How Attachment Is Actually Measured: ECR, AAI, RQ

You can find out your attachment style in four minutes on a phone. Doing it the way the research does takes an hour of interview, a verbatim transcript, and a coder who has been trained and tested for reliability. Those two things produce results that agree less often than you would expect, and the gap between them is the most interesting fact in this whole area.

There are three families of measure. They ask different questions, of different people, at different ages, and they are not interchangeable.

Three Families, Not One Method

Self-report questionnaires ask you to describe your own experience in close relationships and score the answers. Interview measures ask you to talk about your childhood and score the structure of how you talk. Observational procedures watch behaviour directly, and the famous one watches infants.

Most confusion about attachment measurement comes from treating results from one family as though they were results from another. A great deal of popular writing quotes an infant observation category, applies it to an adult who filled in a questionnaire, and then draws a conclusion about that adult's childhood. Each step of that is a leap.

Self-Report: The ECR and Its Relatives

The Experiences in Close Relationships scales are the workhorse of adult attachment research. You rate a few dozen statements about how you feel and behave in close relationships, and the scoring produces two continuous numbers: one for attachment anxiety, one for attachment avoidance. A later revision refined the items and is widely used alongside the original. What you get out is a position on two dimensions, not a category.

The Relationship Questionnaire works differently. It presents four short paragraphs, each describing a way of relating, and asks which fits you best and then how well each one fits. It is quick, it is still used, and it hands you a category directly, which is convenient and throws away detail.

Behind both sits an older lineage of adult self-report scales measuring things like comfort with closeness and comfort with depending on others. The earliest adult measure of all was a single question offering three paragraphs to choose between, which is worth knowing because almost every internet quiz is a descendant of that format rather than of the ECR.

What self-report can establish: how you currently describe your own experience of close relationships, on two dimensions, in a form that can be given to thousands of people and compared. Nearly everything we know about adult attachment at population scale rests on this method.

What it cannot establish: anything you cannot see about yourself, or will not say. That limitation is not evenly distributed. It bites hardest on avoidance, where the whole pattern involves minimising the importance of attachment, and where the honest answer to do you need people is produced by the same system being measured.

Interview: The Adult Attachment Interview

The Adult Attachment Interview is a structured interview about your childhood relationships. It asks for a handful of adjectives to describe the relationship with each parent, then asks for specific memories that support each adjective, then moves through illness, upset, separation, rejection and loss. It runs to around an hour, and it is deliberately designed to be hard to prepare for.

The scoring is the part that surprises people. It is not the content of the childhood that is coded. It is the structure of the account.

The interview is not asking what happened to you. It is asking whether your account of it holds together when pressed.

So a person who describes a genuinely hard childhood clearly, with detail that matches the adjectives they chose, and without either idealising or becoming lost in it, can be classified as secure. A person who describes an idyllic childhood in glowing general terms and then cannot produce a single supporting memory, or contradicts themselves, is not. The main classifications map onto the familiar names, with a separate designation used where talk about loss or abuse shows lapses in reasoning or discourse.

What it can establish: something about the organisation of a person's memory and thinking about attachment, which is not the same construct as how they rate their current relationships. It also predicts things self-report does not, which is a large part of why it is still used despite the cost.

What it cannot do: be taken online, be self-scored, or be done quickly. Interviews are transcribed verbatim and coded by people who have completed an intensive training and passed a reliability test. Any website offering you an Adult Attachment Interview result from a form is offering you something else with a borrowed name.

Observation: The Strange Situation

This is the procedure most people have heard of, and it is not an adult measure by any stretch. An infant of roughly one year is brought to an unfamiliar room and taken through a short series of episodes involving a stranger entering, the caregiver leaving, and the caregiver returning. The whole thing takes around twenty minutes.

The coding rests almost entirely on the reunions rather than the separations. What matters is what the infant does when the caregiver comes back: whether they seek contact and settle, direct their attention elsewhere and stay muted, or escalate and prove hard to comfort. The three original categories came from exactly that. A fourth was added later by other researchers for infants whose behaviour did not organise into any strategy at all.

What it can establish: a classification of the infant's behaviour with that particular caregiver, on that day. Note the phrasing. The same infant can be classified differently with a different caregiver, which is one of the more robust and least reported findings in the field.

What it cannot establish: anything about you now. It has also been argued over for decades on cultural grounds, since what counts as an ordinary reaction to a brief separation is not identical everywhere, and the procedure was developed in one place.

Why the Measures Disagree, and What That Means

Self-report questionnaires and the Adult Attachment Interview correlate modestly at best. Reviews of the two literatures describe the overlap as small. I am deliberately not quoting you a number, because the figure varies by study and by which classification scheme is used, and a precise-sounding coefficient would imply more settled knowledge than exists.

The most common explanation is that they are measuring different things that share a name. One asks how you consciously appraise your current close relationships. The other examines how your mind handles attachment-related memory under pressure. There is no particular reason those two should coincide, and for some people they conspicuously do not: the person who reports comfortable, low-anxiety relationships and produces an interview transcript that falls apart is a recognised pattern rather than an anomaly.

This matters for how much weight any single result deserves. If two respected instruments disagree about the same person, a four-minute quiz is not going to settle the matter, and treating any of them as a verdict misunderstands what is on offer.

Where a Free Online Quiz Sits, Including Ours

An online quiz is a simplified relative of the self-report family, usually shortened, usually converted into one of four names at the end. It inherits every limitation of self-report and adds two of its own: fewer items, so more noise, and a category output, so less information.

That is not a reason to avoid it. A well-written quiz does something genuinely useful and slightly different from measuring: it gives you back a structured description of your own experience in language you did not have. People recognise themselves, and the recognition is the value. What has not happened is a measurement in the sense that word carries in research.

Held that way, our quiz and every other free one is worth taking. Held as a test result, it produces the two failure modes we see most often: people who dismiss an accurate description because the format is casual, and people who build an identity on a label produced by twenty questions. Two related pages are worth reading alongside it: how accurate these quizzes are, and what to make of it when your result changes between sittings. The theory underneath all of it sits in our attachment theory guide.

Frequently Asked Questions

What is the ECR attachment questionnaire?

The Experiences in Close Relationships scales are the standard self-report measure of adult attachment in research. You rate a few dozen statements about how you feel in close relationships, and scoring produces two continuous scores rather than a category: one for attachment anxiety and one for attachment avoidance. A later revision refined the items and is widely used alongside the original version.

Can I take the Adult Attachment Interview online?

No. The Adult Attachment Interview is administered in person or by video by a trained interviewer, recorded, transcribed verbatim, and coded by someone who has completed intensive training and passed a reliability test. It scores the coherence and structure of how you talk about your childhood rather than the content. Any online form offering an AAI result is using the name for something else.

Is the Strange Situation an adult attachment test?

No. The Strange Situation is a laboratory procedure for infants of roughly one year, lasting about twenty minutes, in which brief separations and reunions with a caregiver are observed and the reunion behaviour is coded. It classifies the infant's behaviour with that specific caregiver. It says nothing directly about adults and cannot be applied to one.

Do attachment questionnaires and interviews give the same result?

Often they do not. Self-report measures and the Adult Attachment Interview correlate only modestly, and reviews describe the overlap as small. The usual explanation is that they measure different things sharing one name: conscious appraisal of current relationships in one case, and the organisation of attachment-related memory under pressure in the other.