Research figure: Mary Ainsworth · Written by Artificial Sapiens
Mary Ainsworth's Baltimore Longitudinal Study was an intensive investigation of how relationships between infants and their mothers developed during the first year of life. Beginning in 1963, Ainsworth followed 26 Baltimore families through repeated observations at home and complementary laboratory sessions. The project's defining discovery was methodological as well as psychological: patterns of infant attachment could be understood by comparing many hours of everyday interaction with a child's behavior when the caregiver's availability was briefly challenged. The study established a historically influential empirical basis for the relationship among caregiver responsiveness, infant exploration, proximity seeking, and distinct patterns of attachment behavior. Its original design and findings are documented in Ainsworth's 1974 final report and the four-author monograph Patterns of Attachment.
The Baltimore project belongs to a specific place in the history of developmental psychology. John Bowlby had formulated the broad theoretical account of attachment as a behavioral system serving protection and proximity. Ainsworth developed observational concepts and methods capable of examining that account in ordinary family life. The Baltimore evidence subsequently informed a laboratory procedure, the Strange Situation, and the original secure (B), avoidant (A), and resistant (C) classifications. The later disorganized/disoriented (D) classification was developed in the work of Mary Main and Judith Solomon. These distinct historical contributions matter when reading early attachment research today.
Historical Origins: From Uganda to Baltimore
Ainsworth's earlier study of mother–infant relationships in Uganda during 1954–1955 trained her attention on the diversity of infant signals and the circumstances in which infants could use a familiar adult as a basis for exploration. Her book Infancy in Uganda, published in 1967, described those observations. The Baltimore research program carried forward the commitment to observing ongoing relationships in their everyday ecology, while creating a more intensive record of the first year of development. The move from cross-sectional descriptions and field observation toward carefully sequenced, repeated assessments made it possible to ask whether differences seen late in infancy were related to differences documented months earlier. Historical analysis of the program places this transition within Ainsworth's scientific development after her collaboration with Bowlby and her earlier exposure to William E. Blatz's security theory. See Van Rosmalen, Van der Veer, and Van der Horst (2015).
At Johns Hopkins University, Ainsworth began the Baltimore program in 1963 with support associated with the William T. Grant Foundation. Her research question addressed both the emergence of attachment behaviors and variation among infant–mother relationships. Attachment was studied as an organized relationship across contexts and developmental time, rather than inferred from a single act of clinging, crying, or social approach. The distinctive intellectual step was to connect observation in a familiar setting to behavior in a standardized unfamiliar setting. That connection became central to subsequent assessment work.
The historical term maternal behavior reflects the sampled caregiving arrangements and original publications. In contemporary developmental research, comparable questions apply to mothers, fathers, and other consistent caregivers. This distinction preserves the historical facts while allowing the study's general scientific hypotheses to be tested with more diverse caregiving relationships.
Study Design and Participants
The original longitudinal cohort
Ainsworth's 1974 final report describes a main sample of 26 middle-class families living in Baltimore, Maryland. The infant–mother dyad was the principal observational unit. This was an unusually small cohort for estimating population frequencies, but an unusually intensively observed cohort for describing behavioral development within relationships. The design prioritized repeated, ecologically situated measurement, with laboratory assessments added later. Archival indexing describes the families as White; the project therefore provides a particularly clear example of the limits that arise when an influential measurement tradition begins with a socially and culturally narrow sample. Ainsworth (1974).
The number 26 describes the core longitudinal sample. It should not be conflated with the larger group of 106 infants who contributed to associated Strange Situation research. A later US government review notes that 23 of the Baltimore infants were included among those 106 laboratory observations. The discrepancy is a reminder that different published analyses used different subsets or expanded samples, so a numerical claim must identify the relevant method and report. Infant Attachment: What We Know Now, US Department of Health and Human Services.
The children were observed through infancy as unfolding developmental processes, not assigned to experimentally manipulated caregiving conditions. The investigation was longitudinal and observational. It could identify sequences, associations, and convergences among measures, while causal conclusions required evidence beyond its design. Subsequent research addressed that challenge through broader samples, experimental caregiving interventions, repeated measures, and quantitative synthesis.
Schedule and density of home observation
The principal observational procedure involved approximately four-hour home visits at intervals of about three weeks, spanning the third through the fifty-fourth week of life. The original report also documents variations in the earliest visit schedule: families in a later subsample received additional early observations. Across the project, the main observational record amounted to roughly 72 hours per family, although visit schedules and exposure were not identical for every participant. Describing the schedule as four hours every three weeks accurately conveys the study's organizing pattern; it should not obscure documented differences in the first weeks. Ainsworth (1974).
This density matters analytically. A single short encounter can be strongly affected by hunger, fatigue, unfamiliarity, time of day, or an unusual event. Repeated visits across months permit an investigator to compare responses to varied states and signals and to examine whether particular interaction patterns recur. The accumulation of detailed field narratives was costly and labor intensive, but it offered a stronger basis for understanding a relationship than an isolated observer impression.
The study was also longitudinal in a substantive sense. An infant's signals, mobility, exploration, and proximity-seeking repertoire change across the first year. The researchers therefore observed emerging forms of interaction as the developmental capacities that supported them appeared. Early caregiving observations and later attachment behavior could be evaluated as temporally ordered measurements, a central methodological advantage over a one-time survey.
Five Complementary Sources of Evidence
The 1974 report identifies five forms of data collection. Naturalistic home observation formed the backbone of the study. The other sources were controlled laboratory observation in the Strange Situation, nondirective interviews with mothers and other adults in the household, periodic Griffiths infant developmental assessments, and examination of medical records. These sources answered different questions and supplied contextual information relevant to interpretation. Ainsworth (1974).
Detailed narratives of ordinary interaction
During home visits, trained observers attended to ongoing activities and recorded mother–infant interaction in detail. Feeding, crying, distress, proximity, physical contact, play, exploratory activity, and changes in the caregiver's attention could become observational material. Narrative records helped preserve context: an infant's crying could be interpreted in light of what preceded it, what the caregiver noticed, what the caregiver did, and how the infant responded. The investigators were looking for recurring organization rather than collecting decontextualized counts of affection or physical contact. Later historical and methodological scholarship describes how these records supported the development of Ainsworth's scales of caregiving behavior. Bretherton (2013).
Long home observations can reveal characteristics that a rigid laboratory task misses: the pace of everyday caregiving, interruptions, sibling or household presence, transitions between states, and opportunities for exploration when nothing dramatic happens. Observational detail nevertheless introduces analytic challenges. Coding judgments can depend on training, shared assumptions, context retained in the record, and the observer's decisions about what constitutes a response. The Baltimore study helped build methods for making such judgments explicit, but modern research still requires transparent protocols and checks on interrater reliability.
Interviews, developmental tests, and medical information
Nondirective interviews supplied family context and accounts from the adults who participated in daily care. The Griffiths Infant Intelligence Scale was administered at intervals of approximately nine weeks according to the final report, while medical records offered additional developmental and health information. These measures could help distinguish patterns of attachment-related behavior from broader variations in infant functioning and circumstances. They did not transform the project into a randomized experiment or into a comprehensive study of every possible determinant of child development.
The inclusion of these complementary sources remains scientifically significant because attachment behavior can be misread when it is detached from developmental context. An infant's capacity to crawl toward a caregiver, the circumstances surrounding separation, or the child's current health can affect visible behavior. The project treated such observations as relevant interpretive conditions rather than equating every difference in behavior with a difference in relationship quality.
From Home Observations to the Strange Situation
Why introduce a laboratory procedure?
Ainsworth sought a setting that would make the activation and organization of attachment behavior more observable across dyads. In a familiar home with the caregiver accessible, a securely supported child might spend much of the time exploring. A carefully unfamiliar environment and brief separation episodes could elicit behaviors that ordinary observation did not make equally salient in every family. The Strange Situation therefore complemented the home record by creating a partly standardized context for examining how exploration, distress, reunion, and contact seeking varied. The approach emerged during the Baltimore research and was presented in the 1969 work of Ainsworth and Barbara A. Wittig. Historical reconstructions trace its conceptual and methodological precedents while retaining their coauthorship. Van Rosmalen et al. (2015).
The laboratory assessment was introduced around the child's first birthday; the archival report notes Strange Situation observations beginning at approximately 51 weeks. The procedure used a sequence of episodes involving a novel environment, an unfamiliar adult, caregiver departures, and reunions. Its interpretation emphasized the pattern of behavior across episodes, especially the child's organization of contact, comfort seeking, and return to exploration upon reunion. This design linked a relatively brief laboratory encounter to many hours of prior observations made in a very different setting.
The procedure became a major methodological innovation because the infant's reaction could be interpreted relationally. Crying at separation could indicate protest, distress, temperament, or the stressful character of the setting; its presence alone did not define a secure or insecure relationship. Likewise, an apparently calm reaction did not automatically mean the child felt secure. The combination of reunion behavior, proximity seeking, avoidance, resistance, and exploration allowed more precise distinctions. For the laboratory method in its own right, see the original empirical literature and Patterns of Attachment.
The home–laboratory validation strategy
The study's crucial interpretive move was to compare later Strange Situation behavior with patterns already observed in the home. This was a form of convergent, ecologically informed validation: if a brief procedure identified meaningful relationship differences, those differences should be intelligible in the independently accumulated developmental record. The availability of intensive longitudinal observations gave the original classifications a richer empirical foundation than a laboratory snapshot could provide by itself.
This strategy also created methodological questions that continue to matter. If the same research team developed the concepts, scales, and laboratory classifications, theoretical expectations could influence coding across methods. The small homogeneous sample limited the precision of the observed associations. Subsequent independent research became necessary to evaluate interrater agreement, predictive utility, cultural adaptation, and whether comparable relationships emerged with different researchers, families, and measurement instruments.
Principal Scientific Discoveries
Attachment behavior developed as an organized system
The Baltimore narratives documented changes in how infants oriented to and sought proximity with familiar caregivers as they gained social, motor, and cognitive capabilities. The research treated attachment as the coordination of behavior in relation to a particular adult and a changing environment. An infant might explore when proximity felt accessible, seek contact during alarm, recover, and then resume exploration. Such shifts are the basis for the empirical language of the caregiver as a secure base and a source of comfort. Ainsworth's observational refinement made Bowlby's broader theoretical account available to systematic developmental research. Ainsworth et al. (1978).
Because infants vary in general arousal, sociability, and activity, the organization of their response mattered more than the frequency of any one behavior. The observations supported a relational interpretation of early security: a pattern of coordinated expectations and interactions in a specific dyad, expressed in age-appropriate forms. This interpretive principle remains central to distinguishing attachment classification from global judgments about temperament or personality.
Sensitive and responsive care was associated with attachment security
Ainsworth's team observed meaningful differences in how caregivers detected and responded to infant communication. Sensitive care, in her operational approach, involved awareness of signals, reasonably accurate interpretation, promptness, and a response suited to the infant's needs and developmental state. Responsiveness included respecting the child's exploratory initiative and providing comfort when it was needed. Bretherton's review of the original scales explains how the Baltimore record supported this construct and how it was situated among related caregiving dimensions. Bretherton (2013).
The central research claim was that infant attachment behavior in the laboratory was related to the caregiving patterns documented at home. The magnitude of an observed association in this very small intensive cohort cannot be assumed to hold in every population. Modern meta-analysis supplies a wider estimate: Madigan and colleagues synthesized 174 samples and 230 effect sizes representing 22,914 parent–child dyads and found a pooled positive correlation of r = .25 between caregiving sensitivity and attachment security (95% confidence interval .22–.28). This finding supports a modest, reproducible relationship; it does not show that sensitivity accounts for every difference in attachment or that caregiving outcomes are mechanically determined. Madigan et al. (2024).
The same synthesis found maternal sensitivity and attachment security correlated at r = .26, while paternal sensitivity and security correlated at r = .21, with overlapping interpretive implications despite differences in the evidence base. Longer sensitivity observations and measures with stronger interrater reliability tended to produce larger observed associations. This moderator result directly illuminates the Baltimore project's unusual investment in prolonged naturalistic observation: measurement duration affects what investigators are able to detect. Madigan et al. (2024).
Caregiving quality was multidimensional
The historically influential Maternal Care scales developed from the Baltimore work encompass sensitivity–insensitivity, cooperation–interference, acceptance–rejection, and accessibility–ignoring/neglecting. They described related but distinguishable dimensions of the infant's everyday interpersonal environment. Cooperation concerns the adult's capacity to coordinate with the child's ongoing activity rather than repeatedly imposing a competing agenda. Accessibility concerns practical and psychological availability. Acceptance concerns the overall affective stance toward infant needs and care. These constructs were framed in bipolar rating scales and interpreted through sustained observations rather than a momentary checklist. Bretherton (2013).
The scales' conceptual structure discourages simplistic equations between emotional expressiveness and sensitivity. A caregiver can be affectionate yet sometimes interfere with exploration; another may express affection quietly while responding appropriately to the infant's cues. Cultural customs, working conditions, competing demands, and the child's characteristics can change the form a response takes. Contemporary research must therefore clarify its constructs and the context in which behavior is assessed.
Distinct A, B, and C patterns became observable
The original research program distinguished avoidance (A), secure attachment (B), and resistance or ambivalence (C) as organized infant–caregiver patterns. In the Strange Situation, the child's response to separation and especially reunion helped describe how the relationship operated under stress. The longitudinal home observations supplied essential interpretive context for those laboratory distinctions. Detailed classification procedures, subgroup organization, and the larger subsequent research program were presented in Patterns of Attachment by Mary D. Salter Ainsworth, Mary C. Blehar, Everett Waters, and Sally Wall, first published in 1978. All four authors should be credited for that monograph. Ainsworth et al. (1978).
The later disorganized/disoriented (D) classification emerged from research by Mary Main and Judith Solomon and was introduced after Ainsworth's original three-category program. It addresses behavior that does not organize into a consistent A/B/C strategy under particular circumstances. The D category should not be retroactively described as one of Ainsworth's original classifications. Nor should the original infant categories be casually mapped onto modern adult romantic attachment typologies, which have different procedures, age groups, instruments, and theoretical developments.
The Study's Intellectual and Collaborative Network
The Baltimore study emerged from Ainsworth's historical dialogue with John Bowlby's attachment theory and her earlier intellectual formation through William Blatz. Her empirical program depended on skilled collaborators and successive research reports. Barbara A. Wittig is a coauthor of the 1969 Strange Situation chapter; Silvia M. Bell coauthored the 1970 study of attachment, exploration, and separation and contributed to related research on infant–mother interactions; Donelda J. Stayton coauthored important subsequent work on infant behavior and reciprocal responsiveness. Wellcome Collection's historical record of the Ainsworth, Bell, and Stayton work preserves part of this intellectual network.
The subsequent book Patterns of Attachment integrated evidence from the research program and further methodological analyses under the joint authorship of Ainsworth, Mary C. Blehar, Everett Waters, and Sally Wall. Its first publication date is 1978, confirmed by the original-edition bibliographic record despite discrepancies in descriptions of some later reprints. The collective authorship is part of the historical evidence of how an intensive longitudinal study became a durable measurement and classification tradition. Publisher record.
Authorship, conceptual priority, and research responsibility are distinct questions. Bowlby established the theoretical foundation. Ainsworth advanced observational constructs and the Baltimore research design. Wittig, Bell, Stayton, Blehar, Waters, and Wall made credited contributions to specific publications and phases of the program. Main and Solomon developed a later classificatory extension. This attribution structure prevents the history of attachment research from collapsing into a single inventor narrative.
Methodological Strengths: Why the Study Became a Landmark
The most consequential strength was sustained observation in a familiar environment. The researchers observed caregiver behavior while ordinary life unfolded, giving them repeated opportunities to identify regularities and exceptions. The developmental sequence meant that observations collected early in infancy could be compared with later behavior in a separate laboratory context. Multiple sources of information offered opportunities for contextualization, even though naturalistic observations carried the greatest analytic weight.
A second strength was the reciprocal relation between theory and method. Attachment theory generated questions about proximity, comfort, and exploration; detailed observation challenged investigators to operationalize those questions; the resulting measures allowed hypotheses to be compared across families and later samples. That iterative scientific design helped shift the field from general speculation about early relationships toward a research program with publicly describable procedures and concepts. Historical syntheses place the Baltimore project at the origin of a methodology that later underwent extensive independent testing. Van Rosmalen et al. (2015).
A third strength was methodological complementarity. Laboratory standardization and ecological observation answer different questions. Standardization improves comparability under controlled stress; repeated home observation shows the history and variation of the relationship across settings. The Baltimore project made the contrast scientifically productive rather than relying on either form of observation alone. Contemporary attachment research continues to confront the same question of how short standardized tasks relate to complex everyday interaction.
Limits of Inference and Scientific Debate
Sampling, representativeness, and precision
The original cohort was small, drawn from a narrow middle-class urban context, and centered on mothers. A sample of 26 families is sufficient to develop detailed descriptive hypotheses but provides limited statistical precision and a weak basis for claims about population prevalence. The study cannot by itself establish that proportions of attachment patterns apply to all cultures, socioeconomic conditions, or caregiving arrangements. Later broad syntheses therefore have a different evidentiary role from the Baltimore data: they estimate patterns across many independent samples and make sources of heterogeneity visible.
The 2023 meta-analysis by Madigan and colleagues included over 285 Strange Situation studies and 20,720 dyads, reporting pooled proportions for secure, avoidant, resistant, and disorganized classifications of 51.6%, 14.7%, 10.2%, and 23.5%, respectively. These are distributions across later research using the expanded four-category framework, not estimates from Ainsworth's original Baltimore cohort. The synthesis identified variations associated with geographical region and sociodemographic risk, reinforcing the need to specify the population behind any summary. Madigan et al. (2023).
Observational association and causal explanation
The home observation–laboratory relationship provided evidence for a developmental account connecting caregiver responsiveness with security. It did not randomly assign caregiving quality, isolate every competing influence, or establish a universal causal mechanism. Infant temperament, caregiver stress, household resources, social support, health, the context of separation, and unmeasured selection effects can contribute to outcomes. Later studies and interventions can test more specific causal claims, but historical associations should be described in their original evidentiary form.
A meta-analytic association of r = .25 leaves substantial variation unexplained. That is scientifically compatible with caregiving being meaningful and with attachment security reflecting multiple interacting influences. The result should never be used to assign blame to an individual caregiver on the basis of one infant's observed reaction. The research concerns distributions of relational behavior under specified conditions, and a responsible interpretation recognizes both the contribution of care and the complexity of development. Madigan et al. (2024).
Cultural context and the meaning of separation
The Strange Situation activates behavior through novelty and temporary caregiver unavailability. The meaning and ordinary frequency of separation vary across communities, family routines, and caregiving systems. A specific behavioral response may therefore reflect both relationship history and local expectations about proximity. Cross-cultural interpretation requires careful research into the ecology of care and the measurement procedure, rather than assuming that one Baltimore classification distribution defines the standard for every population.
The classic meta-analysis by van IJzendoorn and Kroonenberg examined approximately 2,000 Strange Situation classifications across eight countries and found both within-culture and between-culture variation. A far larger 2023 synthesis demonstrated the continuing relevance of geographical and population differences. These studies support broad comparative research while cautioning against treating a laboratory distribution as culturally uniform. Van IJzendoorn and Kroonenberg (1988); Madigan et al. (2023).
Observation, coding, and the question of replication
Ainsworth's intensive notes yielded unusually rich material, but the same richness made the approach difficult to replicate at scale. Modern studies frequently assess caregiving over shorter intervals and use different observational scales. The 2024 sensitivity meta-analysis identified both observation length and measurement reliability among moderators of reported associations. A weaker correlation in a later study may reflect substantive population differences, measurement choices, or both. This finding clarifies why replicating a result requires attention to the measurement process, not simply repeating the verbal hypothesis. Madigan et al. (2024).
The laboratory method has been extensively repeated, but an exact duplication of Baltimore's highly intensive home-visit design is a different and resource-intensive project. Consequently, evidence for the Strange Situation's subsequent use and evidence for every original Baltimore interpretation must be evaluated separately. Modern cumulative findings are strongest for the existence of patterned infant–caregiver behavior and a modest positive relationship with observed sensitivity. Precise causal narratives, universal behavioral meanings, and predictions about an individual's adult psychological fate have weaker support.
Ethical and Clinical Interpretation
Contemporary researchers examining historical observational work should distinguish recorded consent procedures, the ethical expectations of its period, and current standards for privacy, safeguarding, and informed participation. The original report describes preliminary contact with families and collection of observational, interview, testing, and medical information. This design would today require explicit institutional review, secure handling of sensitive family information, safeguards surrounding recording and secondary data use, and attention to the burdens of repeated home observation. It is historically appropriate to assess the research in its original institutional setting while applying current ethics to new studies inspired by it.
An attachment classification is a description of behavior in a relationship under specified observational conditions. It is not a psychiatric diagnosis, screening verdict, measure of intelligence, or fixed personality type. A brief episode of crying, apparent independence, resistance, or avoidance cannot establish a disorder or determine whether a caregiver is a good parent. Clinical assessment requires a different evidentiary process and relevant professional standards. It is equally important to distinguish disorganized attachment classifications from diagnoses of attachment disorders, which have separate clinical definitions and criteria.
For families and practitioners, the defensible implication is a general one: observing and responding thoughtfully to infant signals matters, and the broader circumstances of caregiving affect what responses are possible. The research offers a developmental framework for understanding relationship-specific behavior; it is not a license to classify children informally or to forecast their future relationships from a single stressful encounter.
The Baltimore Study in Contemporary Developmental Science
Today the Baltimore project functions as a methodological reference point. Its questions link observational developmental psychology, relational assessment, caregiver–infant interaction, attachment measurement, and cross-cultural research. Its most durable contribution is a research architecture: observe a relationship repeatedly, identify organized behavior, operationalize caregiver responsiveness, compare observations across contexts, and test the proposed associations in new samples.
Contemporary evidence supports several components while sharpening their limits. The 2024 meta-analysis confirms a consistent modest sensitivity–security association across diverse dyads. The 2023 Strange Situation review documents a large, cumulative, geographically varied literature. Systematic reviews extending into the preschool years examine whether caregiver sensitivity and attachment are associated across later developmental periods; this is an extension of the original question, not a direct reanalysis of the Baltimore families. O'Neill and colleagues (2021).
Ainsworth's observations also encourage researchers to attend to context rather than reduce attachment to a popular typology. Infant responses are embedded in emerging capacities, caregiver behavior, broader household life, and social institutions. This is why modern studies can preserve her behavioral insights while developing new designs, including observations of different caregivers, long-term follow-ups, interventions, and methods suited to diverse communities. Ainsworth's historical project provided a fertile beginning to that cumulative enterprise.
Frequently Asked Questions
How many families participated in the original Baltimore study?
The main longitudinal cohort comprised 26 Baltimore middle-class families. Associated laboratory research also included a larger group of infants; the US Department of Health and Human Services reports that 23 of the Baltimore infants were among 106 infants in an associated Strange Situation analysis. Those are different denominators and should not be combined. Ainsworth (1974); ASPE historical review.
What did the researchers actually observe?
Repeated home visits documented ordinary caregiving and infant behavior as they unfolded, emphasizing signals, responses, proximity, play, distress, and exploration. The investigation also incorporated interviews, developmental testing, medical records, and observations in the Strange Situation near the first birthday. Ainsworth (1974).
What was the study's central discovery?
Its central empirical contribution was the relationship between the quality of repeatedly observed caregiving and organized infant attachment behavior, including how children balanced exploration and proximity in familiar and unfamiliar settings. Its central methodological contribution was linking prolonged home observation with a standardized laboratory assessment. Later meta-analyses support a modest association between sensitive caregiving and secure attachment in much larger samples. Madigan et al. (2024).
Was the Baltimore project an experiment?
It was principally a longitudinal observational study. Its laboratory procedure introduced standardized conditions for observing attachment behavior but did not randomly assign infants to different lifelong caregiving histories. Its correlational findings therefore require a carefully bounded interpretation of causality.
Did Mary Ainsworth identify disorganized attachment?
Ainsworth's original program established A, B, and C classifications. The later disorganized/disoriented D category was formulated by Mary Main and Judith Solomon. The contemporary four-category distribution cannot be presented as an original result of the Baltimore cohort.
Can the Strange Situation diagnose a child or predict adult relationships?
No. It is a structured observational method for infant–caregiver attachment relationships under specific conditions. It is not a clinical diagnosis or an individual forecast of adult romantic attachment. Predictions of later outcomes require separate studies, different measurements, and attention to developmental context.
Related Articles
Mary Ainsworth — The researcher's biography, broader intellectual contributions, major works, and place in developmental psychology.
Attachment in Infancy and Early Childhood: Security and Development — The general developmental account of early attachment relationships and secure-base behavior.
What Causes Attachment Styles? Childhood Caregiving, Temperament, and Later Relationships — A thematic analysis of different contributors to attachment patterns beyond the original historical cohort.
Is Attachment Theory Evidence-Based? What Research Supports, What It Does Not, and Common Misuses — A cross-study evaluation of the broader theory, its research evidence, and interpretation limits.
Mary Ainsworth's Uganda Study: Fieldwork and Findings
Mary Ainsworth's Caregiver Sensitivity Scales: Observation and Evidence
Mary Ainsworth's Patterns of Attachment: The 1978 Landmark Book
References
Ainsworth, M. D. S. (1974). The development of infant–mother attachment: A final report of the Office of Child Development. Office of Child Development. ERIC ED122924. https://eric.ed.gov/?id=ED122924.
Ainsworth, M. D. S., and Bell, S. M. (1970). Attachment, exploration, and separation: Illustrated by the behavior of one-year-olds in a strange situation. Child Development, 41(1), 49–67. https://doi.org/10.2307/1127388.
Ainsworth, M. D. S., Bell, S. M., and Stayton, D. J. (1974). Infant–mother attachment and social development: Socialization as a product of reciprocal responsiveness to signals. In M. P. M. Richards (Ed.), The integration of a child into a social world. Cambridge University Press. Wellcome archival record.
Ainsworth, M. D. S., Blehar, M. C., Waters, E., and Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Lawrence Erlbaum Associates. https://doi.org/10.4324/9781315802428.
Bretherton, I. (2013). Revisiting Mary Ainsworth's conceptualization and assessments of maternal sensitivity–insensitivity. Attachment and Human Development, 15(5–6), 460–484. https://doi.org/10.1080/14616734.2013.835128.
Duschinsky, R. (2020). Mary Ainsworth and the Strange Situation Procedure. In Cornerstones of Attachment Research. Oxford University Press. Oxford Academic.
Madigan, S., Deneault, A.-A., Duschinsky, R., Bakermans-Kranenburg, M. J., Schuengel, C., van IJzendoorn, M. H., Ly, A., Fearon, R. M. P., Eirich, R., and Verhage, M. L. (2024). Maternal and paternal sensitivity: Key determinants of child attachment security examined through meta-analysis. Psychological Bulletin, 150(7), 839–872. https://doi.org/10.1037/bul0000433.
Madigan, S., Fearon, R. M. P., van IJzendoorn, M. H., Duschinsky, R., Schuengel, C., Bakermans-Kranenburg, M. J., Ly, A., Cooke, J. E., Deneault, A.-A., Oosterman, M., and Verhage, M. L. (2023). The first 20,000 Strange Situation procedures: A meta-analytic review. Psychological Bulletin, 149(1–2), 99–132. https://doi.org/10.1037/bul0000388.
O'Neill, M. C., Badovinac, S., Pillai Riddell, R., Bureau, J.-F., Rumeo, C., and Costa, S. (2021). The longitudinal and concurrent relationship between caregiver sensitivity and preschool attachment: A systematic review and meta-analysis. PLOS ONE, 16(1), e0245061. https://doi.org/10.1371/journal.pone.0245061.
US Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. (1991). Infant attachment: What we know now. https://aspe.hhs.gov/reports/infant-attachment-what-we-know-now-0.
van IJzendoorn, M. H., and Kroonenberg, P. M. (1988). Cross-cultural patterns of attachment: A meta-analysis of the Strange Situation. Child Development, 59(1), 147–156. https://doi.org/10.2307/1130396.
Van Rosmalen, L., Van der Veer, R., and Van der Horst, F. C. P. (2015). Ainsworth's Strange Situation Procedure: The origin of an instrument. Journal of the History of the Behavioral Sciences, 51(3), 261–284. https://doi.org/10.1002/jhbs.21729.