Published 2026-08-18
We Analyzed 54 Attachment Assessments From Real Couples — Here's What Stood Out
HalfWay looked at 52 anonymized couple profiles and 54 individual ECR-R assessments. The distribution does not match population baselines—and that gap tells a story about who seeks attachment tools.
People often ask whether attachment style tests reflect the general population. We cannot answer that from our own platform—but we can describe who uses HalfWay and what their results look like. This report summarizes 52 anonymized couple profiles and 54 individual attachment assessments completed on HalfWay through early 2026. It is a usage story, not a demographic census.
The completion funnel: getting both partners in is the hard part
Fifty-two couples created profiles on HalfWay. Fifty-four individuals completed the ECR-R-style assessment—but only 22 couples (42.3%) finished assessments for both partners. The friction is not usually the questionnaire itself; it is getting a partner to opt in. One person often arrives curious or concerned; the other may need time, reassurance that results will not be weaponized, or simply a nudge weeks later. That pattern matches what couples therapists describe anecdotally: the person who initiates growth work is not always the person who completes it with them.
52
Couple profiles created
Starting funnel
54
Individual assessments
ECR-R completed
22
Both partners assessed
42.3% of couples
30
Single-partner only
57.7% of couples
Attachment distribution among completers
Among individuals who finished the assessment, the style breakdown looked quite different from published U.S. population estimates. Anxious attachment was the largest category at 38.9%, followed by secure at 35.2%, dismissive-avoidant at 14.8%, and fearful-avoidant at 11.1%. Mickelson, Kessler, and Shaver (1997), in a nationally representative U.S. sample, reported roughly 59% secure, 25% avoidant, and 11% anxious—the classic three-category distribution before fearful-avoidant was commonly broken out separately in pop psychology.
Side-by-side comparison. Population figures: Mickelson, Kessler & Shaver (1997), nationally representative U.S. sample. Three-category model; fearful-avoidant not broken out separately. HalfWay figures are self-selected completers only (n=54).
| Category | HalfWay (n=54) | U.S. population baseline |
|---|---|---|
| Anxious | 38.9% | 11% |
| Secure | 35.2% | 59% |
| Avoidant (dismissive) | 14.8% | 25%* |
| Fearful-avoidant | 11.1% | — |
The gap is best read as self-selection, not evidence that anxious attachment suddenly became more common. People who seek out an attachment assessment often already suspect or struggle with an anxious pattern—they want language for reassurance-seeking, rumination, or pursue–distance cycles. Securely attached people, by contrast, are less likely to go looking for a label unless a partner asks or a transition (moving in, marriage, parenthood) prompts curiosity. HalfWay’s skew toward anxious results therefore likely reflects why people arrive, not how the whole country attaches.
Who starts the assessment
Among recorded initiators, men accounted for 48.1% (25), women for 44.2% (23), and other gender identities for 7.7% (4). Recipients—the partner invited to complete their side—skewed slightly more female at 53.8%. That asymmetry may reflect who feels motivated to name a pattern first—not a claim about gender and attachment style itself.
HalfWay usage data, early 2026. Describes who sends the invite—not attachment style by gender.
Age and relationship stage
Male initiators averaged 36.5 years, with a bimodal cluster: many between 25–29 and another group between 40–49. Female initiators averaged 31.9 years and skewed younger, with the largest cluster between 18–29. Relationship length was heaviest in the first six months (15 couples)—early-relationship curiosity is the dominant use case on the platform—with a fairly even spread from one to ten or more years after that. New couples often want a map before conflict templates harden; longer-term couples sometimes arrive after a crisis or when considering a major decision.
Location (directional, n=24 with data)
Where self-reported location was available, Europe accounted for roughly 37.5% of profiles, North America about 25%, Oceania about 8.3%, and unclear or other regions about 29.2%. Treat this as directional only: many users skip location fields.
HalfWay usage data, early 2026. Many users skip location fields; treat as directional only.
What we take from this—and what we do not
- Do not use these numbers as population statistics. They describe HalfWay users who completed an assessment.
- The anxious over-representation fits a product built for couples seeking clarity when something feels off—not for confirming that everyone is already secure.
- Partner completion remains the bottleneck. Tools that reduce shame and frame results as shared language—not verdicts—may improve the 42.3% both-partner rate over time.
- Published baselines like Mickelson et al. (1997) remain the right anchor when discussing how common each style is in the general public.
For how ECR-R scoring works and what dimensions mean, read our ECR-R overview. For foundational theory behind the categories, see Bowlby, Ainsworth, and the Strange Situation.
Curious where you and a partner land? Create a couple profile and take the assessment together—or start solo and invite them when the timing feels right.