based measurements of the same people over months. Each stat is on its own line so it's easy to quote; if you cite one, please link the research page or the full whitepaper (PDF).

Method in one line: journal entries scored on the ECR-R attachment scale (AI-rated, three passes averaged, rescaled 0–1), analyzed in aggregate; observational, from a help-seeking community — not population prevalence, not causal claims.

The dataset

• 11,793 people were analyzed in the State of Attachment 2026 dataset.
• 51,676 attachment measurements were taken across those people — repeated measures, not one-off surveys.
• Members who stick with journaling write about twice a week on average.
• The community is help-seeking, mostly English-speaking, and mostly adult — a portrait of people actively working on attachment, not the general public.

How common each attachment pattern is (help-seeking community)

• 90.4% of members score above the midpoint on attachment anxiety.
• 43.1% score above the midpoint on attachment avoidance.
• Average attachment anxiety is 0.74 on a 0–1 scale — well above the 0.5 midpoint.
• At the standard 0.5 cutoff, 51.3% present as anxious.
• 39.1% present as fearful-avoidant (disorganized).
• Only 5.6% present as secure — the mirror image of general-population studies, where secure is the most common style.
• 4.0% present as avoidant (dismissive).
• Where you draw the cutoff matters: moving it from 0.4 to 0.6 shifts the "anxious" share from 38.7% to 57.9% — style labels are a rough map over two continuous dimensions.

Can attachment change? The 2026 numbers

• Attachment anxiety fell 0.15 on average (95% CI −0.166 to −0.139) between early and late measurements among consistent journalers.
• 38% of consistent journalers showed a statistically reliable decrease in attachment anxiety (Reliable Change Index).
• 21% moved from an insecure to a secure pattern over the measurement period.
• The change is dose-responsive — more practice, bigger drop:
• 2–4 journaling sessions: −0.070 average anxiety change.
• 5–9 sessions: −0.144.
• 10–19 sessions: −0.189.
• 20–39 sessions: −0.255.
• 40+ sessions: −0.281.
• Effect sizes reach Cohen's d = 0.82 for the most active group versus the least active.
• Attachment avoidance mostly holds over the same period — anxiety shifts with practice; avoidance is stickier, exactly as attachment theory predicts.
• Important caveat, in the researchers' own words: no control group, heavier users started more anxious, and people often journal most during hard stretches — so this is a real, consistent association, not proof of cause.

Mood and attachment style

• People with a secure pattern report pleasant moods 69% of the time — the highest of any group.
• Avoidant: 59% pleasant.
• Fearful-avoidant: 56% pleasant.
• Anxious: 55% pleasant — secure members report pleasant moods roughly 14 points more often than anxious members.
• Attachment anxiety tracks daily mood distress (r ≈ 0.22–0.25) while avoidance doesn't — the divergence attachment theory predicts.

Self-perception vs. measured pattern

• Only 2% of people describe themselves as secure at intake — measurement later classifies 6% that way.
• 64% self-describe as anxious at intake; journaling-derived measurement puts it at 51%.
• Just 12% self-describe as fearful-avoidant — but 39% measure that way, the largest self-knowledge gap of any style.
• Self-image and measured pattern correlate only moderately: r ≈ 0.36.
• About 18% of people say they're "not sure" of their attachment style at intake. (Take the free attachment style quiz if that's you.)

Attachment by relationship situation

• People going through a breakup show the highest anxiety: 55% anxious, 35% fearful-avoidant, 7% secure.
• Long-term relationships: 52% anxious, 37% fearful-avoidant, 6% secure.
• Dating: 48% anxious, 45% fearful-avoidant, 4% secure.
• Single is the only group where fearful-avoidant (48%) outnumbers anxious (38%).

Measurement reliability

• A single journaling-based measurement is noisy (r ≈ 0.29 reliability).
• Averaging six entries reaches ~0.71 reliability; ten entries reach ~0.80 — every change figure above uses averaged measurements.

Citation: Uribe, G., Goicouria, L., & Cheung, S. (2026). The State of Attachment 2026. Attached / SkyPorch. attachedapp.com/research/state-of-attachment · Whitepaper PDF

Journalists and researchers are welcome to reuse these figures with attribution. For questions about the methodology, contact us.

Explore what the patterns mean: anxious attachment · avoidant attachment · disorganized attachment · secure attachment