change the world badge

change the world badge


Child Psychology Blogs

Concerned About Unconventional Mental Health Interventions?

Concerned About Unconventional Mental Health Interventions?
Alternative Psychotherapies: Evaluating Unconventional Mental Health Treatments

Friday, September 30, 2011

Authority versus Evidence: Arguing About Adoption and Psychological Interventions

Off and on for a couple of years, but especially in the last few months, I’ve found myself upholding the idea that there must be an evidence basis for treatment, against others who believe that statements by people who claim authority are the most powerful of all arguments. I’ve referred to this conflict as a “culture war”, and I am convinced that it is an argument between ideological, a priori assumptions about the world, on the one hand, and positions based on systematic observations and evidence, on the other-- same old Plato versus Aristotle, if you like.

Von and other adoption bloggers argue strongly for a Primal Wound, on the basis of their own experience and on the authority of Nancy Verrier (who, incidentally, has not yet answered the questions I asked a month or so ago). They reject the systematic research evidence provided by Michael Rutter and the English-Romanian Adoptees Project or by other investigators, showing that most adopted children, even those adopted late and after intense social deprivation, do quite well in the long run. For Von and friends, the vividness of personal experience and the statement of an authority establish a set of assumptions that do not need to be tested against other evidence. For Rutter and other researchers, evidence is to be explored carefully as a test of existing assumptions.

Valle Oberg, a proponent of Ronald Federici’s methods of dealing with post-institutionalized children, also appeals to authority as the foundation of her argument. She states (in comments on this blog) that Federici has worked with thousands of children (although the arithmetic on this does not seem to work out very well) and has “saved” them, and that her own children were among those. Therefore, she argues, what she says, and what Federici has said, must be correct. In addition, she proposes that peer-reviewed publication of outcome research is not evidence that methods are effective. Oberg dismisses the view that Federici needs to report his evidence to the public before his methods are said to be effective.

While mulling over these disagreements, I came across a letter to the editor published in Science in 2004 (a silent testimony to the number of papers on my desk). That LTE was in response to discussion at that time about hormone replacement therapy and the way it failed to provide the benefits to heart health that had been expected of it. The authors, Philip Guzelian and Christopher Guzelian, pointed out that it was not surprising or anomalous that the predicted results did not occur. They commented that the outcome was a “dramatic example of the difference between authority-based conclusions (arising from opinion, experience, intuition, judgment, and scientific inference)… and evidence-based conclusions (derived from an objective, unbiased, and systematic analysis of scientific knowledge)… The lesson is quite generalizable. Uncritical acceptance of authority-based opinions as conclusive evidence is pervasive, even though top authorities unsuccessfully predict what scientific knowledge will be preserved as ‘fact’ “. Guzelian and Guzelian noted that there are times when decisions need to be made without adequate scientific evidence, but warn against confusing them with evidence-based conclusions and propose that “the obvious solution is to explicitly acknowledge when shortcomings in the amounts or quality of evidence necessitate a reversion to authority”.

An important point in the contribution of Guzelian and Guzelian is the acknowledgment that evidence-based conclusions are not always available. Those authors were not talking about psychological interventions, but that acknowledgement is an important one in discussion of psychological treatment, where design and implementation of research can be extraordinarily challenging. Although Guzelian and Guzelian did not mention levels of evidence (the idea that some forms of research offer stronger arguments than others), they did imply the need to balance evidence and authority differently in different situations. When evidence is strong, it should be weighed far more heavily than authority; when evidence is weak or non-existent, authority and personal experience are better to rely on than flipping a coin or casting the I Ching.

Perhaps the most important message in the Guzelians’ letter is the need for explicit statement that in the absence of systematic evidence, one is appealing to authority for support of a claim. This is only appropriate, of course, if there is no evidence or if the existing evidence is weak or open to interpretation, and if the maker of the claim can show that this is the case. It is not sufficient to do as Von, Valle Oberg, and many others have done-- to ignore the existing evidence and put forward instead a contradictory claim based on authority, and not only authority, but the authority whose views are welcome.

But, of course, if your way of thinking is to appeal to authority, this will make sense to you only if stated by an authority of your choice.

Sunday, September 25, 2011

You Can Pick Your Friends, But You Can't Pick Your Monkeys: Origins of Attachment Theory

One of the greatest temptations in the field of psychology is the urge to leap from research on animal behavior to a conclusion about human beings. Unfortunately, unless research about human beings substantiates the conclusion, the landing may be a bumpy and unreliable one. This danger does not always stop leapers; Alan Schore’s “modern attachment theory” is based on a number of such leaps, and as I have pointed out elsewhere, Schore does not make much of an effort to insure that readers realize which animals were the actual subjects of the research he cites.

The history of attachment theory, right back to John Bowlby’s time, has been characterized by generalization from animal behavior to human beings’ child care and early development. Bowlby originally compared human emotional attachment to the “imprinting” of ducks on the first moving objects they saw, and their later following of the imprinted object (usually the mother) and eventual courtship and mating with another member of their own species. Imprinting had an enormous impact on the duck’s life, determining that it might spend its early life following a toy train, or in adulthood “fall in love” with a human caretaker and ignore available mates of its own species. Considering the facts of duck behavior, Bowlby at first thought that human attachment would have similar mechanisms and effects, including “monotropy” (an inability to make an attachment to more than one human being). But further observation showed that attachment did not occur in the first days of life as imprinting did, that human beings had the ability to develop more than one attachment, and that a child who lost the first attachment relationship was capable, under the right circumstances, of developing a strong emotional attachment to a new caregiver.

Bowlby realized that the leap had gone too far and had landed on the wrong spot. Although human attachment was in some ways analogous to imprinting, it was not the same thing, and it was not possible to draw safe conclusions about human development simply from observations of duck behavior.

Bowlby and his colleagues then moved on to consider another type of animal behavior-- this time, an animal much more like human beings than ducks are. As is well known, Bowlby was much interested in the work of the comparative psychologist Harry Harlow on rhesus monkeys. Harlow demonstrated that rhesus babies, when separated from their mothers, sought out soft fabric model “mothers” and clung to them even though those “mothers” offered no nourishment. They would leave the soft mother to drink from a wire-framed “mother” that contained a bottle of milk, then scurry back to the soft model. Both Bowlby and Harlow were taken with the fact that the rhesus babies did not seem “attached” to the models that supplied milk; this result was in contradiction to Freud’s belief that the experience of feeding created emotional attachment, and it supported Bowlby’s argument that there was something in the social experience with the caregiver that triggered the development of attachment. Bowlby and Harlow were also interested in the later difficulties of the rhesus babies as adults, when they were reluctant to mate and did a very poor job of caring for their own infants if they had them. This outcome seemed to suggest that attachment experiences determined a great deal about personality development, an idea that was later built into Bowlby’s attachment theory and that has informed much of the later research on attachment.

Here’s an important question, though: What if Harlow had chosen some other kind of monkey? This was an issue which in fact received much attention from psychologists and zoologists in the couple of decades following Harlow’s reports and Bowlby’s adoption of the Harlow studies as evidence for attachment theory. The discussion at the time-- largely forgotten now that everyone has seen films of sad-looking rhesus babies on wire “mothers”-- focused on the fact that monkeys exhibit species differences in maternal-infant behavior. A monkey is not “just a monkey”, but a particular kind of monkey. (Much less work on these differences is done today because of the great expense of field studies and current concerns about the ethical treatment of lab animals.)

Harlow worked with rhesus monkeys, Macaca mulatta, a monkey that has frequently been employed in various types of research and which gave its common name to the Rh factor well-known to influence human reproductive success. But other researchers pointed out that the results of behavioral work with rhesus monkeys might not be the same as similar work with other types of monkeys. In one review article (Seay & Gottfried [1975]. A phylogenetic perspective for social behavior in primates. Journal of General Psychology, 75, 5-17), the authors described typical maternal-infant behavior and responses to separation in several monkey species in addition to Macaca mulatta.

Seay and Gottfried noted that in spite of the developmental difficulties of the “surrogate”-reared rhesus infants, rhesus monkeys who had somewhat less deprivation all eventually did well, whether they were reared in small groups of infants, reared alone but had a play period with other infants daily, were reared by brutal or abusive mothers, or had normal rearing by their mothers. They concluded that only an extremely unfavorable environment would alter these animals’ play, aggression, or sexual behavior. In addition, they stated that “even the organism maturing in an environment inadequate to support species-typical behavior will not develop unique techniques for coping with that environment. Rather, his behavioral repertoire will consist of fragments of the behavior patterns which ensure individual and species survival for most of his species-mates” (p. 10).

Comparing the rhesus monkey to other species, these authors noted that their own work on a different monkey, Macaca fascicularis, showed results rather similar to that on Macaca mulatta. However, the African red monkey (patas) showed a different pattern of social behavior. If rhesus monkeys were housed in cages such that the infants could go out to a central playground, when one came out, one or more others would join him. When an infant patas monkey came out, though, other patas infants did not join him—or, if one did, the first baby went back to his mother. As patas monkeys got older, they played chasing games, but unlike rhesus monkeys did not contact each other much.

Rhesus mothers object to physical disturbance from their infants, and will bite or hit an annoying baby. However, they will allow “strange” babies to come into the cage, as long as they “behave”. Patas mothers allow their babies to explore or play-attack them, and either do not respond or respond playfully-- but they will not allow a strange baby to come in. When nursing, rhesus babies are belly-to-belly with their mothers; patas babies do not have much of the body surface touching the mother. All monkey babies are disturbed by separation from the mother, but patas babies recover more quickly than rhesus babies do.

Who are we humans like? Should we accept a theory of personality development based on rhesus monkeys, or are we more like patas monkeys? The best guess is that we are like ourselves. Like other species, we have our own patterns of social interaction and of response to environmental changes. We may resemble other species, and their behavior may seem familiar and appealing to us. But when we come down to it, the behavior and development of other species can only provide a hint about what human beings might be like. We need to be cautious about our leaps from animal evidence to theories about humans--- however, cute, sad, and sympathetic those baby monkeys look to us.

Saturday, September 17, 2011

What is Reactive Attachment Disorder? Further Discussion

I’ve made a number of statements to the effect that violent or aggressive behavior, lying, stealing, and so on, were not diagnostic of Reactive Attachment Disorder. Some readers disagree strongly with this perspective. Reader Valle Oberg, for example, has been arguing against my position for several days at

I’d like to call people’s attention to a paper by Charles H. Zeanah and Anna Smyke from Tulane University, arguably the leading U.S authorities on Reactive Attachment Disorder (Zeanah, C.H., & Smyke, A.T. [2008]. Attachment disorders in family and social context. Infant Mental Health Journal, 29, 219-233). This paper discusses criteria for attachment disorders and compares the DSM and ICD-10 versions, as well as referring to categories suggested by Zeanah in previous work.

In none of this material is there any reference to violent or aggressive behavior, self-injury, lying, cheating , stealing, refusing eye contact, or any of these issues so much stressed by therapists and parent groups outside the mainstream of psychological and psychiatric thought. Zeanah and Smyke come no nearer to these issues than referring briefly to the possibility of risk-taking as one aspect of a RAD-like category.

Zeanah and Smyke discuss measures of inappropriate behavior suggesting a disorder of attachment. They point out three ways in which such measurement has been approached. One method looked at whether children wandered away from caregivers without becoming distressed, whether they approached strangers, whether they were never shy with new adults, whether they were friendly with new adults, and whether they would go off with strangers. A second approach considered whether children failed to differentiate among adults (that is, treated all adults the same way and did not have a preferred caregiver), readily went with a stranger, and failed to check back with a caregiver (i.e., by looking back to them or calling to them as distance between them increased or separation became likely). The third method looked at not having a preferred caregiver, lack of reticence with a stranger, failure to check back, and willingness to go with a stranger. (Once again, none of these methods looks at aggressive behavior, lying, self-injury, etc., etc.)

Zeanah and his colleagues several years ago wanted to develop a technique of assessing preschool children’s attachment to caregivers without depending on parent or teacher reports. They developed what they called the “Stranger at the Door” procedure. One of the researchers knocked at the door of the child’s home, and the caregiver and the child answered the door together. The stranger looked at the child and said, “My name is ____. What is your name? Let’s go for a walk.” Observers coded the child’s behavior in response (if the child was willing to go, they walked a few feet, then came back into the house). Children’s willingness to go with the stranger was greatest in the group who had been institutionalized, but there was also atypical willingness in a comparison group of children who had been in foster care.

The Zeanah and Smyke paper noted that in a group of children adopted from institutions, there were no cases of the inhibited form of RAD, but a “substantial minority” showed the disinhibited type, with less avoidance of strangers and less preference for familiar caregivers than is typical among family-reared children.

In discussing “self-endangering” behavior as an aspect of RAD (the nearest thing mentioned to the claims of aggression and self-injury made by some advocates of an unorthodox view of attachment disorders), Zeanah and Smyke emphasize the possibility that risk-taking is part of a “two person” disorder which is difficult to describe in present language. They raise the issue of possible relational attachment disorders (different from RAD as now defined) that are evident only in the context of the association between the child and a specific caregiver. I would add that this is a very interesting idea with respect to the claims made by members of ATTACh and similar groups that children with RAD behave angelically outside their homes, but are difficult or even dangerous in interaction with their mothers.

Monday, September 12, 2011

What Is Attachment?

What IS attachment, anyway? This word is used constantly by parents and teachers, attorneys and judges, as well as psychologists and social workers. What do they mean? Is attachment one of those things like obscenity, that we can’t define, but “we know it when we see it”?

Actually, one of the problems with understanding attachment is that many people don’t know it when they see it. As a discussion of attachment and child custody decisions pointed out a couple of years ago, “few custody evaluators recognized the distinction between a highly dependent or demonstratively affectionate relationship between a parent and a child and a secure attachment” (Calloway & Erard [2009]. Introduction to the special issue on attachment and child custody. Journal of Child Custody, 6, 1-7). Some quasi-professionals who emphasize attachment believe that the attachment of a child to a caregiver is indicated by physical demonstrations of affection, eye contact “on the parent’s terms”, obedience, and gratitude.

Many textbooks define attachment as an “emotional bond” or “tie” between two people, especially a parent and a child. The organization ATTACh (Association for the Treatment and Training of Attachment in Children) gives the following lengthier definition of attachment:
“Attachment is a reciprocal process by which an emotional connection develops between an infant and his/her primary caregiver. It influences the child’s physical, neurological, cognitive, and psychological development. It becomes the basis for development of basic trust or mistrust, and shapes how the child will relate to the world, learn, and form relationships throughout life.” (

But let’s parse these definitions. Is it of any help to say that something is an “emotional bond”, without saying what an emotional bond is? Actual bonds involve close physical proximity, enforced by physical methods like ropes or handcuffs. Attachment, then, is being said to be like a bond in that it maintains proximity, but the way it does this is through emotional factors rather than with ropes. This is an analogy rather than a definition, and it is inadequate for two reasons. One is that although attachment does maintain a young child’s physical proximity to a caregiver, continuing development alters the way in which attachment is expressed. In any case, the seeking of proximity is not the same at all times, but occurs when the child feels threatened or uncomfortable. The other problem is that this description of attachment leaves it unclear whether both members of the pair behave in the same way or have the same emotions or motivations relative to maintaining proximity. In fact, in the course of development, parent and child trade several times the position of being the one that wants more proximity; the child who has a toddler clung to the parent who was leaving for work grows to the teenager whose parent is anxious to know where he is going and when he will be back

Now let’s look at the ATTACh version. It’s longer and more complicated, but is it any better? It begins by saying that “attachment is a … process”. No doubt there are processes leading to attachment, but if attachment is a process, it’s difficult to see how we can determine whether attachment is secure or even present. Surely attachment is some sort of end state, and not simply an ongoing process. Perhaps the “emotional connection” which is said to result from the attachment process is the actual attachment, but it too remains undefined.

ATTACh also says that attachment is not just a process, but a “reciprocal process”. I interpret this as meaning that attachment develops in the course of interactions with other people and could not exist in a social vacuum, and this is certainly correct. ATTACh may also intend to say that attachment (emotional connection?) in a child develops in tandem with some emotional change in the adult caregiver, which may be true, but seems to be irrelevant to the later emphasis placed on the child’s own development.

ATTACh speaks of connections between an infant and the primary caregiver and appears to be limiting attachment to this relationship. Here we encounter some problems, because all the evidence is that young children can have many attachment relationships, some more obvious than others, and that these relationships may or may not share qualities like security. On the other hand, caregivers tend to be monotropic and have one or two preferred infants, even if they work with a group of children of similar ages; perhaps the ATTACh authors were thinking of the adult’s connection with a “primary” baby when they wrote this.

ATTACh goes on to cite the various aspects of development that are influenced by attachment (although whether as a process or as a loosely-defined emotional connection is not clear). It is not stated how attachment affects the child’s physical development, or how psychological development is different from cognitive development, but even if these points are set aside, these remarks do not contribute to the definition of attachment, but simply say more about things that are related to the undefined process or end state. Similarly, in the last sentence, references to basic trust and to shaping relationships do nothing more than to say that something called attachment is involved here. (As to the statement about learning, it is far from clear what the authors meant.)

It would seem that neither the textbook definition nor the ATTACh effort is of much use in clarifying what attachment actually means. I’d like to offer an alternative definition that I believe cuts through some of the difficulties we’ve seen here.

I would suggest that attachment is best thought of as an attitude, or a readiness to behave in particular ways toward the object of the attitude. It is an attitude toward human beings that involves the readiness to behave differently toward familiar and unfamiliar people, but to do so in different ways when there is perceived threat than when there is not. Like other attitudes, attachment changes with development, and although young children show their attachment by seeking proximity to familiar people when uncomfortable, they gradually move toward other related behaviors. Also, like other attitudes, attachment involves characteristic emotional responses that may occur together with or instead of observable behavior, and these too change with development. Attachment attitudes vary with experience as well as with age and culminate in an internal working model of human relationships in general as well as the specific reaction to threat which is their foundation.

This may not be the perfect definition, but at least it actually is a definition. Comments and editing are invited.

Sunday, September 11, 2011

Infant and Toddler Overnights: A Focus of Divorce Disagreements

When couples have children of infant and toddler age, the breakdown of marriage is often followed by intense and bitter argument over allowing the child to have an overnight visit and sleep away from the familiar home. In most cases, the mother is the custodial parent and continues to live in the marital home, while the father is more likely to be the one who moves out, lives elsewhere, and does the “visiting”.

Mothers are very likely to resist overnight visits for young children. They are concerned that the father is unaware of the child’s bedtime rituals or needs, that the child may be distressed by the experience, or even that a father who has behaved violently in the past may lose his temper if the child cries or fusses in the unfamiliar night-time setting. Fathers, on the other hand, may be concerned that the child spends less time with them than with the mothers and may therefore be less attached to the fathers. They may also believe that if the child does not form a strong attachment to the father in the first couple of years, no real relationship will ever be possible.

However the estranged couple work things out, this problem will pass (and of course will be replaced by different issues). The child will get older and reasons to resist overnight visits will gradually disappear. At the beginning, though, the overnight issue seems to be of overwhelming importance. In a recent special issue of the journal Family Court Review (2011, Vol. 49), the issue editor noted that when she surveyed readers to ask what they wanted to know about attachment and child custody, 60% of them asked what to do about overnights.

Several well-known psychiatrists and psychologists who contributed to the issue made a point of commenting on the overnight question. And all of them said about the same thing: overnight visits are problematic for about the first two years of the child’s life. Depending on the child, it may be best to wait as late as age 4 before beginning overnights (George, Solomon & McIntosh [2011]. Divorce in the nursery: On infants and overnight care. Family Court Review, 49, 521-528). The reasoning behind this advice is that children have real individual differences in temperament (for instance, the ease with which they accept a new situation) and in language development (which allows them to understand what they are told about the length of a visit). Children who are easily distressed, who have little concept of time, and who are still immature in their understanding of language may interpret an overnight visit as permanent abandonment by the custodial parent and may not be able to communicate their fears to the other parent-- who may have no wish to hear of the child’s longing for the divorced spouse. When children have handicapping conditions that affect communication and cognition, their level of development may be much more important than their chronological age.

But what if there are no overnights? Does this mean that the child’s relationship to the noncustodial parent is negated from the beginning? Charles Zeanah, the well-known attachment researcher, commented about this: “it’s not necessary for both parents to have attachment relationships with the child in the early years. That can happen later, when the child has more sophisticated abilities to sustain attachment relationships over time and place. Where it is possible, it does make sense for the child to keep contact at the level of comfort and familiarity until they are ready for more” (Lieberman, Zeanah, & McIntosh [2011]. Attachment perspectives on domestic violence and the law. Family Court Review,49, 529-538). Alan Sroufe, another leading attachment researcher, said, “I think that if parents, judges, lawyers, and so on took the view that attachment is a gradual building process, and that each relationship is built on its own terms, there would be less paranoia about this… even if they had no overnights for the first 2 years” (Sroufe & McIntosh [2011]. Divorce and attachment relationships: The longitudinal journey. Family Court Review, 49, 464-473).

These commentators are stressing the best interests of the child, a vague standard indeed, but one which we can clearly distinguish from an emphasis on parents’ rights. They seem to agree that it’s in the child’s best interest to have as calm an infancy and toddlerhood as possible and not to be asked to adjust to unnecessary changes and transitions. They also agree that two caring parents-- even if separated-- are better than one, but that there is no natural process that demands that both sets of relationships must develop simultaneously and early on. The fact that no “window of attachment” closes at age 2 means that it’s possible both to preserve family connections and to support a child’s early needs for a peaceful life.

Separated parents with very young children would do well to take these cautions to heart, but at the same time they may want to consider whether they want to try out visits with an eye to overnights. A gradual approach, including nap-time at the noncustodial residence, may give them an idea whether a child is ready to take the step to an overnight stay. If the child copes well with one overnight visit, it is wise to wait a while for the next one rather than rushing to a new schedule of frequent changes and risking overwhelming him or her. But none of this will work well unless the parents are able to concentrate on the child’s needs and reactions rather than hurrying to blame each other for any distress the child shows.

When overnight visits begin, one point that can be difficult for separated parents to deal with is co-sleeping. Plenty of parents let infants and toddlers sleep with them part or all of the time (and I am far from criticizing this practice). In addition, if a father leaves the household, the mother is likely to respond to the child’s distress and her own by letting the child sleep in her bed. What happens, then, if the child stays overnight with the father? Can the co-sleeping mother deal with having the child share a bed with the father, or will this trigger concerns about sexual behavior? How does the father comfort the child who is accustomed to sharing a bed, except by co-sleeping? This highly emotional situation is one in which straightforward discussion with neutral parties is essential in order to avoid distress, fears, and accusations that are in the great majority of cases completely unrealistic.

P.S. The comments of Alan Sroufe about attachment as a gradual building process are also highly relevant to adoption issues.

P.P. S. The issue of Family Court Review I'm referring to is available for free on line. Google the journal name and you'll see how to get to the articles.