Attachment Theory in the Classroom: Evidence and Practice

Attachment theory is often used in education to discuss trust, help-seeking and the conditions under which learners explore. That use requires care. The theory began as an account of infant-caregiver relationships, and its best-known classifications come from a specific observational procedure. They are not labels that teachers can validly assign from behaviour in lessons.

This article therefore treats attachment theory as a research lens, not a classroom diagnosis. Readers seeking a fuller history of Bowlby, his developmental phases and internal working models should use our specialist guide to Bowlby's attachment theory. For a wider comparison of psychological perspectives, see our overview of learning theories.

‌

What Does Attachment Theory Mean in the Classroom?

In education, attachment theory can shape questions about how a learner seeks help, faces doubt or returns to a task. It may help scholars form a testable idea about the quality of a bond. It does not show that a given act in class has an attachment cause.

The distinction matters because the same observation can have many explanations. A learner who does not ask for help may not understand the task, may fear peer judgement, may have a communication difference, may not yet trust the adult, or may simply prefer to work independently. Behaviour alone cannot select among these possibilities.

Terms such as secure base and safe haven also need precision. In attachment research, a secure base supports exploration, while a safe haven refers to seeking a familiar person for comfort or protection under distress. A calm and available teacher can sometimes perform similar functions. That functional similarity is not enough to conclude that the relationship is an attachment bond or that the teacher is replacing a caregiver.

‌

Bowlby, Ainsworth, Main and Solomon

Bowlby (1958) described attachment as a system of action that may aid survival. In his later account, a young child seeks a known caregiver when alarm rises. A sense of safety can then support play and exploration.

Internal working models are growing views about the self, other people and close bonds. They can change; they are not fate.

Ainsworth and Bell (1970) studied infant-caregiver behaviour in the Strange Situation, a structured sequence of separations and reunions. Work by Ainsworth and colleagues distinguished secure, avoidant and resistant or ambivalent organised patterns. These classifications described behaviour in a specified dyad and procedure. They did not establish four universal kinds of child.

Main and Solomon (1990) later introduced criteria for disorganised or disoriented behaviour during the Strange Situation. This category was not one of Ainsworth's original three. It identifies contradictions or disruptions in the organisation of attachment behaviour under the assessment conditions. The term does not mean generally chaotic conduct, and it should not be used as a synonym for trauma, abuse or attachment disorder.

These roles matter. Bowlby did not create a set of four learner types. He set out the main theory. Ainsworth and her team led the work behind the first three research groups.

Main and Solomon later added rules for coding disorganised or disoriented acts.

‌

What Attachment Classifications Mean

An attachment classification is a research judgement made within a defined relationship, age range, procedure and coding system. It is not a trait score that follows a person unchanged into every relationship. A child may have different relational experiences with different caregivers, and the meaning of observed behaviour depends on context and method.

Granqvist et al. (2017) warned against using disorganised attachment as an individual screening tool for maltreatment. The classification has multiple possible pathways and is neither a diagnosis nor proof of a particular family history. Even trained research classification is not the same task as clinical assessment. Informal classroom observation is further removed still.

The table separates four evidence layers that are frequently collapsed in popular accounts.

Swipe horizontally to compare the evidence boundaries.

‌

Stability, Development and Prediction

Attachment research deals in chance, not fate. Opie et al. (2021) found a fair level of stability in early childhood, as well as signs of publication bias. A group trend does not make one child's result permanent. It also cannot let a teacher trace a later act back to an infant bond.

Long-term studies can test whether an early measure is linked with a later social or emotional result. Such links can inform debate. Yet later bonds, current settings, growth and the choice of measure still matter. Fearon and Roisman (2017) thus treat attachment as one part of development, not a full account of later life.

Educational claims should preserve that uncertainty. “Attachment explains this learner” is too strong. A more defensible statement is that relationship history may be one of several influences worth considering in research or multidisciplinary formulation. It cannot be inferred from a single classroom episode.

‌

Culture and the Limits of Classification

Attachment theory has generated substantial cross-cultural research and debate. The cross-cultural meta-analysis by van IJzendoorn and Kroonenberg (1988) found variation both within and between samples. That pattern is incompatible with simple national stereotypes. It also shows why a classification distribution cannot be treated as a ranking of families or cultures.

Keller (2013) argues that major attachment accounts can favour certain views of freedom, face-to-face contact and care by one main adult. Those views do not fit all groups. Care may involve several known adults. Closeness, time apart, felt emotion and independence can also take different forms.

Mesman et al. (2018) make a useful point. Care can matter across settings without looking the same in each place. Staff should therefore pause before they interpret a cultural difference. They should not turn it into an attachment label or judge a family bond from an unfamiliar way of speaking or acting.

‌

Teacher-Child Relationships

Verschueren and Koomen (2012) developed an attachment perspective on teacher-child relationships while drawing an important boundary. A teacher may provide ad hoc secure-base and safe-haven functions, particularly for younger or more vulnerable children. Yet, for most children, the teacher-child relationship is probably not an attachment bond.

This wording avoids two opposite mistakes. The first is to dismiss relationships as irrelevant to learning. Availability, fairness and repair can affect whether a learner seeks help and re-engages after difficulty.

The second mistake is to describe every warm teacher bond as attachment, or to cast teachers as substitute parents. Schools are places for learning and care. They are not clinical attachment services.

A relationship lens also differs from a style lens. The useful unit of attention is often an interaction or relationship in context: when help was offered, how a rupture was repaired, whether expectations were predictable, and what changed across adults or tasks. Calling the learner “avoidant” or “disorganised” shifts attention from observable conditions to an unvalidated personal label.

‌

What the Education Evidence Shows

Reviews find links between warm teacher-learner bonds and school outcomes. Roorda et al. (2011) brought together 99 studies. The link was stronger for engagement than for attainment as a whole. Its size also varied with age and other features of the studies.

An updated meta-analysis by Roorda et al. (2017) again found links with engagement and achievement. García-Rodríguez et al. (2023) reviewed teacher-learner relationships and school engagement across later work. Together, these syntheses support taking relationship quality seriously. They do not show that a teacher can identify an infant attachment classification from classroom conduct.

Most included studies are correlational or use non-randomised designs. Relationship measures may also overlap with ratings of behaviour or engagement, and direction of influence can run both ways. A supportive relationship may enable participation, while participation and fewer conflicts may also improve the relationship. Shared classroom, family or individual factors can influence both.

The safe claim is clear. The quality of a bond matters in school and deserves care. The evidence does not support a diagnosis by style.

It does not prove that attachment caused the result or that staff should match help to a style. Work on social and emotional learning needs its own evidence too.

‌

Attachment-Aware School Programmes

“Attachment-aware” can describe staff development, whole-school language or relational approaches. The label does not identify one standard intervention. Programmes vary in content, intensity, professional input and outcome measurement. As a result, evidence for one local model cannot automatically be transferred to every programme using the term.

A study of how a programme was put in place can tell us if staff liked the training, if schools used it and what people said. This helps us judge whether the plan can work in real life. But without a fair control group, sound outcome tools and follow-up, it cannot show that the training caused gains.

School leaders should ask what was delivered, who took part and how the study measured each result. They should check for a fair control group and an independent review.

Follow-up time, harms, people left out and other likely causes also matter. A sound account of how a plan may work is not the same as proof that it did work.

A programme may include sound steps, such as clear routines and calm repair. These steps may have good grounds outside attachment theory. Leaders can judge each part by its own evidence, safe use, staff skill and cost. Our guide to trauma-informed schools covers the separate claims made for that form of provision.

‌

Professional Implications

Attachment theory does not require teachers to assign styles. Staff should describe what happened, seek the learner's view and use ordinary support routes. A theory-led guess about hidden causes is less useful than clear records, predictable help and a review of other factors that may affect access or participation.

1. Describe the event. Record the behaviour, timing, task, people present and support already tried. Avoid inserting a hidden motive or family explanation.
2. Ask without leading. Invite the learner's account of what made access, participation or help-seeking difficult. Do not ask questions designed to confirm an attachment hypothesis.
3. Make support predictable. Clarify routines, boundaries, how to request help and how the learner can return after an error or conflict.
4. Repair day-to-day strains. Name what took place, restate the rule and rebuild a sound working bond. Do not press for a private account.
5. Check other causes. Review access to the work, language and speech, special educational needs, health, peers, absence and current stress.
6. Use established routes. Share concerns with the relevant pastoral, SEND, mental-health or safeguarding lead rather than attempting an informal assessment.

These steps are for all and do not diagnose. They can help a school act as one team without claiming that attachment caused the need. They also keep access to the work and fair rules in view. A focus on bonds must not lead to lower aims or the loss of safe limits.

Where recurring conflict is the main concern, the relevant evidence may sit with classroom management and school behaviour systems. Where learners need structured support with returning to a regulated state, our account of co-regulation should be read as a separate practice literature, not proof of an attachment classification.

Read the attachment theory in education study-note transcript

ATTACHMENT THEORY IN EDUCATION

A research lens for relationships, not a classroom label

CORE CONCEPTS [1, 7]

• Attachment behavioural system: Bowlby proposed that seeking proximity to a familiar caregiver becomes especially important under alarm and supports protection.
• Secure base: caregiver availability can support exploration.
• Safe haven: a familiar caregiver may be sought for comfort or protection when distress rises. [7]
• Internal working models: expectations about self, others and relationships can develop through experience. They are not fixed destinies.

SECURE-BASE LOOP

Alarm or uncertainty → seek availability → regain enough security → explore

WHO ESTABLISHED WHAT?

BOWLBY [1]

Proposed the attachment behavioural system, secure-base account and internal working models. His developmental phases are not attachment styles.

AINSWORTH AND COLLEAGUES [2]

Developed and used the Strange Situation with infants and caregivers. They identified secure, avoidant and resistant or ambivalent organised classifications.

MAIN AND SOLOMON [3]

Later introduced criteria for disorganised or disoriented behaviour within the Strange Situation. This was not one of Ainsworth's original three classifications.

ATTACHMENT CLASSIFICATIONS

WHAT THEY MEAN [2, 3]

• Patterns coded in a specified relationship and structured assessment context.
• Research classifications requiring age-appropriate procedures and trained coding.
• Groupings that can support research questions about relationships and development.

WHAT THEY DO NOT MEAN [4, 5, 9]

• Fixed types of children or permanent personalities.
• Clinical diagnoses or a severity scale leading to attachment disorder.
• A classroom checklist based on eye contact, help seeking, anger, clinginess or attainment.
• Proof of trauma, maltreatment, neglect or a particular family history.

Evidence is probabilistic, not destiny. A meta-analysis found moderate stability across early childhood and evidence of publication bias. [5]

EDUCATION EVIDENCE [7, 8]

• A teacher may provide ad hoc secure-base or safe-haven functions, particularly for younger or more vulnerable children.
• For most children, the teacher-child relationship is probably not an attachment bond.
• Across 99 studies, affective teacher-student relationship quality was associated more strongly with engagement than with achievement overall.

ASSOCIATION IS NOT INTERVENTION PROOF

Relationship evidence does not show that teachers can identify infant attachment classifications or that attachment-aware training causes attainment gains.

OBSERVE AND DO

• Record the specific behaviour, timing, context and support already tried.
• Ask the learner what made access or participation difficult, using non-leading language.
• Make routines, help, boundaries and return to learning predictable.
• Consider curriculum access, SEND, communication, health, peers, attendance and mental health.
• Share concerns through the appropriate pastoral, SEND, mental-health or safeguarding route.

DO NOT INFER

• An attachment style or attachment disorder.
• Trauma, abuse, neglect or a caregiver cause.
• A hidden feeling, intention or “attachment need” from one behaviour.
• A style-specific intervention or reduced safety, curriculum or behaviour expectations.

CULTURE [6]

Care may involve several familiar adults. Separation, proximity, autonomy and emotional expression can carry different meanings across settings. Do not rank families or turn culture into an individual profile.

CLINICAL [9]

Secure, avoidant, resistant or ambivalent and disorganised or disoriented classifications are not Reactive Attachment Disorder or Disinhibited Social Engagement Disorder. NICE requires comprehensive assessment, age-appropriate validated tools, skilled professionals and specialist referral when needed.

SAFEGUARDING [10]

Act on observed concerns. Follow school policy and report to the designated safeguarding lead or deputy without delay. Attachment theory does not replace safeguarding procedures.

SOURCES

[1] Bowlby, 1958; 1969/1982. [2] Ainsworth & Bell, 1970; Ainsworth et al., 1978. [3] Main & Solomon, 1990. [4] Granqvist et al., 2017. [5] Opie et al., 2021. [6] van IJzendoorn & Kroonenberg, 1988; Keller, 2013; Mesman et al., 2018. [7] Verschueren & Koomen, 2012. [8] Roorda et al., 2011. [9] NICE NG26. [10] Department for Education, current KCSIE.

MEMORY LINE: Observe context. Support relationships. Do not classify the learner.

‌

Clinical and Safeguarding Boundaries

The secure, avoidant, resistant or ambivalent, and disorganised or disoriented research groups are not diagnoses. They are not Reactive Attachment Disorder or Disinhibited Social Engagement Disorder. A clinical review looks at the child's history, care, signs, daily life and other causes. NICE calls for a full review, sound tools, skilled staff and specialist referral when needed.

A teacher should not use an online checklist, training slide or classroom observation to diagnose either a research classification or a disorder. If a learner may need mental-health assessment, schools should use their established pathway. Our guide to mental health in schools covers that broader system boundary.

Safeguarding is distinct again. If a member of staff is concerned about harm, neglect or a disclosure, the response is governed by current statutory guidance and school policy. Staff should act on the concern and report it to the designated safeguarding lead or deputy without delay. They should not investigate the family, delay action to observe an attachment pattern, or treat attachment language as evidence that harm has occurred.

‌

Common Misinterpretations

Most classroom errors come from treating a research classification as a visible learner type. Attachment terms cannot reveal a learner's past, diagnosis or likely future. They also cannot show which response will work. The claims below mark the main limits that scholarship and professional practice should preserve.

‌

“Attachment style explains why this learner behaves this way”

No single classroom behaviour identifies an attachment classification or its cause. Describe the behaviour and context, ask the learner, and consider plausible educational, developmental, health and social explanations.

‌

“Disorganised attachment means trauma or abuse”

No. Granqvist et al. (2017) explicitly caution against this inference. Disorganisation has multiple possible pathways and cannot be used as an individual maltreatment screen. Safeguarding action rests on observed concerns and disclosures, not an attachment label.

‌

“A securely attached learner will always learn well”

No attachment group can promise good learning, calm conduct or good health. Learning also rests on teaching, prior knowledge, task design, time, health, language and close bonds. Our guide to child development theories sets out other views.

‌

“The teacher should become a corrective attachment figure”

This overstates both theory and role. Teachers can be reliable, respectful adults and may provide secure-base-like functions. They are not replacement caregivers or therapists, and school support must remain within professional competence and safeguarding structures.

‌

Frequently Asked Questions

Teachers often ask whether attachment groups can guide a response to one learner. The short answer is no. The theory can shape broad research questions about bonds, trust and exploration. It cannot turn acts seen in class into a style, a diagnosis or a set treatment.

‌

Can teachers identify attachment styles?

No. The familiar classifications come from structured research with specified relationships, age ranges and trained coding. Classroom observation can identify support needs and concerns, but it cannot validly assign an attachment classification.

‌

Is attachment theory useful in education?

It can be useful as a bounded conceptual lens for relationship, exploration and help-seeking. Education research also supports attention to teacher-learner relationship quality. Its usefulness falls away when the theory is used to label learners, infer family history or claim a causal classroom treatment.

‌

Are attachment classifications permanent?

No. Research finds both change and some stability. One review found a fair level of stability in early childhood. A group trend does not set the path for one child. It cannot support a guess about early bonds from an act seen years later in class.

‌

Does disorganised attachment mean attachment disorder?

No. Disorganised or disoriented behaviour is a research classification within the Strange Situation. Attachment disorders are clinical diagnoses assessed through specialist procedures and criteria. One does not form a severity step leading automatically to the other.

‌

What should a school do when relationship difficulties persist?

Document the specific pattern, seek the learner's perspective, review access and support, and involve the appropriate pastoral, SEND or mental-health professional. Any safeguarding concern must go directly through the school's designated safeguarding route.

‌

Further Reading

• NICE NG26: Children's attachment, recommendations, for clinical assessment and referral boundaries.
• Keeping Children Safe in Education, for current statutory safeguarding duties in England.
• Bowlby's Attachment Theory, for the specialist historical and theoretical account.

‌

References

• Bowlby, J. (1958). The nature of the child's tie to his mother. International Journal of Psychoanalysis, 39, 350 to 373. PubMed record.
• Bowlby, J. (1969/1982). Attachment and Loss, Volume 1: Attachment. London: Hogarth Press.
• Ainsworth, M. D. S., & Bell, S. M. (1970). Attachment, Exploration, and Separation: Illustrated by the Behavior of One-Year-Olds in a Strange Situation. Child Development, 41(1), 49 to 67. https://doi.org/10.2307/1127388.
• Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment in the Preschool Years. Publisher record.
• Granqvist, P., et al. (2017). Disorganized attachment in infancy: A review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 19(6), 534 to 558. https://doi.org/10.1080/14616734.2017.1354040.
• Opie, J. E., et al. (2021). Early childhood attachment stability and change: A meta-analysis. Attachment & Human Development, 23(6), 897 to 930. https://doi.org/10.1080/14616734.2020.1800769.
• Fearon, R. M. P., & Roisman, G. I. (2017). Attachment theory: Progress and future directions. Current Opinion in Psychology, 15, 131 to 136. https://doi.org/10.1016/j.copsyc.2017.03.002.
• IJzendoorn, M. H. van, & Kroonenberg, P. M. (1988). Cross-cultural patterns of attachment: A meta-analysis of the Strange Situation. Child Development, 59(1), 147 to 156. https://doi.org/10.1111/j.1467-8624.1988.tb03202.x.
• Keller, H. (2013). Attachment and Culture. Journal of Cross-Cultural Psychology, 44(2), 175 to 194. https://doi.org/10.1177/0022022112472253.
• Mesman, J., van IJzendoorn, M. H., & Sagi-Schwartz, A. (2018). Universality without uniformity: A culturally inclusive approach to sensitive responsiveness in infant caregiving. Child Development, 89(3), 837 to 850. https://doi.org/10.1111/cdev.12795.
• Verschueren, K., & Koomen, H. M. Y. (2012). Teacher-child relationships from an attachment perspective. Attachment & Human Development, 14(3), 205 to 211. https://doi.org/10.1080/14616734.2012.672260.
• Roorda, D. L., Koomen, H. M. Y., Spilt, J. L., & Oort, F. J. (2011). The influence of affective teacher-student relationships on school engagement and achievement. Review of Educational Research, 81(4), 493 to 529. https://doi.org/10.3102/0034654311421793.
• Roorda, D. L., Jak, S., Zee, M., Oort, F. J., & Koomen, H. M. Y. (2017). Affective teacher-student relationships, engagement and achievement: A meta-analytic update. School Psychology Review, 46(3), 239 to 261. https://doi.org/10.17105/SPR-2017-0035.V46-3.
• García-Rodríguez, L., et al. (2023). Teacher-student attachment relationship, variables associated, and measurement: A systematic review. Educational Research Review, 38, 100488. https://doi.org/10.1016/j.edurev.2022.100488.