By Vera Auerbach, Clinical Psychologist


The attachment focus began with the theoretical works of John Bowlby published some 30 years ago. Based on a psychoanalytic tradition, particularly the British Object Relations School, he challenged basic tenets for psychoanalysts and other clinicians. Attachment as a theory was revitalised as a science by Mary Ainsworth (1967) who created an experimental research paradigm (SSP) to study attachment theory in infancy. The introduction of video tapes where the infant could be videoed led to a landmark achievement in being able to reliably assess and classify mother-baby dyad attachment relationships.

With the later development of the Adult Attachment Interview (AAI) one could start researching issues of intergenerational transmission of attachment patterns. This also opened up an area of having scientific evidence for studying outcome of psychoanalytic psychotherapy, as now one can assess attachment patterns before and after therapy. Further we can now research how representation of attachment relationships in the caregiver relate to behavioural patterns of attachment in the child. It also leads us to be able to identify pregnant mothers (babies) who may be at risk due to their insecure attachment patterns, as we can tell which infant may be at risk being born to mothers with an insecure attachment pattern on the AAI. Insecure attachment can be seen as a risk factor for later disorder in childhood as shown by Emde, (1990).

Since Ainsworth the attachment world has grown with different theories emerging, and just like Melanie Klein and Anna Freud had divergent ideas, so too did the world of attachment have a separation in the development of attachment theory. Mary Ainsworth trained Inge Bretherton, Mary Main, Jude Cassidy, Robert Marvin, and Patricia Crittenden, among others.

Mary Main with Judith Solomon (1990) developed the concept of disorganisation (the D classification). They then trained Peter Fonagy and Target. It seems that the falling out in the attachment world was between the theories and systems developed by Mary Main and Patricia Crittenden. Crittenden (1995) had trained with an abnormal population, while Main was using middle class population samples. Crittenden (1999) argued that attachment theory needed to be able to be applied to clinical groups and developed her Dynamic-Maturational system.

Mary Main and Goldwin stayed with simply having an A, B, C, (and Ds1, Ds3, F1-5, E1-3 and CC) system. Mary Ainsworth is said to have commented that the world of attachment theory should now be grown-up enough to tolerate both theoretical approaches. Implications for training are important though and discussed below after the different classification systems are introduced.

CARE – Index 0-18 months (30 months)

The CARE index was developed by Patricia Crittenden (2003). It is the simplest and most versatile of the attachment measures. It assesses mother-infant interaction from birth to about two years of age based on a short, videotaped play interaction of 3-5 minutes. Once the coder is trained, coding of an interaction takes at least 15-20 minutes.

The measure assesses mothers on three scales: sensitivity, covert and overt hostility, and unresponsiveness. There are also four scales for infants: cooperativeness, compulsive compliance, difficultness, and passivity. These scales tend to be related to the maternal scales in the order listed. The scales

  • are highly correlated with the infant Strange Situation (SSP) assessment patterns of attachment
  • differentiate abusing from neglecting, abusing-and-neglecting, marginally maltreating, and adequate dyads
  • can be used during intervention, and
  • can be used to assess the effectiveness of an intervention.

The information derived from the CARE-Index can be used as quasi-continuous or categorical data by researchers (Crittenden, 2003). In addition the procedure is easily applied to “live” observations made by nurses, infant teachers, clinicians, and social workers. Needless to say, such observations cannot be treated as reliable when video is not used, but the method is effective in the same way that other guided observations or interviews are. The specific training of what to look for sharpens both one’s observation and one’s ability to tell others exactly what the dyad did and why it should be interpreted in a particular manner.

Training Courses

Crittenden (2003) runs a 5-day course taught from videotapes, which depends upon both adult and infant behaviour. Non-verbal behaviour, interpersonal strategies, and developmental processes are emphasised. Adult patterns consist of sensitivity, controlling, and unresponsivity; the infant patterns are cooperative, difficult, compulsive compliant, caregiving or attention, and passive. The course includes a preliminary reliability test based on a standardised set of American mother-infant dyads. At the conclusion of the course, participants generate a set of videotapes that reflect their intended applications and culture(s). These tapes are then used to create a course-specific reliability test.

A second 4-day course on toddlers (15-30 months) is now also offered. The patterns include a wider array of compulsive patterns and coercive patterns (both aggressive and passive) as well as combinations.

As with all courses offered by Dr. Crittenden, participants are given a written and signed statement of their percent agreement with the standard. This reliability can be reported in research articles. Evidence of reliability should be requested if the participant will code data for another researcher.

Ainsworth Strange Situation Procedure (SSP) – (11 – 15 (17) months)


The Infant Strange Situation is an instrument for assessing quality of attachment in 11-17-month-old infants. The procedure was developed by Mary D. Salter Ainsworth (1967) and has been expanded by both Mary Main and Judith Solomon (1990) (who added the notion of disorganisation) and also by Patricia Crittenden (who added the A/C and pre-compulsive classifications). The Ainsworth Infant Strange Situation involves a 20-minute laboratory procedure and takes three people to carry out (plus the child and mother). Classification is based on careful review of the videotape and takes at least one-half hour per videotape.

Dynamic-Maturational Classifications in Infancy – Crittenden (1999)

This classificatory system includes all of the categories of the original infant procedure (i.e. A1-2, B1-4, C1-2), plus several patterns that were identified by others, specifically, disorganisation, A/C, and pre-compulsive patterns. These additions make the system suitable for a wider range of cultures and home environments, especially those at risk for psychopathology.

Training Courses

Crittenden: The 5-day course is taught from videotapes, and classifications depend upon both maternal and child behaviour. Non-verbal behaviour, interpersonal strategies, and developmental processes are emphasised. The course includes a preliminary reliability test based on a set of American mother-child dyads. At the conclusion of the course, participants again generate a set of videotapes which will be used to create a course-specific reliability test.

Main: I believe Mary Main also offers training in the SSP but I have been unable to find her details.


Preschool Assessment of Attachment PAA – (18-21 months)

The Preschool Assessment of Attachment (PAA) is an instrument for assessing patterns of attachment in 18-month to five-year-old children. Like the Ainsworth Strange Situation for infants, the PAA involves a 20-minute laboratory procedure and takes three people to carry out (plus the child and mother). Classification is based on careful review of the videotape and takes approximately one-half hour per videotape.

Dynamic-Maturational Classifications in the Preschool Years (Crittenden, 1988, 1992)

This classificatory system includes all of the categories of the infant procedure (i.e., A1-2, B1-4, C1-2, A/C), plus several patterns that develop during or after the preoperational shift at the end of the second year of life. Specifically, Ainsworth’s three basic strategies for negotiating interpersonal relationships are modified to fit preschoolers’ more sophisticated mental skills and organisations of behaviour (Crittenden, 1992). Thus, the patterns are renamed as secure/balanced (Type B), defended (Type A) and coercive (Type C), (Crittenden, 1992).

With the addition of several new subpatterns, the PAA sets the stage for the even more complex patterns of behaviour that are observed at later ages, especially in the school years. Furthermore, a particular advantage of the expanded classificatory system used in the PAA as compared to other classificatory systems is that it differentiates “unendangered” A1-2 and C1-2 children from those using the compulsive (A+) and obsessive (C+) subpatterns (Crittenden, 1988). These patterns, i.e, the compulsive caregiving A3, compulsive compliant A4, aggressive C3 and feigned helpless C4 subpatterns, identify the children most likely to come from troubled homes and most likely to experience psychological, social, developmental, and learning problems. These coercive and defended subpatterns as well as the compulsive/obsessive A/C pattern have implications for emotional and behavioural problems that develop during the preschool and school years, Crittenden, (1995).

Training Courses

Crittenden (2003) offers a two-week course taught from videotapes and depends upon both maternal and child behaviour. Non-verbal behaviour, interpersonal strategies, and developmental processes are emphasised. The course includes a preliminary reliability test based on a standardised set of American mother-child dyads. At the conclusion of the course, participants generate a set of videotapes that reflect their intended applications and culture(s). These tapes are used to create a course-specific reliability test.

School-Age Assessment of Attachment (6 – 13 years)

Separation Anxiety Test (5 – 12 years) – SAA/SAT

I have not been trained in this but know it exists, and was developed by Stern.

Family Drawings (4 years on)

Adolescent (AAI) 16 – 21 years

As teenagers still live at home they are thought to be not as fully integrated as yet as an adult would be. Thus this is a somewhat changed version of the AAI. (See AAI)

Adult Attachment Interview AAI – (25 years plus)

Crittenden: Overview of the Dynamic-Maturational Method

The Adult Attachment Interview is a procedure for assessing adults’ strategies for identifying, preventing and protecting the self from perceived dangers, particularly dangers tied to intimate relationships. The course offered by Dr. Crittenden (2003) is based on an expansion of the Bowlby-Ainsworth theory (Crittenden, 1995) and an extensions of the Main and Goldwyn procedure (Main & Goldwyn) as applied to the Adult Attachment Interview developed by George, Kaplan & Main (1986 & 1996).

The dynamic-maturational (Crittenden, 1998) approach to the Adult Attachment Interview (AAI) is both a useful research tool and also a potential guide for psychotherapists. Therapists, in particular, learn not only new ways to conceptualise disturbed development, but also ways to identify in adults’ distortions of the mental processing of information, particularly information relevant to disorders of feelings, thought and behaviour. The techniques for interpreting speech can be useful even if the therapist does not formally use the interview itself in practice. For those interested in research applications, a 15-day training is available in the major classifications and subclassifications.

The Dynamic-Maturational approach to attachment theory emphasises the dynamic interaction of the maturation of the human organism, across the life-span, with the contexts in which maturational possibilities are expressed around the functions of self-protection, reproduction, and protection of progeny (Crittenden, 1998).

Major tennants of Crittenden’s theory are:

  • Maturation is both neurological mental and physical. It involves both the increase in functional potential during childhood and adulthood and the ultimate decrease in functioning in old age.
  • Contexts include both the people and places that affect development, e.g., family, school, and also the intra- and interpersonal challenges of different periods of the life-span. These challenges include:
  1. preschoolers’ need to learn safe forms of self-reliance for short periods of time;
  2. school-aged children’s need to establish symmetrical attachments with best friends while concurrently maintaining affiliative peer relationships;
  3. adolescents’ need to transform these same-sex attachments into heterosexual, reciprocal attachments with a sexual component;
  4. adults’ establishment of symmetrical and reciprocal spousal attachments that foster both partners’ development, the nurturance of children in non-reciprocal, non-symmetrical attachment relationships in which the adult is the attachment figure;
  5. attachments in later life when the adult is becoming less physically and mentally competent and in need of protection once again.
  • The outcome of this interaction is the organisation of mental and behavioural strategies for protection of the self and progeny, i.e., patterns of attachment.

The Dynamic-Maturational approach hypothesises that, as maturation makes new and more complex mental and behavioural processes possible, changes in context provide the occasion for using these processes (Crittenden, 1998). Specifically, there is a need for maturing individuals to attribute meaning to ambiguous, incomplete, and deceptive information in ways that promote self-protective behaviour; the interaction of maturation with experience provides a basis for this. The particular organisation of self-protective behaviour that develops reflects the strategies that most effectively identify, prevent, and protect the self from the dangers of particular contexts while concurrently promoting exploration of other aspects of life.

Because exposure to danger differs by age as well as by person, family, and cultural group, individuals’ patterns of attachment will reflect:

  • individual developmental history (with individual periods of change in pathway alternating with periods of stability and distributions of patterns differing by developmental period);
  • familial and cultural influences (with the distributions of patterns varying by family and culture).

Training Courses

Crittenden’s training (2003) involves 15 days of full-time effort coding transcripts; the days are usually arranged in 2 or 3 segments. The first 5 days cover most of the Ainsworth-based patterns that form the basis for the Main and Goldwyn system. The second 5 days cover most of the atypical patterns in the dynamic-maturational approach. The third 5 days cover the A/C and AC combinations and the modifiers of patterns, e.g., lack of resolution of trauma and loss, depression, disorganisation or reorganisation.

Course participants read, code and classify two transcripts each day for discussion the following day. The routine, therefore, is 4 hours of training per day followed by independent reading and classifying of the next day’s transcripts. As each transcript takes 2-3 hours to prepare it becomes a 10 hour day while undergoing training.

Following the training, a set of 20 practise transcripts is given to each participant for classification; then 20 transcripts are provided and these transcripts constitute the reliability test, where one needs to get over 85% agreement to pass.

The Dynamic-Maturational method (Crittenden, 1999) for analysing Adult Attachment Interviews differs from the Main and Goldwyn method in several ways (Crittenden, 2003):

  • Intent: The intent of the Dynamic-Maturational method is to describe the self-protective strategies and patterns of mental processing of speakers; the intent of the Main and Goldwyn method is to predict infants’ patterns of attachment retrospectively.
  • Outcome classifications: The set of outcome classifications is larger in the Dynamic-Maturational method and better suited to differentiating among individuals with psychological disorder than the set of classifications used by the Main and Goldwyn method.
  • Treatment of non-Ainsworth classifications: The Dynamic-Maturational method uses 6 compulsive Type A subpatterns (A3-8) and 6 obsessive Type C subpatterns (C3-8), plus a full array of combinations of these. In the Main and Goldwyn method, most non-normative individuals fall in three classifications (E3, U/E3, and “Cannot Classify”).
  • Patterns vs ratings: The Dynamic-Maturational method depends upon patterns within and among memory systems, whereas the Main and Goldwyn system depends upon ratings of constructs.
  • Functions vs defined meanings: The Dynamic-Maturational method uses the function of discourse markers to define meaning, whereas the Main and Goldwyn method assigns meanings to discourse markers.
  • Memory systems: The Dynamic-Maturational method systematically assesses 5 memory systems (procedural, imaged, semantic, episodic, and working memory), whereas the Main and Goldwyn method considers only 3 (semantic, episodic, and working memory).
  • Modifiers: In the Dynamic-Maturational method, there are four modifiers (depressed, disorganised, reorganising, and unresolved with regard to trauma or loss), with 8 different forms of lack of resolution of trauma or loss (dismissed, displaced, vicarious, blocked, preoccupied, anticipated, imagined, and hinted); the Main and Goldwyn method has only preoccupied lack of resolution of loss or trauma.
  • Validity: Validity for the Main and Goldwyn method is primarily based on normative samples and prediction from mothers to infants, whereas the validity of the Dynamic-Maturational method is primarily based on clinical samples and differentiation among disorders. However only a very small number if people have achieved reliability in this system


Implications for Training:

I was totally unaware about the distinctions between the different attachment training courses and theories when I enrolled for training via the NSW Institute of Psychiatry in 1999. I trained in the AAI and SSP with Patricia Crittenden which has been an interesting experience, some of which I will share briefly here. I have now been trying to get reliability (over 85% agreement) in the AAI and SSP for 3 years. Colleagues are also still waiting (3 years on) to sit for the CARE Index.

When I started, 14 people in the world had AAI reliability. Of 15 people in my course there are only 2 left trying to get reliability. Most people have found it too time consuming and difficult to achieve and unfortunately dropped out. I have at least 5 more transcripts to do, of which one is due every fortnight and takes a whole weekend (10 hours) to do, so it is a big time commitment.

With the SSP we were trained in the Ainsworth system, but as it took so long to get the test tapes, Crittenden had by then further developed her manual. So when we all (six of us) sat for reliability in her and Ainsworth’s system in 2003, we all failed the test. At present we are not sure what will happen next. We may be offered to repeat the test we may not be. There is general consensus that if reliability is so difficult to get, is it really a useful research tool?

On the other hand I have learned a lot about theories of attachment and the various attachment classifications. Would I ever use it if I had reliability? Not so much in private practice, but it would be a great tool to have as part of a unit that is doing research. One could use it for court decisions as to whether to remove a child or not. The cost of $4000 was certainly a lot.

Learning more about attachment theory could be achieved by looking at the theory rather than sitting for reliability in one of its assessment measures, yet these courses did not exist in 1999. I have heard that the NSW Institute of Psychiatry now offers an Introduction to Attachment Theory Course. If one wants to do research in Attachment however, reliability will be needed. At present I would advise someone who is interested to check how many participants get reliability per course, and compare Main’s “Graduation” rate to Crittenden’s. It may help in finding the better course / training.


Ainsworth (1967). Infancy in Uganda: Infant Care and the growth of love. Baltimore: Johns Hopkins University Press.

Emde R.N. (1990) in Greenberg,M.T., Cicchetti, D., Cummings, E.M. (Edn) Attachment in the Preschool years: Theory, Research and Intervention. University of Chicago Press: Chicago

Crittenden, P. Website (2003) – Family Relations Institute

Crittenden, P.M. (1998) Adult Attachment Interview: Coding Manual for the Dynamic-Maturational Method. Unpublished manuscript, available from the author (AAI Manual Update)

Gullestad, S.E. (in press). Adult Attachment Interview and psychoanalytic outcome studies. The International Journal of Psychoanalysis.

Crittenden, P.M. (1988-1994) Preschool Assessment of Attachment Manual. Unpublished manuscript available from the author.

Crittenden, P.M. (1992) Quality of Attachment in the preschool years. Development and Psychopathology, 4, 209-241

Crittenden, P.M. (1999) Danger and Development: The Organisation of self-protective strategies. In J. I Vondra and D Barnett (Eds). A typical attachment in infancy and early childhood around children at developmental risk. Monographs of the Society for Research on Child Development (pp145-171)

Crittenden, P.M (1995) Attachment and Psychopathology: In S. Goldberg, R. Muir & J Kerr (Eds). John Bowlby’s attachment theory: Historical, Clinical and Social Significance New York: Analytic Press (pp367-406)


George, C, Kaplan, N & Main, M (1986). Adult Attachment Interview. Unpublished manuscript, Department of Psychology, University of California, Berkeley.

George, C, Kaplan, N & Main, M (1996). Adult Attachment Interview. Unpublished manuscript, Department of Psychology, University of California, Berkeley (third edition).

Main, M & Goldwyn, R (under contract). Adult attachment scoring and classification systems. In M Main (Ed), Assessing attachment through discourse, drawings, and reunion situations (working title). New York: Cambridge University Press.

Main, M & Solomon, J (1990). Procedures for identifying infants as disorganised/disoriented during the Ainsworth Strange Situation. In M Greenberg, D Cicchetti, and E.M. Cummings (Eds), Attachment in the Preschool Years (pp161-182). Chicago: University of Chicago Press