Child Development

Attachment Theory Is All Wrong. Here’s What The Science Really Says.

It's high time that parents ditched their secure attachment styles, a renowned Harvard psychologist says.

Updated: 
Originally Published: 
Two dads in a kitchen, one holding their baby.
Willie B. Thomas/Getty

People with insecure, anxious, disorganized attachment styles can rest easy. The science behind the 60-year-old theory of infant attachment is vanishingly thin and being dismissed by an increasingly large body of psychology researchers and clinicians, such as Judith Rich Harris and Tiffany Field. But how has the theory, which states that the first attachment style a child experiences will stick with them for life, persisted among parents and professionals for so long?

Psychoanalyst John Bowlby first proposed attachment theory in 1958 after looking at animal research that attributed most distress to absent mothers in the first year of life. The theory was later applied to humans, hypothesizing that if an infant successfully bonds with their primary caregivers, they will be able to have largely secure, emotionally stable relationships throughout their lives — and by extension, superior mental and emotional health. More importantly, if they’re not able to connect or bond, they’re doomed to a life of instability and much-needed therapy.

Bowlby’s evidence for such a theory was thin and included a mixture of personal and cultural confirmation biases, including his own childhood, in which he was disconnected from his mother but close with his governess, who quit when he was only 4.

Despite its shallow suppositions, the theory took root in a post–World War II America, in part because it appealed to fears about wives going to work. The notion that what a mother does during the first few years of life psychologically makes or breaks a child caught on because it told people what they wanted to hear.

“In retrospect, his suggestion that what happens in the first year of life influences in a significant way how you’ll be for the rest of your life is an unreasonable idea,” said late psychologist Jerome Kagan, Ph.D., who was a Professor Emeritus at Harvard University and a leading critic of the attachment theory. “The social class in which a child is reared is today in many nations the best predictor of depression, anxiety, addiction, crime, and occupations. It’s better than any set of genes or observation of a child.”

And yet, half a century later, people still subscribe to attachment theory despite ample evidence that social class, temperament, and culture are much more accurate predictors of future outcomes. But Kagan, who is listed by the American Psychological Association as the 22nd most eminent psychologist of the 20th century, is certain that within 10 to 15 years, attachment theory will be a historical footnote. He shared with Fatherly where his colleagues went wrong and where the field of psychology is moving to correct past mistakes, albeit very slowly.

How did attachment theory become the standard in psychology with very little research?

Bowlby and then his student Mary Ainsworth introduced the idea of attachment at a time right after World War II, during the 1950s, when America is the moral power of the world and everything is peaceful. There’s no big city gangs or opioid crisis or climate change, and unemployment is low. And during this time the question of why some children become successful in life and others don’t was of interest to most Americans. The favorite answer — which came from a time when ways to measure the brain had not been discovered — has been that it’s what the mother does in the first years of life that matters.

So that would make attachment theory a very American idea then?

Yes. Bowlby’s ideas were very popular in America but not in other places of the world, because he was telling them what they wanted to believe — that if a mother is loving and affectionate and consistent in the first year or two of life, then like a vaccine, the child will be protected from things like anxiety and depression for the rest of their lives.

Attachment is a far less popular explanation in 2019 than it was in the 1960s, and in 10 to 15 years, it’s going to be rare to find anyone defending the theory. It’s just dying out slowly.

How did Bowlby come up with this idea, exactly? What research did he use and why is it not accurate?

Bowlby was trained in psychoanalysis, but he visited psychologist Harry Harlow in Wisconsin and saw the behavior of the monkeys who had been separated from their mothers. And he said, “You see what happens if you don’t have your mother?”

And then he was very impressed by Konrad Lorenz’s discovery of imprinting in ducks. He assumed imprinting in ducks was similar to an infant bonding with their mothers, which it isn’t.

Finally, his colleague John Robertson, a pediatrician at a London hospital, told Bowlby that some 2-year-olds would come to the hospital for surgery and cry when no one was around. He thought, “Well, that’s it.” Even though Robertson told him that they wouldn’t cry if a nurse was there, or if they were 3 years old, Bowlby had made up his mind on the importance [of] the caretaker in the opening years.

From this, he makes the bold claim that if a mother is sensitive in the first few years of a child’s life, they will be protected from stress and mental illness for the rest of their lives.

And how did other psychologists build on this thin evidence?

His student Mary Ainsworth invented the Strange Situation, where she found that when mothers left the room and children would cry and couldn’t be quieted, then these children came from homes with less sensitive mothers. That seemed to Bowlby to be the final nail, and that’s when it became popular. In the 1950s and ’60s, you could not pick up a journal without finding an article on attachment.

How long did it take for people to question attachment theory and why aren’t psychologists more critical of it today?

By the ’80s, this castle had begun to crumble for several reasons: First, some scientists found that the temperament of the child is a major determinant to how they behave in a strange situation. Children with a more irritable temperament cry when the mother leaves and cannot be soothed. In Ainsworth’s theory, those children cry and cannot be soothed because they’re insecurely attached. Other children who don’t cry easily because of their temperament tend to be called securely attached. It’s temperament that determines how children respond.

Second, scientists found that children who grew up securely attached in their first year didn’t grow up to be protected from anxiety and depression. So that evidence made people question if secure attachment in that first year predicted anything.

But it seems like there had to be some pushback in the field of psychology if attachment theory is still popular today. How did that happen?

Those who supported attachment theory got defensive. Mary Main at the University of California Berkeley invented the Adult Attachment Interview, and said we don’t have to look at childhood. We can interview a person and tell if they were securely attached. The problem is the interview used the coherence of the person’s speech as an index of secure attachment, but that’s correlated with social class. Better educated people give more coherent interviews, and of course, better educated, wealthier people are always happier that poorly educated, poorer people. So the interview was not a good substitute.

In the end, attachment theory was just too simple. It ignored the temperament and social class of a child’s family, and it ignored the cultural setting. His beautiful but simple idea was slain by ugly facts. Attachment is a far less popular explanation in 2019 than it was in the 1960s, and in 10 to 15 years, it’s going to be rare to find anyone defending the theory. It’s just dying out slowly.

Yes, what happens to you in the first year or two of life has an effect, but it’s tiny.

Of course, abuse and neglect during early childhood are obviously bad for kids. How is acknowledging that not the same as attachment theory?

Abuse and neglect in the first years are bad, but they’re associated with class. Children who are abused and neglected are far more likely to come from poor families than wealthy families. If you’re raised in a poor, single-parent family, you’re more likely to be abused. If you’re abused, you’re more likely to have problems when you’re 20. Now we’re quick to say it must be the abuse, but if you’ve been raised in poverty, you can’t dismiss that.

I’m suggesting that an abused child from a wealthy family of privilege would be far more likely to not have problems, because to be born into a disadvantaged class means you’re going to encounter different teachers, different schools, different peers, different values for the rest of your life. So we can’t just blame the abuse.

Even though it’s not based on any facts, why do people want to believe attachment theory is real so badly? What makes it so appealing?

Attachment theory is attractive because Americans want to believe two things: that what happens in the opening years is critical and that a mother’s love has a special power, more than a father’s love. That belief is still powerful among Americans and the British, that her love has a special effect on a child. Even with gender equality, that idea is still very present in society.

A big part of this seems to come from fears about women entering the workforce. Is that accurate?

When mothers began to go to work in the 1960s, there were news articles saying this is going to be terrible, and of course, it wasn’t. Children who went to good daycare centers were fine. But the protests against women working were so powerful that when Nixon was president and thinking of having national daycare centers, it never happened because the protests were so strong.

If attachment theory is a fallacy, then what helps predict how children will turn out and how can parents use that to help their kids? Is it just about balancing class, culture, and other factors?

Of the more than a hundred million children who were born in 2018, they’re born with different biologies, different temperaments, different potentials, and it’s like a hundred million different seeds of plants. Imagine we scatter those across the world, some won’t grow at all if it’s too cold, and some will thrive. It’s the same thing for humans. I have these hundred million infants with different genes, and I put them in different families in different parts of the world at different times, and these variances shape them. The final factor is now what happens to those children growing up in families, what are the local circumstances.

Consider a child who is perfectly happy until a tsunami hits and kills their mother. No one could have predicted the tsunami. You could be a happy child in Syria before a civil war and suddenly they’re bombing every house. Look at the anxiety of America’s high school kids. When I was in school, I didn’t worry about getting into college because a lot of people didn’t go. Back then, no one would have predicted that today there would be so many young people worried about getting into college.

Historical circumstances are changing what you worry about. Most psychological theories focus on the person or the person’s family, not on what history is doing in changing the environment.

So is it fair to say that attachment theory is not real, but how children are cared for in their first two years is one of many factors involved in what you’re describing?

Attachment theory, as Bowlby stated it, is just not right. Let’s rephrase it: Yes, what happens to you in the first year or two of life has an effect, but it’s tiny. If I take a 1-year-old child who is securely attached, and the parents die and the child is adopted by a cruel foster parent, that child is in trouble. Their secure attachment is useless. When you think about it, it’s silly that after the first year you could predict with any confidence what this person is going to be like in 20 years. It’s a ridiculous idea.

This article was originally published on