The Adult Attachment Interview (AAI) (Pt.2 of 2): Mary Main’s Scary Parent Study

Scary Parents

Mary Main BerkeleyPart 1 of this blog Jan. 23 concluded that by 1978, Dr. Mary Ainsworth’s estimate for U.S. babies was that (A) 23% were Avoidant insecure (avoid parents); (B) only 69-70% were Securely Attached; and (C) 8% were Ambivalent insecure (unhappy and indecisive).  [FN1, 11]

By 1988, her Strange Situation study had been done with 2,000 infant-parent pairs in 32 studies in 8 countries. By 1999 it  had been done globally with 6,282 infant-parent pairs. Some countries varied, but worldwide results averaged out the same as Ainsworth’s 1978 original. [FN1, 9, 11]

In 1973 Mary Main [above] became Ainsworth’s grad student at Johns Hopkins in Baltimore, working on the Strange Situation experiments from the start. After her doctorate, Main moved to Berkeley, to see if Ainsworth’s Kampala and Baltimore findings would replicate.[FN8]  In 1978 she ran a Strange Situation study of 189 Bay Area infant-parent pairs and made the same findings. [FN10]

But by 1979 Dr. Main was making her own discoveries—out of concern about the parents. “In none of Ainsworth’s original observations was the possibility considered that some mothers… could also be frightening,” Main notes. “For my dissertation at Johns Hopkins, I watched 50 children in the Strange Situation… Using Ainsworth’s three-part classification (secure, avoidant and ambivalent), I found at least five infants could not be classified.”  Ainsworth was concerned, too; in fact, she’d left some babies in her Secure set only since they didn’t fit her other two sets. [FN8]

By 1982, Main decided to “extend attachment theory to include the import of infant exposure to anomalous fear-arousing parental behaviors… The mother is the haven of safety that must be approached in times of danger. However, when the infant’s biological haven of safety has simultaneously become a source of fright, the infant is placed in an irresolvable and disorganizing approach-flight paradox,” she said.  [FN8]

Soon after 1982, Main and Ainsworth agreed that some 15% of babies actually formed a new, fourth group: (D) Insecure Disorganized. Their mothers were so frightening that the babies couldn’t develop any consistent response at all. These infants “exhibited a diverse array of inexplicable or overtly conflicted behaviors in the parent’s presence” including “disorganization, disorientation, and confusion.” This includes crying loudly then suddenly freezing; ignoring the parent to rock on hands and knees; moving away; raising hand to mouth in fear; or even swiping at the parent’s face. They seemed vulnerable to dissociation.   [FN8, 11]

This finding that a whopping 15% of average U.S. babies are so insecure they’re almost incoherent was so shocking it was checked for years. [FN11, 12] But the number held. Worse, among children of American adolescent mothers the rate is over 31%, and is over 25% in many Third World nations. [FN13]

To remove the 15% of disorganized babies from Ainsworth’s original 69-70% Secure, reduced the Secure set to only 54-55%.  What was up with parents that 45-46% of their kids couldn’t manage secure attachment?  If so, Main would have  four categories: Avoidant 23%, Ambivalent 8%, Disorganized, 15%, Secure 55%.  [FN1, 11]

That was how researchers, including Dr. Benoit in footnote 1, read Main’s documents for 20 years 1990-2010. See also footnote 11.

In numerous writings, Main and colleagues called the disorganized group a new “fourth category,” a category in addition to Ainworth’s original three, and added that most disorganized subjects had been classed “secure” under Ainworth’s original three categories, but in fact belonged elsewhere.
FN12 Main M, & Solomon J (1986): ” …the majority of these…disorganized-disoriented infants would have been identified as secure… had we forced them into the standard classification system.
FN11 vanIjz1999, p.226 :  “the limits of the traditional Ainsworth et al. (1978) coding system became apparent because many children with an established background of abuse or neglect nevertheless had to be forced into the secure categoryDisorganization of attachment is usually considered a type of insecure attachment...”

Dr. Ainsworth herself, co-authoring a volume with Dr. Main in 1990, writes on page 480:  “… it is acceptable to consider adding…the fourth D category that Main identified as disorganized… Now that so much careful work has been done to explore the ramifications of this new category, it may be accepted as a valuable extension of the Strange Situation classificatory system.”

Recently, however, there has apparently been over-use and abuse of the term “disorganized” by social services and courts, such that Dr. Main’s group began around 2011 to criticize the “reification” of disorganized as a category:

“Main, Hesse, and Hesse (2011, p.441) have criticized the “widespread” and “dangerous” presumption that infants can be divided into four categories of comparable status… Solomon expresses particular concern:
“\’ The reification of our work from its context—and a lack of awareness of the grounding of our ideas in the behavioral and theoretical contributions of Bowlby and Ainsworth—has lead readers to treat D as a category equivalent in kind to ABC, rather than recognizing it as a phenomenon that runs orthogonal to the basic Ainsworth patterns. (personal communication, April 2013) ‘/ ”
from: “The Emergence of the Disorganized/Disoriented (D) Attachment Classification, 1979–1982,” by Robbie Duschinsky, Hist Psychol. 2015 Feb; 18(1): 32–46.
Cited within which: “Attachment theory and research: Overview with suggested applications to child custody,” Family Court Review, 49, 426–463, Main M., Hesse E., & Hesse S. (2011).

The Adult Attachment Interview (AAI)

Mary Main '09 Bowlby-Ainsworth award(Inge Bretherton,Everett Waters)Either way, by 1982 Main had seen enough to begin developing the Adult Attachment Interview (AAI).  She’d seen plenty of disorganized babies – that meant a lot of scary parents. She needed a way to document the behavior of parents. [FN14] (Right: Dr. Main receives the Bowlby-Ainsworth Attachment Award, 2009.)

The AAI was designed to document the level of secure, loving attachment the parents had, during their own childhoods with the babies’ grandparents. Parents were asked “both to described their attachment-related childhood experiences – especially their early relations with parents – and to evaluate the influence of these experiences on their development and current functioning.”

Parents were studied not just on facts they gave, but on how coherent a narrative they could produce quickly. That’s easy for folks who had a secure childhood, but difficult for folks who did not. The AAI questions are designed to “surprise the unconscious” to yield information about the “state of mind with regard to attachment” that might not otherwise show up. Each AAI was taped and transcribed verbatim. Then transcripts were classified by specific patterns by independent trained specialists. [FN15, 10]

In 1983 Main also created the “Berkeley Longitudinal Study” to take the Bay Area infant-parent pairs in her 1978 Strange Situation, and study them for a generation. In 1978, the babies were 12-18 months old; Main and her team re-studied these pairs when the kids reached age 6 in 1983, and studied the kids again when they reached 19 in 1996.  [FN8].   In 1983 Main and her team gave three completely different tests to the Bay Area pairs:

— 1. The six year olds were again studied with their parents in the Strange Situation (as in 1978 when they were infants).

— 2. The “sixes” were also individually (without parents) given a new Separation Anxiety Test (SAT): they were shown pictures of children being separated from their parents, then asked how they felt. This was taped, transcribed and sorted by Main’s student Nancy Kaplan into Ainsworth’s three sets: secure, avoidant and ambivalent. [FN16, 8]

— 3. The parents were given the AAI (without their kids). Responses were sorted into three set of adult attachment matching Ainsworth’s three 1978 categories of infant attachment: Secure-autonomous (matching infant Secure), Dismissing (infant Avoidant)  and Preoccupied (infant Ambivalent). (Main’s new 4th category wasn’t in use until after 1990.) [FN 17, 8, 10]

Astonishing Results You’ve Never Heard

Mary Main, Erik Hesse '09 Bolwby-Ainsworth AwardDr Main’s first 1983 results were so astonishing that attachment researchers have been buzzing ever since.  (Dr. Main and Dr. Erik Hesse, right.) Her results were also so important that it’s outrageous that your doctor never learned this in medical school; your therapist (and mine) never heard of this; you’ve never heard of it; and so you have to read about it here, since the media doesn’t report it. It’s hard to even find this story on the internet; I had to sleuth it out.  Her results were:

First: the six year olds’ 1983 responses with parents in the Strange Situation correlated strongly to their 1978 responses in the Strange Situation as infants five years earlier. In 1978 the infants were Securely attached 69%;  Avoidant 23%, and Ambivalent 8%. In 1983 the same kids at six were the same as they had been as infants, in the same percents. [FN7]

Second: the six year olds’ 1983 solo responses to the SAT photos also produced the same results and percentages.

Third: The parents’ 1983 solo responses to the Adult Attachment Interview correlated strongly with just how their own kids had behaved as infants, five full years earlier.  The parents also turned out to be Secure-autonomous (matching infant “Secure”) 69%;  Dismissing (matching infant “Avoidant”) 23%;  and Preoccupied (matching infant “Ambivalent”) 8%.

Fourth:  The match of the parents’1983 AAI security with how securely their babies behaved five years back in 1978, was at an unheard-of level. It correlated more than 70% of the time, in a field where a 20% correlation is highly significant. “A marked relation between a parent’s discussion of his/her own attachment history (AAI), and the offspring’s Strange Situation behavior 5 years previously, had been uncovered.”  [FN8, 10]

Fifth: When the 1978 infants reached age 19 in 1996, they too were given the Adult Attachment Interview. Again results correlated strongly: the 19 year olds’ responses in the AAI in 1996, correlated precisely to their infant behavior in the Strange Situation in 1978, their behavior at six in 1983, and to their parents’ 1983 AAI responses.

That means the Strange Situation predicts an infant’s behavior for life, and the AAI proves it.  [FN8, 10]

Sixth: When after 1990 Mary Main’s four categories were used, the results of all the tests were just as strongly correlated across Main’s four categories, as when data were sorted into only three sets. Main’s four categories did become Avoidant 23%, Ambivalent 8%, Disorganized, 14-15%, and Secure 54-55%.  Whenever responses to the Strange Situation, SAT, and AAI were sorted into these four sets,  the same percents were found. [FN1, 11]

By 2009, over 10,500 subjects globally had been given the AAI and the results continued to average out the same.  [FN18]

“Researchers worldwide have replicated the relation originally uncovered in the Bay Area study between a parent’s status in the Adult Attachment Interview and an infant’s Strange Situation response to that same parent… The same average parent-to-child, secure/insecure match of 75% holds even when the interview is conducted before birth of the first child…

“Describing the strength of this relation across studies conducted several years ago, van I Jzendoorn (1995) calculated that it would take 1,087 further attempted replications, every one yielding insignificant results, to reduce the present relation between adult and infant attachment status to insignificance,” Main concluded. [FN7, p.1091]

Kathy’s news blogs expand on her book “DON’T TRY THIS AT HOME: The Silent Epidemic of Attachment Disorder—How I accidentally regressed myself back to infancy and healed it all.” Watch for the continuing series each Friday, as she explores her journey of recovery by learning the hard way about Attachment Disorder in adults, adult Attachment Theory, and the Adult Attachment Interview.


FN1  NIH: Benoit, Diane, MD, FRCPC, “Infant-parent attachment: Definition, types, antecedents, measurement and outcome,” Paediatr Child Health, Oct 2004; 9(8) p. 541–545 at:  4th subhead “Measurement” reports:
“The three ‘organized’ strategies (secure, avoidant and resistant) are assessed in the Strange Situation (SS), a 20-minute laboratory procedure where patterns of infant behaviour toward the caregiver following two brief separations are categorized… “Infants with secure attachment greet and/or approach the caregiver and may maintain contact but are able to return to play, which occurs in 55% of the general population… Infants with insecure-avoidant attachment fail to greet and/or approach, appear oblivious to their caregiver’s return… avoiding the caregiver, which occurs in 23% of the general population. Infants with insecure-resistant [ambivalent] attachment are extremely distressed by separations and cannot be soothed at reunions,  displaying much distress and angry resistance to interactions with the caregiver, which occurs in 8% of the general population.” [This NIH article earlier reports that the remaining “approximately 15% suffer insecure ‘disorganized’ attachment,” citing their own footnote which states “In normal, middle class families, about 15% of  infants develop disorganized attachment.” [23%+8%+15%  = 46% not securely attached.]

FN2  Felitti VJ, MD; Anda RF, MD, et. al, 1998, “Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study,” American Journal of Preventive Medicine 1998;14:245–258.  Detailed article on the ACE Study:

FN3  Karr-Morse, Robin, Wiley, Meredith,  “Scared Sick,”  Penguin Basic Books, 2012

FN4  Porges, Stephen, PhD, “Body, Brain, Behavior: How Polyvagal Theory Expands Our Healing Paradigm,” 2013,
“Beyond the Brain: Vagal System Holds the Secret to Treating Trauma,” 2013,
–“The Polyvagal Theory for Treating Trauma,” 2011,
–“Polyvagal theory: phylogenetic substrates of a social nervous system,” International Journal of Psycho-physiology 42, 2001,  Dept. of Psychiatry, Univ. Illinois Chicago,

FN5   Earned secure attachment occurs when we experience harmful parenting, so we start with insecure attachment, but find ways to “rise above” childhood trauma and “are now securely attached… What’s more important than what happened to us, is how we’ve made sense of our own childhood,” Dr. Dan Siegel says. “When we make sense of our past… we become free to construct a new future for ourselves and for how we parent our children. Research is clear: If we make sense of our lives, we free ourselves from the prison of the past.”  (Source: video by Dr. Mary Main, Dr. Erik Hesse, Dr. Daniel J. Siegel, Dr. Marion Solomon: )
The question is: what ways?  How do we “rise above” and “make sense” of our childhood trauma?
“Mindfulness has been shown to be effective in healing insecure attachment,” say Siegel’s recent writings. “The purpose of both psychotherapy and mindfulness practice is to provide this internalized secure base. Attunement, whether it is internal in mindfulness, or interpersonal in attachment, is what leads to a sense of secure base.” (The Mindful Therapist: A Clinician’s Guide to Mindsight and Neural Integration, W.W. Norton, 2010;   Mindsight: The New Science of Personal Transformation,  Random House, 2010).
“The regular exercise of mindful awareness seems to promote the same benefits–bodily and affective self-regulation, attuned communication with others, insight, empathy, and the like–that research has found to be associated with childhood histories of secure attachment,” Siegel wrote earlier. “ Mindfulness and secure attachment alike are capable of generating… the same invaluable psychological resource: an internalized secure base.” (Siegel, 2007, Wallin, 2007, p. 5-6).
In his 1999 book “The Developing Mind: How Relationships and the Brain interact to shape who we are,” (Guilford Press),  Siegel defined “earned secure/autonomous attachment” as a pattern noticed by therapists doing the Adult Attachment Interview:  “individuals whose experiences of childhood… [were] likely to produce insecure attachment (avoidant, ambivalent, or disorganized),” but their AAI interview responses instead show “a fluidity in their narratives and a flexibility in their reflective capacity, such that their present state of mind with respect to attachment is rated as secure/autonomous. These individuals often appear… to have had a significant emotional relationship with a close friend, romantic partner, or therapist, which allowed them to develop out of  insecure…into a secure/autonomous AAI status.”

FN6  Bowlby, John, “The Nature of a Child’s Tie to His Mother,” British Psychoanalytical Society, London, 1958; “Attachment and Loss,” New York, Basic Books, 1969

FN7  Main, Mary,  2000, “The Adult Attachment Interview: Fear, attention, safety and discourse processes;” also titled “The Organized Categories of Infant, Child, and Adult Attachment: Flexible vs. Inflexible Attention Under Attachment-Related Stress,” Jour of Amer Psychoanalytic Assoc, 48:1055-1095; 2000.  *p.1091: “The same average parent-to-child, secure/insecure match of 75% holds even when the interview is conducted before birth of the first child…”

FN8  Main, Mary, 2005, with Hesse, Erik & Kaplan, Nancy, “Predictability of Attachment Behavior and Representational Processes at 1, 6, and 19 Years of Age – The Berkeley Longitudinal Study,” Chapter 10 of “Attachment from Infancy to Adulthood: The Major Longitudinal Studies,” edited by Klaus E. Grossmann, Karin Grossmann, and Everett Waters, pp. 245–304, New York: Guilford Press. Main refers to it as “Regensburg.” Regensburg 2005 .pdf
–Main’s summary of the Strange Situation in this document:
“Ainsworth structured the Strange Situation procedure to include three of Bowlby’s ‘natural clues to danger’ in eight episodes:  1. Introduction to the room.  2. Mother and infant are left alone in a toy-filled environment whose unfamiliarity supplies the first natural clue to danger.  However, the mother’s presence is expected to provide the infant with security sufficient for exploration and/or play.  3. Providing a second clue to danger, a stranger joins the mother and infant.  4. The mother leaves the infant with the stranger, providing two combined clues to increased danger. 5. The mother returns, and the stranger departs…. Many infants initially seek proximity but then, reassured of their mothers’ nearness, resume play. 6.  The mother leaves, and the infant remains entirely alone in the unfamiliar setting. Infant distress can be strong at this point, and this episode is often terminated rapidly. 7. The stranger, rather than the mother, enters the room. 8. The mother returns… By now, most infants are expected to be crying, and actively not only seeking proximity to mothers, but also… indicating a strong desire to be held… Nonetheless, they are expected to settle and renew interest in exploration and play by the end of this 3-minute period…
“Somewhat surprisingly, Ainsworth found that infant responses to separation and reunion in this procedure fell into three distinct, coherently organized patterns of attachment (“secure,” “insecure-avoidant,” and “insecure-ambivalent” (Ainsworth, Blehar, Waters, & Wall, 1978). … Given the length and complexity of this chapter, we suggest individuals divide reading to its three central parts (secure attachment, pp. 261–273; avoidant attachment, pp. 273–279; and disorganized attachment pp. 279–288).”

FN9  (1988 van Ijzendoorn: on global proof of Ainsworth’s 3 categories; written before Main’s 4 category put in use)
van IJzendoorn, Marinus H.; Kroonenberg, Pieter M.  “Cross-Cultural Patterns of Attachment: A Meta-Analysis of the Strange Situation,” Child Development,Vol 59 No 1, Feb 1988, p.147–56.   Abstract:

FN10   Hesse, E., (2008) “The Adult Attachment Interview: Protocol, Method of Analysis, and Empirical Studies,” Chap. 25 of Cassidy, Jude &  Shaver, Phillip R. (Eds), “Handbook of Attachment: Theory, research, and clinical applications,” 2nd edition, 2008, p. 552-598, New York, Guilford Press. It was online (I downloaded it Aug. 2014) but was taken down or has web issues; try here:   [His 1999 version in Cassidy & Shaver’s 1st edition, Chap. 19,  “The Adult Attachment Interview: Historical and current perspectives,” p. 395-433 wasn’t online.]

FN11  van IJzendoorn MH, Schuengel C, Bakermans-Kranenburg MJ, Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants and sequelae. Dev Psychopathol. 1999; 11:225–49. [PubMed] at

FN12  Main, Mary, & Solomon, Judith, (1986), “Discovery of an insecure disoriented attachment pattern: procedures, findings and implications for the classification of behavior,” in Brazelton T, Youngman M. Affective Development in Infancy, Ablex, Norwood, NJ
Main, M., & Solomon, J. (1990). “Procedures for identifying infants as disorganized/disoriented during the Ainsworth strange situation,” in Greenberg, M. T., Cicchetti, D., & Cummings, M. (Eds.),. Attachment in the preschool years: Theory, research, and intervention (pp. 121-160), University of Chicago Press

FN13  “Among children of American adolescent mothers, the rate is over 31% (Broussard 1995). Disorganized attachment is also common among the Dogon of Mali (~25%, True et al 2001), infants living on the outskirts of Cape Town, South Africa (~26%, Tomlinson et al 2005) and undernourished children in Chile (Waters and Valenzuela 1999),” reports NIH [op cit FN1; NIH cites v. Ijzendoorn 1999 and the other authors just mentioned]

FN14  George, Carol, Kaplan, Nancy, & Main, Mary, “Adult Attachment Interview,” Unpublished MS, Department of Psychology, University of California at Berkeley, third ed. 1996.  Original 74-page MS dated 1984, 1985, 1996.  Described in FN10 op cit Hesse 2008. ( I have it, but it was removed from the web during 2014.)

FN15  Main, Mary B., “Adult Attachment Interview Protocol,” 11 pgs, 20 questions, no date or publisher. Dr. Main requires intensive training for use of the AAI. The questions per se, marked “Do not reproduce this material without permission of the author,”  are here:   UCLA’s Lifespan Learning Institute in Los Angeles holds AAI workshops and has an extensive CD  lectures on the AAI at

FN16  Kaplan, Nancy, (1987), Separation Anxiety Test (SAT): “Individual differences in six-year-olds’ thoughts about separation: Predicted to actual experiences of separation,”  Unpublished doctoral dissertation, University of California, Berkeley.
Kaplan, N. (2003, April), “The development of attachment in the Bay Area study: One year, six years, nineteen years of age.” Paper at the biennial meeting of the Society for Research in Child Development, Tampa, FL.

FN17  Main 1985; Main & Goldwyn 1984

FN18  Bakermans-Kranenburg MJ, van IJzendoorn MH, “The first 10,000 Adult Attachment Interviews,” Attach Hum Dev. 2009 May; 11(3): 223-63. doi: 10.1080/14616730902814762 at

More by Mary Main on the AAI:
Main, Mary, PhD, “Introduction to the special section on attachment and psychopathology: 2. Overview of the field of attachment,” Journal of Consulting and Clinical Psychology, 64, 237-243, 1996
Steele, Howard and Miriam, Editors, “Clinical Applications of the Adult Attachment Interview,” The Guilford Press,  New York, 2008

5,263 total views, 5 views today

Share Button

6 responses to “The Adult Attachment Interview (AAI) (Pt.2 of 2): Mary Main’s Scary Parent Study

  1. Thank you for this terrific info! This is very important. It links with (jargon-free) research I did and a resource website I have which you and others may find useful:
    Here’s a 2-minute read article about my research:

  2. Pingback: Name Memories Without Words – Sundown Healing Arts

  3. It would be nice if we had a way to assess our own AAI score, but it’s dicey if people start assessing each other. It took me months to pull together this picture, detective-style, as to how Mary Main discovered all this. Mary Ainsworth has passed away. None of them wrote anything clear to the layman.
    Why? The more I read their documents, the more it seems they don’t want to encourage the average person using the AAI.
    Perhaps Dr. Main and co. are concerned that lay persons might abuse the AAI to judge or classify each other? They do insist in most things they write, that even professional therapists must get special training to use the AAI on clients.
    Yes, all therapists should be getting the training and using it! Yes, I’m upset most of them never heard of it.
    But us un-trained folks using it on each other? Dicey.
    I didn’t have time to find Dr. Main et. al. in person to ask them exactly what they’re thinking. I believe Mary Main is still at Berkeley. As Footnote 15 in my blog says, UCLA’s Lifespan Learning Institute in Los Angeles holds AAI workshops and has an extensive CD lectures on the AAI at
    To follow this further, check out these other footnotes:
    FN10 Hesse, E., (2008) “The Adult Attachment Interview: Protocol, Method of Analysis, and Empirical Studies,” 2nd edition, 2008, p. 552-598, New York, Guilford Press. It was online (I downloaded it Aug. 2014) but was taken down or has web issues; try here:
    FN15 Main, Mary B., “Adult Attachment Interview Protocol,” 11 pgs, gives 20 AAI questions, no date or publisher. Dr. Main requires intensive training for use of the AAI. The questions per se, marked “Do not reproduce this material without permission of the author,” are here:

  4. Have you seen One Love’s phone AP that helps us (or our loved ones) assess our risk of being harmed in domestic violence?
    I wonder if we could decrease spanking and increase awareness of positive parenting with an AP that allows you to self assess your own AAI score. Parents seek primary support from family, friends, and colleagues and secondarily the internet. Friends could also do an AAI in order to figure out if their loved one is at risk of having problems with attachment to their child. It could have all sorts of resources connected to it.
    Try to download the One Love AP and just go through it. It is amazing! The Ravens gave her $400,000 to support this project.
    Do you have any idea who I might contact to talk about this idea? Do you know how to reach Mary Ainsworth?

  5. I find this so absorbing and interesting. I’m currently getting my Masters degree in Psychology and would love to take the training for the Adult Attachment Interview. I live in Ontario, Canada – do you have any links or information about training for AAI ? Many thanks for sharing this important article!

Leave a Reply

Your email address will not be published. Required fields are marked *

 characters available