Methodology

How the 35-question attachment style test works

The test organizes current relationship responses across two dimensions: attachment anxiety and attachment avoidance.

Question structure

The test contains exactly 35 questions. Twenty-eight questions contribute to the two scores: 14 attachment-anxiety items and 14 attachment-avoidance items. Seven additional questions organize current context, preferred sources of security, conflict response, desired growth area, relationship situation, and response consistency.

Answer formats

The test uses five-point agreement scales, scenario choices, sliders, and non-scored context choices. Scenario values and slider conversion remain in the same scoring direction as the original test. Reverse-scored items are converted before averaging.

Two-dimensional scoring

Each dimension is calculated as the arithmetic mean of its 14 scored items on a 1–5 scale. Attachment anxiety reflects sensitivity to uncertainty, rejection, and loss of connection. Attachment avoidance reflects sensitivity to dependence, vulnerability, and loss of autonomy.

Primary result mapping

Secure attachment

Attachment anxiety below 3.2 and attachment avoidance below 3.2.

Anxious attachment

Attachment anxiety at or above 3.2 and attachment avoidance below 3.2.

Avoidant attachment

Attachment anxiety below 3.2 and attachment avoidance at or above 3.2.

Fearful-avoidant attachment

Attachment anxiety at or above 3.2 and attachment avoidance at or above 3.2.

Borderline results

A cut point is a practical way to organize a continuous score. When either score is close to 3.2, a small answer change could produce a neighbouring type. The report therefore presents both dimension scores and the highest-response situations rather than treating the label as absolute.

Consistency questions

Two non-scored sliders revisit the ability to interpret quiet communication calmly and to ask for support. They are compared with the two average scores to describe whether the overall response pattern was highly consistent, broadly consistent, or more situation-dependent. They do not change the type result.

Limitations

The test is a self-report snapshot. Results can be affected by the relationship you had in mind, current stress, recent experiences, interpretation of the wording, and ordinary measurement variation. It is not a validated clinical diagnostic instrument and should not be used to diagnose an attachment disorder or infer another person’s mental state.

Version information

The English localization preserves the existing 35-question order, answer values, reverse scoring, score averages, and 3.2 result thresholds. Only user-visible language and English UX copy are localized.

Methodology version: 2026-08-06. Future calibration should be based on aggregated completion data and qualified review, without changing historical results silently.