Maternal wounds & attachment

Your Attachment Style Is Not Your Star Sign

By Dr. Donetta D. Quinones, PhD, LMHC, LPC  ·  7 min read

She had done the reading before she took her coat off.

She sat down, declined water, and told me she was anxious-preoccupied, that her mother’s unpredictability had produced hypervigilance, and that she was drawn to avoidant men because inconsistency felt like home. Ninety seconds, and roughly true.

Then: “I did the quiz four times.”

Three different answers. Anxious. Anxious again. Then disorganized, which ruined a weekend. Then secure with anxious tendencies, “which I think is what they say when they want you to buy the workbook.”

She laughed, then asked the question I have been turning over ever since.

“So which one am I?”

Call her Yvette — a composite, like everyone I describe. In thirty-four years nobody had handed her a word for any of this. Not her mother, not two fiancés, not the school counselor who saw her twice in 2004.

Her phone did. That is not nothing.

The generous case, which I mean

People are not fools for finding these words useful. They are reaching for a name for something that had no name. If you grew up in a house where nothing was labelled — where the mood arrived and left like weather and no adult ever said she is unwell or this is not about you — then the first phrase that fits is not a fad. It is the first accurate sentence anyone has ever spoken about your life.

And the internet found real things. Childhood emotional neglect. The loneliness of the child nobody worried about. It found people competent and unreachable, faster and in far greater numbers than my profession, which requires an appointment and a copay.

So the argument is narrower than the internet is dumb: the vocabulary spread faster than the evidence, and it is now doing work the evidence cannot support.

The mechanism is not lying

In 2025, Karasavva and colleagues had two licensed clinical psychologists rate 98 popular ADHD videos on TikTok against the diagnostic criteria, claim by claim. Fewer than half the symptom claims lined up. Mean quality rating: 1.78 out of 5. No video scored perfectly. Psychiatrists rating the top hundred ADHD videos in 2022 had already called 52 percent misleading.

But here are the two numbers.

Only 4.1 percent of the videos noted the claims might not apply to everyone with the condition.

Only 1.4 percent noted the symptoms also occur in people who do not have it.

Sit with the second. One video in seventy mentioned that ordinary people also lose their keys and cannot start a task until the day it is due.

That is the mechanism, and it is not malice. Almost nobody there was inventing symptoms. They were omitting the qualifier, the boring and algorithmically fatal sentence that says this might not be you. A video with that sentence removed is not a lie. It is a true statement wearing a costume, and the costume is universality.

Then the study that settles it. Also in 2025, Aragon-Guevara and colleagues analyzed 133 autism videos with 198.7 million combined views: 27 percent accurate, 41 percent inaccurate, 32 percent overgeneralized. Then they checked whether accuracy predicted engagement.

It did not. Truth is not one of the variables.

The part I would leave out if I were being tidy

Nick Haslam named what happens to a concept when it travels. Concept creep, from his 2016 paper: harm concepts expand horizontally, sideways into new kinds of things, and vertically, downward into milder versions. In 2023, Baes with Haslam and colleagues tested this across 825,628 psychology abstracts and found exactly that for “trauma.”

Now the inconvenient part. The same group ran the same method on “anxiety” and “depression,” across a psychology corpus of over 133 million words and a general American one of over 500 million.

They found the opposite. The severity of the words appearing alongside them went up. More clinical, not less. Pathologization, not dilution.

The tidy essay says every word is being watered down and the kids have medicalized ordinary sadness. Satisfying, and not what the data say. I could have skipped that finding and you would never have known.

I would rather you trust me than agree with me.

Which brings me to therapy-speak, and the most honest sentence available: nobody has studied it. There is no peer-reviewed empirical literature on it. None. The closest thing to evidence concerns one word — in 2025, Darke, Paterson and van Golde found 63 percent of 595 participants used “gaslighting” spontaneously, before researchers introduced it, and well outside the intimate-partner situations it was built for.

Eldest daughter syndrome is birth-order astrology

There is no validation study for eldest daughter syndrome. What exists is two peer-reviewed papers studying it as internet culture: one on a Taylor Swift song, one on a hashtag. It has never been studied as a condition because it is not one.

Set that against what birth order does. In 2015, Rohrer, Egloff and Schmukle examined three national panels and found no birth-order effects on extraversion, emotional stability, agreeableness, conscientiousness or imagination. That same year, Damian and Roberts ran it on 377,000 American high school students: personality correlations of about .02, and an IQ difference favoring firstborns of about one point.

One point. In three hundred and seventy-seven thousand people.

And yet the thing the phrase reaches for is real, and it is not birth order. It is parentification — the child handed a parent’s job before she was built to hold it. Lisa Hooper and colleagues pooled 12 studies covering 2,472 people and found an association with adult psychopathology of r = .14. Real, reliable, small, and all retrospective self-report.

So which one is she?

People searching for anxious avoidant attachment in relationships are looking for a box. The research offers a pair of dials. When Kelly Brennan and colleagues factor-analyzed 482 attachment items in 1998, the variance collapsed onto two dimensions, anxiety and avoidance, with secure meaning low on both.

The instruments also disagree. In 2007, Glenn Roisman and colleagues compared interview-based attachment measures against self-report romantic questionnaires and found the overlap trivial to small. The interview and the quiz you took at eleven at night are not measuring the same thing.

And none of it is fixed. R. Chris Fraley’s meta-analytic work puts stability from infancy to adulthood at moderate, broadly in the .25 to .39 range. Real continuity, nowhere near a birth chart.

Yvette was not misdiagnosed by the internet. She was correctly described and then left there, holding an accurate label and no instruction, which is like being handed a good X-ray and no doctor.

Something to do, instead of another quiz

One week. Carry something to write on.

Every time the alarm fires — an unanswered message, a shift in someone’s tone — write three things down before you do anything else. What you concluded. How confident you were, as a number. And how old the conclusion feels: whether it is a thought about this person, or one you have been having since you were nine.

Then, when you find out what actually happened, write that next to it.

At the end of the week you have something no quiz can give you: a private accuracy record for your own alarm. Most people find it is not uniformly wrong. It is wrong in a particular direction, about a particular kind of silence. That is usable. “I’m anxious-preoccupied” is not.

This is my clinical adaptation of a standard cognitive practice, not a tested protocol — a habit of mind, not a treatment. If the week turns up something that knocks the wind out of you, that goes better with a person in the room than alone at midnight.

Now the part where I turn this on my own book

No study has ever tested whether learning your attachment style improves your relationships.

Not one randomized trial, not one controlled study, not one outcome evaluation. The most popular self-help framework of the decade is entirely unevaluated. That attachment insecurity predicts lower relationship satisfaction is well supported — Candel and Turliuc pooled 132 studies on exactly that in 2019 — that knowing your style improves anything has never been tested. A measurement predicting an outcome is not an intervention producing a change, and conflating the two is the central error of this genre.

It gets worse for me. In 2003, Marian Bakermans-Kranenburg, Marinus van IJzendoorn and Femmie Juffer meta-analyzed 70 studies of attachment interventions and put the answer in the title: less is more. What worked was brief and behavioral — insensitive parenting reduced at an effect size of 0.33, infant attachment insecurity at 0.20. Changing what parents did. Longer, insight-oriented work aimed at understanding one’s own history did worse.

The usual defense is bibliotherapy. Pim Cuijpers pooled six studies with 272 participants in 1997 and found a large effect against a waiting list. But look at what was in those books: structured cognitive-behavioral workbooks, for depression, with exercises and schedules. Media-recruited volunteers, likely publication bias.

Different product category.

So: mine is an insight-oriented conceptual intervention with no outcome data, in a category whose closest tested relative underperformed the behavioral version. I believe it is useful. I have watched something like it be useful in a room, hundreds of times. That is clinical experience — real knowledge, famously unreliable, not a trial.

Treat everything in it as hypotheses about yourself. Not findings, not a diagnosis, not a category you belong to. A hypothesis is a sentence that could be wrong, and its value is that you can check it.

Keep the words. Just hold them the way you would a first draft written by somebody who never met you.

Then judge me by one thing. Whether anything is different in six months. Not whether you felt understood — that is the easiest effect in publishing to produce. Whether you said something you would not have said. Whether you stayed in a room you would have left. Whether one person got closer.