attachment ao3

Caregiver-infant interactions: Filmed data

One strength of the research on this topic is that caregiver infant interactions are usually filmed in a laboratory. This means that other activity, that might distract a baby, can be controlled. Also, using films means that observations can be recorded and analysed later, therefore it is unlikely that researchers will miss seeing key behaviours. Having filmed interactions also means that more than one researcher can record data and establish inter-rater reliability of observations. Finally, babies don't know they are being observed, so their behavior does not change in response to observation. Therefore the data collected in such research should have good reliability and validity.

Caregiver-infant interactions: Interpreting babies behaviour

One limitation of research is that it is hard to interpret a baby’s behaviour. Young babies lack co-ordination and much of their bodies are almost immobile, the movements being observed are just small hand movements or subtle changes in expression. It is difficult to be sure, for example, whether a baby is smiling or just farting. It’s also difficult to determine what is taking place from the babies perspective. For example, we cannot know whether a movement such as a hand twitch is random or triggered by something the caregiver has done. This means that we cannot be certain that the behaviours seen in caregiver-infant interactions have a special meaning.

Caregiver-infant interaction: Feldman + counter

A further limitation is that simply observing a behaviour does not tell us its developmental importance. Feldman (2012) points out that ideas like synchrony and reciprocity simply gives names to patterns of observable caregiver and baby behaviours. They can be reliably observed, but they still may not be particularly useful in understanding child development as it does not tell us the importance or purpose of these behaviors. This means that we cannot be certain from observational research alone that reciprocity and synchrony are important for a child’s development. 

However, there is evidence from other lines of research to suggest that early interactions are important. For example, Isabella et al (1989) found that achievement of international synchrony predicted the development of a good quality attachment. This means that, on balance, caregiver infant interaction is important in development. 

Caregiver-infant interaction: Papusek

Papusek et al (1991) Showed that the tendency to produce caregiverese (high pitched baby talk) is common across American, Chinese, and German mothers. This suggests that some aspects of caregiver-infant interactions in the development of attachment are not culturally biased but are universal to all or many cultures and thus may be innate within infants and caregivers

Stages of attachment: External validity + counter

One strength of Schaffer and Emerson’s research is that it has good external validity. Most of the observations were made by parents during ordinary activities and reported to the researchers. The alternative would have been to have researchers present to record observations. This might have distracted the babies or made them feel anxious. This means it is likely that the babies behaved naturally while being observed. 

However, as it was self report from mothers, it’s unlikely that they were objective observers. They may have been biased in terms of what they noticed and what they reported due to a social desirability bias. For example, they may exaggerate separation anxiety to present their attachment to their baby as stronger. This means that even if babies had behaved naturally their behaviour may not have been accurately recorded.

Stages of attachment: Validity of measures

One limitation of Schaffer and Emerson’s stages is the validity of the measures they used to assess attachment in the asocial stage. Young babies have poor co-ordination and are fairly immobile. If babies less than two months old felt anxiety in everyday situations they might have displayed this in subtle, hard to observe ways. This made it difficult for mothers to observe and report back to researchers on signs of anxiety and attachment in this age group. This means that babies may actually be quite social but because of flawed methods they appear asocial.

Stages of attachment: RWA

Another strength of Schaffer and Emerson;s stages is that they have practical application in day care where babies are cared for outside of their homes by a non-family adult. In the asocial and discriminate attachment stages day care is likely to be straightforward as babies can be comforted by any skilled adult. However, Schaffer and Emerson’s research tells us that day care, especially starting day care with an unfamiliar adult, may be problematic during the specific attachment stage. This means that parents’ use of day care can be planned using Scaffer and Emerson’s stages.

Stages of attachment: External and temporal validity

This study used only white, working class Scottish babies from Glasgow, so cannot be generalised to babies of other classes or cultures (or even babies living in other cities in Scotland) so lacks external validity.

Study was carried out in the 1960s so now may lack temporal validity as childrearing practices have changed significantly over the last 50 years.

Role of the father: Clarity

One limitation of research into the role of fathers is the lack of clarity over the question being asked. The question ‘What is the role of the father?’ In the context of attachment, is complicated. Some researchers attempting to answer this question want to actually understand the role of the fathers as a secondary attachment figure. But others are concerned with fathers as a primary attachment figure. The former have tended to see fathers as behaving differently from mothers and having a distinct role. The latter have found that fathers can take on a ‘maternal’ role. This makes it difficult to offer a simple answer to the role of the father as it depends on what specific role is being discussed

Role of the father: Methodology + counter

A further limitation of research into the role of fathers is that findings vary according to the methodology used. Longitudinal studies such as that of Grossmann et al have suggested that fathers as secondary attachment figures have an important and distinct role in their children’s development, involving play and stimulation. However, if fathers have a distinctive and important role we would expect that children growing up in a single mother and lesbian families would turn out differently from two parent heterosexual families. In fact studies consistently show that these children do not develop differently from children in two parent heterosexual families. This means that the question as to whether fathers have a distinctive role remains unanswered. 

However, these lines of research may not in fact be in conflict. It could be that fathers typically take on distinctive roles in two-parent heterosexual families, but that parents in single-mother and lesbian families simply adapt to accommodate the role played by fathers. This means that the question of a distinctive role for fathers is clear after all. When present, fathers tend to adopt a distinctive role, but families can adapt to not having a father.

Role of the father: RWA

One strength of research into the role of the father is that it can be used to offer advice to parents. Parents and prospective parents sometimes agonise over decisions like who should take on the primary caregiver role. For some this can even mean worrying about whether they should have children at all. Mothers may feel pressured to stay at home because of stereotypical views of mothers and fathers roles. Equally, fathers may be pressured to focus on work rather than parenting. In some families, this may not be economically the best decision. Research into the role of the father can be used to offer reassuring advice to parents. For example, heterosexual parents can be reassured that fathers can become a primary attachment figure and lesbian parents can be informed that not having a father around does not affect a child’s development

Lorenz: Regolin and Vallortigara

One strength of Lorenz’s research is the existence of support for the concept of imprinting. A study by Regolin and Vallortigara (1995) supports Lorenz’ idea of imprinting. Chicks were exposed to simple shape combinations that move, such as a triangle with a rectangle in front. A range of shape combinations then moved in front of them and they followed the original most closely. This supports the view that young animals are born with an innate mechanism to imprint on a moving object present in the critical window of development.

Lorenz: Generalisability

One limitation of Lorenz’ studies is the ability to generalise findings and conclusions from birds to humans. The mammalian attachment system is quite different and more complex than that in birds. For example, in mammals attachment is a two way process and mothers also show an emotional attachment to their young, This means that it is probably not appropriate to generalise Lorenz’s ideas to humans.

Harlow: RWA

One strength of Harlow’s research is its important real world applications. For example, it has helped social workers and clinical psychologists understand that a lack of bonding experience may become a risk factor in child development allowing them to intervene to prevent poor outcomes. We also now understand the importance of attachment figures for baby monkeys in zoos and breeding programmes in the wild. This means that Harlow’s research has both theoretical and practical value.

Harlow: Generalisability

One limitation is the ability to generalise findings from monkeys to humans. Rhesus monkeys are much more similar to humans than Lorenz’ birds, and all mammals share some common attachment behaviours. However, human brains and human behaviour is still much more complex than that of a monkey. This means that it may not be appropriate to generalize Harlow’s findings to humans.

Learning theory: Animal studies

One limitation of learning theory explanations for attachment is lack of support from studies conducted on animals. For example, Lorenz’s geese imprinted on the first moving object they saw regardless of whether this object was associated with food. Also, if we consider Harlow’s research with monkeys, there is no support for the importance of food. When given a choice, Harlow’s monkeys displayed attachment behaviour towards a soft surrogate mother in preference to a wire one which provided milk. This shows that factors other than association with food are important in the formation of attachments.

Learning theory: Schaffer and Emerson

A further limitation of learning theory explanations is lack of support from studies of human babies. For example, Schaffer and Emerson (1964) found that babies tended to form their main attachment to their mother regardless of whether she was the one who usually fed them. In another study, Isabella et al found that high levels of interaction synchrony meant better quality attachment, these factors have nothing to do with feeding. This again suggests that good is not the main factor  in the formation of human attachments.

Learning theory: Elements of conditioning + counter

One strength of learning theory is that elements of conditioning could be involved in some aspects of attachment. It seems unlikely that association with food plays a central role in attachment, but conditioning may still play a role. For example, a baby may associate feeling warm and comfortable with the presence of a particular adult and this may influence the baby’s choice of their main attachment figure. This means that the theory may still be useful in understanding the development of attachments. 

However, both classical and operant conditioning explanations see the baby playing a relatively passive role in attachment development, simply responding to associations with comfort or reward. In fact research shows that babies take a very active role in the interactions that produce attachment (Feldman and Eidelman). This means that conditioning may not be an adequate explanation of any aspect of attachmen

Bowlby monotropy: Schaffer and Emerson

One limitation of Bowlby's theory is that the concept of monotropy lacks validity. Rudolph Schaffer and Peggy Emerson (1964) found that although most babies did attach to one person at first, a significant minority formed multiple attachments at the same time. Also, although the first attachment does appear to have a particularly strong influence on later behaviour, this may simply mean it is stronger, not necessarily different in quality from the child's other attachments. For example other attachments to family members provide all the same key qualities (emotional support, a safe base etc). This means that Bowlby may be incorrect that there is a unique quality and importance to the child's primary attachment.

Bowbly monotropy: Social releasers

One strength of Bowlby's theory is the evidence supporting the role of social releasers. There is clear evidence that cute baby behaviours are designed to elicit interaction from caregivers. T. Berry Brazelton et al. (1975) observed babies trigger interactions with adults using social releasers. The researchers then instructed the babies' primary attachment figures to ignore their babies' social releasers. Babies (who were previously shown to be normally responsive) became increasingly distressed and some eventually curled up and lay motionless. This illustrates the role of social releasers in emotional development and suggests that they are important in the process of attachment development.

Bowlby monotropy: IWM support + counter

A further strength of Bowlby’s theory is support for the internal working model. The idea of the internal working model predicts that patterns of attachment will be passed from one generation to the next. Heidi Bailey et al. (2007) assessed attachment relationships in 99 mothers and their one-year-old babies. The researchers measured the mothers' attachment to their own primary attachment figures (i.e. their parents). The researchers also assessed the attachment quality of the babies. They found that mothers with poor attachment to their own primary attachment figures were more likely to have poorly attached babies. This supports Bowlby's idea that mothers' ability to form attachments to their babies is influenced by their internal working models (which in turn comes from their own early attachment experiences). 

However, there are probably other important influences on social development. For example some psychologists believe that genetic differences in anxiety and sociability affect social behaviour in both babies and adults. These differences could also impact on their parenting ability (Komnienko 2016). This means that Bowlby may have overstated the importance of the internal working model in social behaviour and parenting at the expense of other factors.

Strange situation: Later development + counter

One strength of the Strange Situation is that its outcome predicts a number of aspects of the baby's later development. A large body of research has shown that babies and toddlers assessed as Type B (secure) tend to have better outcomes than others, both in later childhood and in adulthood. In childhood this includes better achievement in school and less involvement in bullying (McCormick et al. 2016, Kokkinos 2007). Securely attached babies also tend to go on to have better mental health in adulthood (Ward et al. 2006). Those babies assessed as having insecure-resistant attachment and those not falling into Types A, B or C tend to have the worst outcomes. This suggests that the Strange Situation measures something real and meaningful in a baby's development.

However, the Strange Situation clearly measures something important that is associated with later development. However, not all psychologists believe this something is attachment. For example, Jerome Kagan (1982) suggested that genetically-influenced anxiety levels could account for variations in attachment behaviour in the Strange Situation and later development. This means that the Strange Situation may not actually measure attachment.

Strange situation: Inter-rater reliability

A further strength of the Strange Situation is good inter-rater reliability (the agreement between different observers). Johanna Bick et al. (2012) tested inter-rater reliability for the Strange Situation for a team of trained observers and found agreement on attachment type in 94% of cases. This high level of reliability may be because the procedure takes place under controlled conditions and because behaviours (such as proximity-seeking and stranger anxiety) involve large movements and are therefore easy to observe. For example, anxious babies cry and crawl away from strangers. This means that attachment type as assessed by the Strange Situation does not depend on subjective judgements.

Strange situation: Cultures

One limitation of the Strange Situation is that it may not be a valid measure of attachment in different cultural contexts. The Strange Situation was developed in Britain and the US. It may be culture-bound, i.e. only valid for use in certain cultures (in this case Europe and the US). One reason for this is that babies have different experiences in different cultures and these experiences may affect their responses to the Strange Situation. For example, in one Japanese study by Keiko Takahashi (1986), babies displayed very high levels of separation anxiety and so a disproportionate number were classified as insecure-resistant. Takahashi (1990) suggests that this anxiety response was not due to high rates of attachment insecurity but to the unusual nature of the experience in Japan where mother-baby separation is very rare. This means that it is very difficult to know what the Strange Situation is measuring when used outside Europe and the US.

Cultural variations: Indigenous psychologists + counter

One strength of the research is that most of the studies were conducted by indigenous psychologists. Indigenous psychologists are those from the same cultural background as the participants. For example, van Uzendoorn and Kroonenberg included research by a German team (Grossmann 1981) and Keiko Takahashi (1986) who is Japanese. This kind of research means that many of the potential problems in cross-cultural research can be avoided, such as researchers' misunderstandings of the language used by participants or having difficulty communicating instructions to them. Difficulties can also include bias because of one nation's stereotypes of another. This means there is an excellent chance that researchers and participants communicated successfully, enhancing the validity of the data collected.

 However, this has not been true of all cross-cultural attachment research. For example, Morelli and Tronick (1991) were outsiders from America when they studied child-rearing and patterns of attachment in the Efé of Zaire. Their data might have been affected by difficulties in gathering data from participants outside their own culture. This means that the data from some countries might have been affected by bias and difficulty in cross-cultural communication.

Cultural variations: Confounding variables

One limitation of cross-cultural research is the impact of confounding variables on findings. Studies conducted in different countries are not usually matched for methodology when they are compared in reviews or meta-analyses. Sample characteristics such as poverty, social class and urban/rural make-up can confound results as can the age of participants studied in different countries. Environmental variables might also differ between studies and confound results. For example the size of the room and the availability of interesting toys there - babies might appear to explore more in studies conducted in small rooms with attractive toys compared to large, bare rooms. Less visible proximity-seeking because of room size might make a child more likely to be classified as avoidant. This means that attachment behaviour in different non-matched studies conducted in different countries may not tell us anything about cross-cultural patterns of attachment.

Cultural variations: Imposing culture bound tests

A further limitation of cross-cultural research is in trying to impose a test designed for one cultural context to another context. Cross-cultural psychology includes the ideas of emic (cultural uniqueness) and etic (cross-cultural universality). Imposed etic occurs when we assume an idea or technique that works in one cultural context will work in another. An example of this in attachment research is in the use of babies' response to reunion with the caregiver in the Strange Situation. In Britain and the US, lack of affection on reunion may indicate an avoidant attachment. But in Germany such behaviour would be more likely interpreted as independence rather than insecurity. Therefore that part of the Strange Situation may not work in Germany. This means that the behaviours measured by the Strange Situation may not have the same meanings in different cultural contexts, and comparing them across cultures is meaningless

Bowlby maternal deprivation: 44 thieves + counter

One limitation of the theory of maternal deprivation is the poor quality of the evidence it is based on. Bowlby's 44 thieves study is flawed because it was Bowlby himself who carried out both the family interviews and the assessments for affectionless psychopathy. This left him open to bias because he knew in advance which teenagers he expected to show signs of psychopathy. Other sources of evidence were equally flawed. For example, Bowlby was also influenced by the findings of Goldfarb's (1943) research on the development of deprived children in wartime orphanages. This study has problems of confounding variables because the children in Goldfarb's study had experienced early trauma and institutional care as well as prolonged separation from their primary caregivers. This means that Bowlby’s original sources of evidence for maternal deprivation had serious flaws and would not be taken seriously as evidence nowadays. 

However, a new line of research has provided some modest support for the idea that maternal deprivation can have long-term effects. Frederic Lévy et al. (2003) showed that separating baby rats from their mother for as little as a day had a permanent effect on their social development though not other aspects of development. This means that there are other sources of evidence for his ideas.

Bowlby maternal deprivation: Rutter

Another limitation of Bowlby's theory of maternal deprivation is his confusion between different types of early experience. Michael Rutter (1981) drew an important distinction between two types of early negative experience. Deprivation strictly refers to the loss of the primary attachment figure after attachment has developed. On the other hand privation is the failure to form any attachment in the first place - this may take place when children are brought up in institutional care. Rutter pointed out that the severe long-term damage Bowlby associated with deprivation is actually more likely to be the result of privation. So the children studied by Goldfarb may actually have been 'prived' rather than deprived. Similarly, many of the children in the 44 thieves study had disrupted early lives (e.g. spells in hospital) and may never have formed strong attachments. This means that Bowlby may have overestimated the seriousness of the effects of deprivation in children's development

Bowlby maternal deprivation: Critical period

A further limitation of the theory is Bowlby's idea of a critical period. For Bowlby, damage was inevitable if a child had not formed an attachment in the first two-and-a-half years of life. Hence this is a critical period. However, there is evidence to suggest that in many cases good quality aftercare can prevent most or all of this damage. For example Jarmila Koluchová (1976) reported the case of the Czech Twins. The twins experienced very severe physical and emotional abuse from the age of 18 months up until they were seven years old. Although they were severely damaged emotionally by their experience they received excellent care and by their teens they had recovered fully. This means that lasting harm is not inevitable even in cases of severe privation. The 'critical period' is therefore better seen as a 'sensitive period’.

Romanian oprhan studies: RWA

One strength of the Romanian orphanage studies is their application to improve conditions for children growing up outside their family home. Studying the Romanian orphans has improved psychologists' understanding of the effects of early institutional care and how to prevent the worst of these effects (Langton 2006). This has led to improvements in the conditions experienced by looked-after children, i.e. children growing up in the care system. For example children's homes now avoid having large numbers of caregivers for each child. Instead the children tend to have one or two 'key workers' who play a central role in their emotional care. Also institutional care is now seen as an undesirable option for looked-after children. Considerable effort is made to accommodate such children in foster care or to have them adopted instead. This means that children in institutional care have a chance to develop normal attachments and disinhibited attachment is avoided.

Romanian orphan studies: Confounding variables

Another strength of the Romanian studies is the lack of confounding variables. There were many orphan studies before the Romanian orphans became available to study (e.g. orphans studied during the Second World War). Many of the children studied in orphanages had experienced varying degrees of trauma, and it is difficult to disentangle the effects of neglect, physical abuse and bereavement from those of institutional care. However the children from Romanian orphanages had, in the main, had been handed over by loving parents who could not afford to keep them. This means that results were much less likely to be confounded by other negative early experiences (higher internal validity). On the other hand, studying children from Romanian orphanages might have introduced different confounding variables. The quality of care in these institutions was remarkably poor, with children receiving very little intellectual stimulation or comfort. This means that the harmful effects seen in studies of Romanian orphans may represent the effects of poor institutional care rather than institutional care per se.

Romanian orphan studies: Adult development

One limitation of the Romanian orphanage studies is the current lack of data on adult development. The latest data from the ERA Study looked at the children in their early- to mid- 20s. This means that we do not currently have data to answer some of the most interesting research questions about the long-term effects of early institutional care. These research questions include the lifetime prevalence of mental health problems and participants' success in forming and maintaining adult romantic and parental relationships. It will take a long time to gather this data because of the longitudinal design of the study, i.e. the same participants are followed over a long period. This means it will be some time before we know more completely what the long- term effects are for the Romanian orphans. It is possible that late-adopted children may 'catch up’

Influence of early attachment: Supporting evidence + counter

One strength of the research into attachment and later relationships is supporting evidence. We have looked at studies linking attachment to later development. Reviews of such evidence (e.g. Fearon and Roisman 2017) have concluded that early attachment consistently predicts later attachment, emotional well-being and attachment to own children. How strong the relationship is between early attachment type and later development depends both on the attachment type and the aspect of later development. So whilst insecure-avoidant attachment seems to convey fairly mild disadvantages for any aspect of development, disorganised attachment is strongly associated with later mental disorder. This means that secure attachment as a baby appears to convey advantages for future development while disorganised attachment appears to seriously disadvantage children. 

However, not all evidence supports the close links between early attachment and later development. For example the Regensburg longitudinal study (Becker-Stoll et al. 2008) followed 43 individuals from one year of age. At age 16 attachment was assessed using the adult attachment interview and there was no evidence of continuity. This means that it is not clear to what extent the quality of early attachment really predicts later development. There may be other important factors.

Influence of early attachment: Retrospective assesment

One limitation of most research into the influence of attachment is that early attachment is assessed retrospectively. Most research on the link between early attachment and later development are not longitudinal (ie. they don't assess attachment in early life and then revisit the same person later in life). Instead researchers usually ask adolescent or adult participants questions about their relationship with parents, and identify attachment type from this. This causes two validity problems. First, asking questions relies on the honesty and accurate perception of the participants. Second, it means it is very hard to know whether what is being assessed is early attachment or in fact adult attachment. This means that the measures of early attachment used in most studies may be confounded with other factors making them meaningless.

Influence of early attachment: Confounding variables

A further limitation of studies into the influence of early attachment on later development is the existence of confounding variables. Some studies do assess attachment in infancy (e.g. McCarthy) which means that the assessment of early attachment is valid. However, even these studies may have validity problems because associations between attachment quality and later development may be affected by confounding variables. For example parenting style may influence both attachment quality and later development. Alternatively genetically-influenced personality may be an influence on both factors. This means that we can never be entirely sure that it is early attachment and not some other factor that is influencing later development.