If you've been reading about attachment styles for more than an afternoon, you've probably run into the abbreviation "ECR-R" — usually in a sentence like "the gold-standard measure of adult attachment." And if you searched for a way to take it online, you've probably also noticed something confusing: some sites offer "the ECR," some offer "the ECR-R," some offer unnamed attachment quizzes, and almost none of them explain what the R stands for or whether it matters.
It does matter — though probably not in the way you'd expect. This article explains what the ECR-R actually is, what the revision changed and why, what a legitimate online version should look like, and where our own test sits in that landscape. If you want the story of the original ECR from the beginning, our overview of the ECR scale covers that; this piece focuses specifically on the revised instrument and on taking ECR-style tests online.
A Quick Recap: What the ECR-R Revised
The original Experiences in Close Relationships scale was published in 1998 by Kelly Brennan, Catherine Clark, and Phillip Shaver. Its construction was unusually empirical: the authors pooled 323 items from essentially every self-report attachment measure in existence and factor-analyzed them, finding that two dimensions kept emerging from the data — attachment anxiety (worry about abandonment and rejection) and attachment avoidance (discomfort with closeness and dependence). The ECR consisted of the 36 items that measured those two dimensions best, 18 per dimension.
Two years later, in 2000, Chris Fraley, Niall Waller, and Kelly Brennan published the revision — the ECR-R — in the Journal of Personality and Social Psychology. The headline facts are easy to state: the ECR-R still has 36 items, still measures the same two dimensions with 18 items each, and still uses a 7-point response scale from "Disagree strongly" to "Agree strongly." Your score on each dimension is the average of its items, a number between 1 and 7.
So if the format barely changed, what was revised? The items themselves — and, more importantly, the method used to choose them.
Why the Revision Happened: The IRT Story
The original ECR items were selected using classical psychometric techniques, which evaluate how well items hang together overall. That approach produces a reliable scale, but it has a known blind spot: items tend to get chosen because they discriminate well among people at the high end of a trait. A statement like "I worry a lot about my relationships" separates very anxious people from moderately anxious people quite well — but it does a poor job distinguishing between people at the secure end of the spectrum, all of whom simply disagree with it.
Fraley, Waller, and Brennan approached the item pool with item response theory (IRT), a psychometric framework that evaluates each item individually: how much measurement information does this item contribute, and at which levels of the underlying trait does it contribute it? Their analysis showed that the original ECR, like most attachment measures of its era, measured the insecure end of each dimension with much more precision than the secure end.
The ECR-R was assembled to fix that. Drawing from the same large item pool, the authors selected the 36 items that maximized measurement precision across the full range of both dimensions — including the low, secure end where earlier scales went blurry.
Why does this matter to an ordinary test-taker? Because a lot of interesting attachment questions live at the secure end. Whether someone is moderately secure or very secure, whether a person's scores have shifted from mild insecurity toward security after a good relationship or therapy — these are exactly the distinctions the original scale was weakest at capturing and the ECR-R was engineered to capture. If a researcher wants to track subtle change over time, precision across the whole range is not a luxury.
The result is that the ECR-R is now the default instrument in adult attachment research. It's worth stressing that this doesn't make the original ECR a bad measure — the two versions overlap heavily and produce strongly related scores. The revision is a refinement in measurement engineering, not a change in what's being measured.
What the ECR-R Does and Doesn't Tell You
Like its predecessor, the ECR-R measures your default strategies in close romantic relationships along two continuous dimensions. The familiar four attachment styles are just the quadrants those dimensions form: secure (low on both), anxious-preoccupied (high anxiety, low avoidance), dismissing-avoidant (low anxiety, high avoidance), and fearful-avoidant (high on both).
Three boundaries are worth restating, because they apply to every version of the instrument, online or on paper:
- It's dimensional, not categorical. The ECR-R gives you two numbers, not a box. Any style label attached to your result is a convenience drawn over continuous scores — our scoring guide walks through why the numbers carry more information than the label.
- It's a self-report measure. It captures your perception of your patterns, which is honest data but not omniscient data. The strengths and limits of that are covered in our piece on attachment test accuracy.
- It's not a diagnostic tool. No ECR-R score, however high, diagnoses a mental health condition. It describes relational tendencies studied in social and personality psychology.
Taking an ECR-Style Test Online: What to Look For
Here's where the practical questions start. The ECR-R was built for research settings, and its 36 items — many of them near-paraphrases of each other, by design — make it a long sit for casual use. So the online landscape splits into three tiers, and it's worth knowing which one you're looking at:
Full research versions. A few academic sites administer the complete 36-item ECR-R with 7-point ratings, sometimes as part of ongoing studies. These give you the closest thing to a research-grade score. The trade-off is length and, often, an interface built for data collection rather than for explaining your result.
Validated-lineage adaptations. Many serious attachment tests — including ours — are shortened adaptations drawn from the ECR tradition. A well-built short form keeps the essential architecture: two separately scored dimensions, items sampling the core content of each (abandonment worry, reassurance-seeking, and rejection sensitivity for anxiety; discomfort with closeness, self-reliance, and emotional withholding for avoidance). What it trades away is some measurement precision — fewer items means a coarser estimate, particularly at the extremes the extra ECR-R items were chosen to sharpen.
Unanchored quizzes. Tests with no named instrument behind them, no separation of anxiety and avoidance, and results written for shareability. These aren't versions of the ECR or ECR-R at all, whatever their titles claim.
A simple check works across all three tiers: does the test name its lineage, score anxiety and avoidance separately, and tell you what it can't do? If yes, you're in the first or second tier and the result is worth taking seriously as a starting point.
Where Our Test Fits
To be precise about our own instrument, since this article is about precision: the test on this site is an 18-item adaptation based on the original ECR item pool — nine anxiety items and nine avoidance items, rated on a 5-point scale — not the full 36-item, 7-point ECR-R. We report both dimension scores along with your quadrant, because the dimensions are where the real information lives.
That design is a deliberate trade. Eighteen items take about three minutes, which is short enough that people actually finish thoughtfully rather than click through the second half of a 36-item form. What you give up relative to the full ECR-R is fine-grained precision — the kind a researcher needs to detect small changes across a study, and the kind the IRT revision specifically added at the secure end of each dimension.
For the question most people bring to an online test — where do I roughly sit on anxiety and avoidance, and which patterns should I be paying attention to? — a careful short form answers it well. If you need a research-precise score, seek out the full ECR-R in an academic context; if your relational patterns are causing real distress, a licensed therapist can assess attachment in the context of your actual history, which no questionnaire of any length can do.
The Bottom Line
The ECR-R is the 2000 revision of the ECR, rebuilt item-by-item with item response theory so that all 36 items measure attachment anxiety and avoidance precisely across the full range of each dimension — not just at the insecure end. It measures the same two dimensions as the original, on the same 7-point format, and it remains the gold standard in adult attachment research.
Online, most tests you'll encounter — including this one — are shorter adaptations of that tradition rather than the full instrument. That's not a flaw as long as the test is honest about it: the two-dimensional logic is what makes an attachment test meaningful, and a short form that preserves it gives you a directionally sound picture in a fraction of the time.
Curious where you land? Our free attachment style test is an 18-question adaptation of the ECR — about three minutes, no sign-up, with both your anxiety and avoidance scores explained. This article is educational and not a substitute for professional care. If you're in crisis, call or text 988 (US Suicide & Crisis Lifeline).