Human socioemotional development is rooted in the early relational bonds forged between an infant and their primary caregivers. The Attachment Q-Set (AQS) stands as one of the most methodologically rigorous, ecologically valid, and influential observational instruments in modern developmental psychology, designed specifically to capture the dynamic organization of young children's attachment behaviors within their natural habitats. By operationalizing John Bowlby's concept of the caregiver as a secure base from which to explore and a safe haven to which to return in times of distress, the instrument bridges the gap between intricate ethological theory and precise psychometric quantification.
Attachment Q-Set (AQS)
1. Concise Definition
The Attachment Q-Set (AQS) is a standardized observational assessment tool and quantitative psychometric methodology designed to evaluate the quality and organization of attachment behavior in infants and young children aged roughly 12 to 60 months within naturalistic, familiar environments. Developed primarily by Everett Waters and Kathleen E. Deane, the instrument translates open-ended behavioral observations of a child interacting with their primary caregiver into a structured, continuous profile of attachment security. It achieves this by employing an ipsative card-sorting procedure that systematically arrays behavioral descriptors against an empirically derived criterion of the hypothetically ideal secure child.
Unlike laboratory-bound, stress-inducing stressor paradigms that rely on artificial separations and reunions, the AQS captures how a child balances the competing motivational systems of environmental exploration and proximity-seeking over extended observational periods. The resulting security score is represented as an objective correlation coefficient ranging from -1.00 to +1.00, reflecting the degree to which a specific child's behavioral repertoire mirrors the prototypical organization of attachment security established by international developmental experts.
2. Etymology & Linguistic Origin
The nomenclature of the Attachment Q-Set derives from two distinct intellectual lineages: developmental psychoanalysis and psychometric methodology. The term attachment traces its linguistic roots to the Old French verb estachier (cognate with the Old English staca, meaning a stake or post), which evolved into the Middle French attachement, denoting a physical fastening, ligature, or tether. In the mid-twentieth century, British psychoanalyst and ethologist John Bowlby appropriated the word to define the primary, biologically evolution-driven affectional bond that connects an infant to a selective protective figure.
The designation Q-Set (frequently interchanged with Q-Sort) originates directly from the psychometric innovation known as Q-methodology, formulated in 1935 by British physicist and psychologist William Stephenson. Stephenson introduced the letter "Q" to provide a formal mathematical contrast to conventional "R-methodology" (named after Charles Spearman's and Karl Pearson's correlation coefficient r). While R-methodology calculates correlations between isolated trait variables across large populations of individual subjects, Q-methodology inverts the analytic matrix to calculate correlations between individual profiles across a comprehensive array of behavioral items. Everett Waters and Kathleen E. Deane fused Stephenson's structural mechanics with Bowlby's theoretical architecture in their seminal 1985 monograph published by the Society for Research in Child Development (SRCD), formally introducing the tool as the Attachment Behavior Q-Set.
3. Pronunciation & Grammatical Form
The instrument is phonetically pronounced in academic English as /əˈtætʃ.mənt kjuː sɛt/ or colloquially abbreviated as the initialism A-Q-S (/ˌeɪ kjuː ˈɛs/). Within formal scientific prose, the term functions primarily as a proper compound noun (e.g., "The Attachment Q-Set was administered at 24 months"). It is frequently employed in an attributive nominal sense to modify related research constructs, appearing in standard formulations such as "AQS security score," "AQS methodology," "AQS observer sort," or "AQS criterion sort."
In clinical and developmental literature, writers often alternate between "Attachment Q-Set" (referring strictly to the physical deck of standardized behavioral descriptors) and "Attachment Q-Sort" (referring to the procedural act of sorting or the resulting data matrix). Both variants are grammatically and conceptually accepted, although "Q-Set" remains the designated formal title favored by its original authors in published operational manuals.
4. Detailed Conceptual Explanation
To comprehend the Attachment Q-Set, one must understand how attachment theory conceptualizes early socioemotional adaptation. Attachment is not viewed merely as the raw quantity of physical affection or maternal proximity, nor is it equated with general dependency. Rather, it is understood as an organized behavioral control system designed by evolutionary natural selection to maintain proximity to a protective caregiver under conditions of threat, illness, or environmental novelty, while facilitating autonomous mastery of the physical and social world when safety is assured.
The AQS operationalizes this dynamic equilibrium through the core construct of the secure base phenomenon. In a naturalistic setting such as the home, a securely attached child treats the caregiver as an anchor of psychological and physical safety. The child ventures forth to manipulate toys, interact with siblings, or inspect unfamiliar domestic objects, but continuously engages in distal communication—exchanging smiles, vocalizing, making visual checks, and showing objects across distances. When startled, fatigued, or frustrated, the child readily truncates exploration, returns promptly to the caregiver for physical comfort or emotional refueling, absorbs that soothing, and rapidly re-engages the environment.
Conversely, insecurely organized attachment manifests in observable disruptions of this regulatory cycle. Some children display avoidant organization, suppressing proximity-seeking behavior even when stressed, maintaining an exaggerated focus on inanimate objects, and treating the caregiver with emotional detachment. Other children manifest resistant or ambivalent patterns, exhibiting pervasive, anxious clinginess, extreme distress upon minor domestic separations, and an inability to use the caregiver's presence to calm down and resume exploration. The AQS measures these dynamic patterns through 90 discrete, finely articulated behavioral items printed on individual cards, capturing nuanced nuances of everyday interactions.
The fundamental conceptual innovation of the AQS is its capacity to assess behavioral organization rather than simple behavioral frequencies. In traditional time-sampling observation, a researcher tallies how many times an infant cries or touches a mother over two hours. However, such absolute counts are easily confounded by transient infant temperament, physical fatigue, hunger, or domestic disruptions. The AQS overcomes this psychometric obstacle by forcing the observer to evaluate the contextual salience and relative prominence of behaviors within the child's overall behavioral repertoire, providing an ecologically valid portrait of how the child organizes self-regulation in relation to a specific caregiving figure.
5. Historical Development
The genesis of the Attachment Q-Set was propelled by methodological dissatisfaction with existing observational approaches during the late 1970s and early 1980s. Historically, Mary Ainsworth and her colleagues had originally conceptualized attachment through exhaustive, naturalistic home observations conducted across the first year of life in Uganda and Baltimore, spending upward of 72 hours per family. However, because such longitudinal field observations were extraordinarily labor-intensive and difficult to replicate across large epidemiological samples, Ainsworth devised the laboratory-based Strange Situation Procedure (SSP) in 1978.
While the SSP became the global gold standard for classifying infant attachment into secure (Group B), avoidant (Group A), and resistant/ambivalent (Group C) patterns, it possessed distinct methodological constraints. It was strictly validated only for infants between 11 and 18 months of age, was highly sensitive to artificial laboratory stress, could not be repeatedly administered across short intervals due to carryover effects, and was criticized by ecological psychologists for divorcing relational assessment from the child's natural habitat.
Recognizing these limitations, Everett Waters—a student and collaborator of Ainsworth—joined Kathleen E. Deane to engineer an alternative methodology that could retain the ecological richness of Ainsworth's original home observations while providing the psychometric standardization and quantitative precision of modern rating scales. In 1985, they published the initial iteration of the Attachment Q-Set in the landmark SRCD Monograph Growing Points of Attachment Theory and Research. This preliminary version contained 100 items but was subsequently refined, pruned of redundant or psychometrically unstable descriptors, and consolidated into the canonical 90-item Version 3.0 in 1995.
Over the subsequent decades, the AQS transitioned from a specialized research instrument into an international standard for assessing toddler and preschool attachment. The publication of a massive meta-analysis by Marinus van IJzendoorn and colleagues in 2004 established the psychometric validity of the AQS across hundreds of independent studies, solidifying its standing as an indispensable counterpart to the Strange Situation Procedure in developmental science.
6. Theoretical Foundations
The Attachment Q-Set rests upon three interlocking conceptual foundations: ethological attachment theory, cybernetic systems control models, and Stephenson's Q-methodological measurement theory.
Ethologically, the AQS adopts John Bowlby's tenet that infant attachment behaviors are biologically canalized survival mechanisms. Rather than viewing attachment as a secondary drive derived from feeding or oral gratification (as posited by traditional psychoanalysis and early drive theories of behaviorism), the AQS assumes that human infants possess an innate behavioral control system designed to regulate proximity to an attachment figure. This ethological framing mandates that behavior must be studied in the natural environment—the environment of evolutionary adaptedness—where the functional utility of the secure base phenomenon can be observed in response to ordinary, everyday contextual variations.
From modern cybernetics and cognitive science, the AQS integrates the concept of the internal working model (IWM). Bowlby postulated that through repeated daily interactions with primary caregivers, children construct cognitive-affective representations of self and others. A child whose caregiver is consistently accessible and responsive constructs an internal model of the self as worthy of care and the caregiver as dependable. In the AQS paradigm, this internal working model is not directly accessed via verbal self-report; instead, it is directly inferred from the structural coherence and flexibility of the child's behavioral choices across hours of unstructured domestic activity.
Finally, the psychometric foundation of the AQS draws heavily from Stephenson's ipsative measurement theory. Traditional normative rating scales suffer from endemic rater biases, including the halo effect (the tendency to rate an attractive or pleasant child highly across all positive traits), leniency bias, and central tendency bias. By requiring the observer to sort a standardized deck of items into a forced, quasi-normal distribution, the AQS neutralizes these systematic errors. The score assigned to any single behavioral descriptor is entirely dependent upon its relative salience compared to all other descriptors within that individual child's behavioral ecology.
7. Key Components, Types & Dimensions
The standard Attachment Q-Set (Version 3.0) consists of 90 meticulously defined behavioral descriptors printed on individual cards. These items span multiple behavioral domains and are systematically categorized across structural dimensions:
- Secure Base and Exploration Items: Descriptors that assess how the child uses the caregiver as a launching pad for autonomous play and returns for comfort (e.g., Item 13: "When the child is upset by mother's leaving, he continues to cry or remains distressed after she returns"; Item 87: "Child easily abandons play activities to follow mother when she leaves the room").
- Affective Sharing and Communication Items: Items detailing the richness, frequency, and spontaneity of positive emotional exchange across physical space (e.g., Item 1: "Child readily shares with mother or lets her hold things if she asks"; Item 9: "Child is lighthearted and playful most of the time").
- Proximity and Contact Seeking Under Stress: Behavioral markers reflecting how the child seeks relief during moments of sudden distress, injury, or unexpected environmental intrusion (e.g., Item 44: "Child asks for and accepts comfort from mother when hurt or frightened").
- Social Orientation Toward Unfamiliar Adults: Items capturing how the child negotiates interactions with visiting observers, strangers, and guests, balancing interest with appropriate wariness (e.g., Item 16: "Child is more interested in people than in things"; Item 33: "Child copies visitors' behaviors or takes suggestions from them").
- Dependency and Executive Agency Items: Descriptors that purposefully distinguish pathological dependency, helplessness, or passivity from genuine attachment security (e.g., Item 18: "Child follows mother around the house, clutching her or trying to be held"; Item 45: "Child enjoys dancing or performing for mother or other adults").
- The Nine-Pile Forced Distribution: The procedural matrix requiring the sorter to allocate exactly 10 cards to each of 9 conceptual piles, ranging from Pile 1 ("Most Uncharacteristic") through Pile 5 ("Neither Characteristic Nor Uncharacteristic") to Pile 9 ("Most Characteristic").
- The Criterion Sorts: Empirically derived arrays established by expert consensus. The primary array is the Security Criterion, generated by leading international attachment scholars who sorted the 90 items to describe the "hypothetically ideal secure child." Secondary criteria, such as the Dependency Criterion, allow researchers to parse apart security from immature clinginess.
8. Examples & Illustrative Cases
To illustrate the operational dynamics of the Attachment Q-Set, consider two contrasting clinical research vignettes evaluated during a standard three-hour domestic observation:
Case Illustration 1: High Attachment Security (Liam, 26 months).
During the home observation, Liam is playing with construction blocks in the living room while his mother prepares food in the adjacent open-plan kitchen. Liam occasionally looks up, catches his mother's eye, smiles, and vocalizes enthusiastically ("Look, tall!"), to which the mother replies with warmth (Item 1: high). When the visiting researcher moves closer to offer a new toy car, Liam pauses, glances at his mother's face for social referencing, receives a reassuring nod, and confidently takes the car (Item 28: high). Halfway through the session, Liam trips over a rug and grazes his knee; he immediately bursts into tears, scrambles directly to his mother, buries his face in her lap, and is gathered into her arms. Within two minutes of being held and soothed, his crying ceases, his physiological respiration regularizes, and he leaves her lap to test the toy car against the baseboard (Item 44: high; Item 13: low). When sorted into the 90-item distribution and correlated with the security criterion, Liam receives an AQS security coefficient of +0.74, reflecting an exceptionally well-organized secure base system.
Case Illustration 2: Insecure-Avoidant Organization (Julian, 28 months).
Julian is observed in a domestic setting with his mother. Throughout the observation, Julian remains deeply engrossed with a battery-operated train set, sitting with his back firmly turned toward his mother. He does not display toys to her, rarely initiates eye contact, and does not monitor her movements as she enters and exits the living area (Item 23: low). When the mother initiates an interaction by crouching near him, Julian avoids looking at her face, redirects his attention to the train tracks, or sidesteps out of reach without exhibiting overt hostility (Item 65: high). When the unfamiliar observer drops a clipboard with a loud crash, startling Julian, he freezes momentarily; however, rather than seeking his mother's proximity or comfort, he turns to an inanimate sofa cushion and rhythmically pats it until his autonomic arousal subsides. When Julian's behavioral matrix is correlated against the expert security criterion sort, his resulting AQS score is -0.31, capturing a prominent avoidant adaptation characterized by the defensive exclusion of the attachment figure during baseline and mildly stressful experiences.
9. Measurement & Assessment
Administering the Attachment Q-Set requires adherence to rigorous methodological protocols to ensure high psychometric reliability. The standard administration unfolds across three distinct methodological phases: naturalistic field observation, the physical sorting procedure, and mathematical correlation analysis.
During the observational phase, an extensively trained observer visits the family home during a period when the child is awake, active, and interacting freely with the primary caregiver. The observation typically spans between two and four hours, or is divided across two separate visits to account for situational fluctuations in the child's state. Observers are instructed to remain unobtrusive, taking field notes while refraining from driving or structuring the interpersonal interaction.
Following data collection, the observer performs the card-sorting procedure. To minimize cognitive overload, sorting is performed through a three-stage sequential technique:
- The observer initially divides the 90 cards into three preliminary piles of 30 cards each: uncharacteristic behaviors (piles 1–3), neutral/ambiguous behaviors (piles 4–6), and characteristic behaviors (piles 7–9).
- Each of these broad piles is further subdivided into three distinct categories based on relative behavioral salience.
- The observer systematically redistributes cards until precisely 10 cards occupy each pile from 1 through 9. This forced, unimodal distribution guarantees uniform statistical variance across all assessed subjects.
In the final mathematical scoring phase, each item is assigned a numerical value equal to the pile number in which it was placed (1 to 9). The subject child's 90-item array is then compared to the standardized expert criterion array (published by Waters) by computing a Pearson product-moment correlation coefficient (r). This calculation produces a continuous score bounded between -1.00 (the polar opposite of security) and +1.00 (a profile perfectly identical to the ideal secure child). Inter-rater reliability between two independent home observers typically exceeds r = .80 when observers have undergone formal coding training.
10. Applications & Practical Significance
The practical utility of the Attachment Q-Set extends across developmental research, clinical infant mental health, child welfare evaluations, and educational settings. In clinical psychology and early intervention, the AQS is widely celebrated for its clinical sensitivity in evaluating programs like the Video-feedback Intervention to Promote Positive Parenting (VIPP). Because the AQS yields a continuous, granular metric rather than a coarse categorical label, clinicians can quantitatively track subtle, incremental improvements in a toddler's secure base behavior following attachment-informed therapeutic interventions.
In foster care and adoption research, the AQS serves as an essential tool for evaluating the repair and reorganization of attachment bonds in children who have suffered early relational trauma, institutional deprivation, or chronic neglect. Because foster parents often report elevated stress during formal laboratory assessments, observing the child in the domestic stability of the foster home allows practitioners to measure whether the child has developed an organized secure base with their new caregiver.
Furthermore, early childhood educators and developmental researchers utilize the AQS to examine secondary attachment bonds formed outside the home, such as those established between toddlers and professional daycare providers or preschool teachers. The flexible nature of the behavioral descriptors permits the evaluation of non-parental secure base phenomena in communal educational classrooms, providing actionable empirical guidance on how classroom caregiver-to-child ratios directly shape infant socioemotional security.
11. Research & Empirical Evidence
The scientific literature regarding the Attachment Q-Set is vast and empirically robust. The definitive milestone in AQS empirical validation was published by Marinus van IJzendoorn, Chris Vereijken, Marian Bakermans-Kranenburg, and Marianne Riksen-Walraven in their 2004 meta-analysis encompassing 139 individual studies and over 13,800 children. This meta-analytic investigation resolved several contentious methodological debates and solidified the instrument's psychometric credentials.
First, the meta-analysis demonstrated that scores derived from trained, independent observer sorts exhibit robust convergent validity with the Strange Situation Procedure. Children categorized as secure in the laboratory Strange Situation scored substantially and statistically significantly higher on the AQS home security dimension than their insecurely classified peers (with an average effect size of d = 0.63). Second, the authors confirmed that observer-administered AQS scores demonstrate substantial predictive validity across childhood, prospectively predicting children's later social competence with peers, emotional self-regulation, language proficiency, and executive functioning into middle childhood and adolescence.
Crucially, the 2004 meta-analysis revealed a vital methodological caveat regarding the identity of the sorter. When the AQS was sorted by trained independent observers, it consistently demonstrated strong convergent and discriminant validity, exhibiting appropriate associations with maternal sensitivity (r ≈ .39) and minimal confounding with child temperament. Conversely, when the cards were sorted by the children's own mothers, the convergent validity with Strange Situation classifications was significantly compromised, and correlations with maternal sensitivity dropped markedly. Mothers systematically rated their children as more secure due to social desirability bias, defensive idealization, and an untrained inability to distinguish normative dependency from authentic attachment security. Consequently, modern psychometric consensus strictly mandates that the AQS be administered by trained external observers in empirical research.
12. Cultural & Cross-Cultural Considerations
A fundamental critique often leveled against attachment theory is its potential Western ethnocentrism, raising concerns that the definition of a "secure child" may simply reflect middle-class Anglo-American parenting ideals. The Attachment Q-Set has played a pivotal role in empirically testing and refuting this assumption.
In an influential international collaboration led by Germán Posada and colleagues (1995, 2013), expert attachment researchers and child-development professionals from diverse nations—including China, Colombia, Germany, Israel, Japan, Norway, and the United States—were independently tasked with completing the AQS criterion sort for the "ideal secure child." Strikingly, the correlation coefficients between the expert sorts across these disparate cultural contexts ranged from .80 to .98, demonstrating an astonishing level of cross-cultural consensus regarding the universal core of secure base behavior.
Nevertheless, cultural nuances must be carefully contextualized when observing specific behavioral items. In socio-cultural ecologies where collective caregiving (alloparenting) is normative, or where physical proximity is sustained continuously (such as the traditional practices of the Gusii of Kenya or indigenous communities in South America), items penalizing physical clinging or scoring distal exploration may require sensitive interpretation. Researchers must differentiate between culturally normative proximity maintenance and individual behavioral insecurity, reinforcing the imperative that observers understand the local socio-cultural context before scoring the AQS.
13. Criticisms, Debates & Limitations
Despite its extensive empirical utility, the Attachment Q-Set faces notable methodological criticisms and operational limitations. Chief among these is its high resource burden. Conducting two to four hours of in-home naturalistic observation per participant, followed by a meticulous 90-card sorting protocol, requires massive temporal and financial expenditures. In large-scale epidemiological studies tracking thousands of mother-infant dyads, the sheer logistics of deploying trained observers to domestic households frequently leads researchers to default to less ecologically valid laboratory procedures or abbreviated parent-report questionnaires.
A second major theoretical limitation is the difficulty of differentiating between discrete subtypes of insecure attachment. While the laboratory Strange Situation explicitly sorts children into discrete classifications—Insecure-Avoidant (A), Insecure-Resistant (C), and Insecure-Disorganized (D)—the primary AQS output is a single, continuous security-insecurity correlation metric (r). Although supplementary criterion sorts have been constructed for dependency and avoidance, the standard AQS struggles to reliably identify disorganized/disoriented attachment behaviors (such as anomalous freezing, asymmetric postures, or apprehension toward the caregiver), which are critical markers of early clinical psychopathology and severe trauma.
Finally, controversies persist regarding the influence of infant temperament. Critics contend that certain items in the 90-card deck (e.g., items measuring fussiness, irritability, fear of novel stimuli, or physical sociability with adult strangers) tap into innate biological temperament rather than relational attachment quality. While Everett Waters and subsequent psychometricians have demonstrated that statistical partialling of temperament maintains the predictive validity of the AQS, researchers must exercise meticulous care during sorting to prevent constitutional reactivity from artificially depressing a child's security index.
14. Related Terms & Distinctions
To avoid conceptual confusion, the Attachment Q-Set must be demarcated from related psychological constructs and assessment instruments:
- Strange Situation Procedure (SSP): A 20-minute structured laboratory paradigm involving mild, cumulative stress through two brief separations and reunions with the mother and a friendly stranger. Whereas the SSP yields categorical classifications (A, B, C, D) based primarily on reunion behaviors, the AQS yields a continuous dimensional score based on extensive naturalistic home observations.
- Adult Attachment Interview (AAI): A semi-structured, 18-question clinical interview developed by Mary Main and colleagues that evaluates an adolescent's or adult's internal state of mind with respect to their childhood attachment experiences, emphasizing discourse coherence. In contrast, the AQS evaluates real-time, non-verbal and pre-verbal behavioral organization in early childhood.
- Maternal Sensitivity Scales (Ainsworth): Four 9-point observational rating scales that evaluate the caregiver's capacity to perceive, accurately interpret, and promptly respond to the infant's communicative signals. The AQS assesses the child's behavioral response and secure base utilization, providing an assessment of the child's relational organization rather than maternal behavior directly.
- Attachment Security vs. Emotional Dependency: A foundational theoretical distinction. Dependency refers to an immature, non-specific clinging to adults and inability to act autonomously regardless of safety. Attachment security involves the targeted, adaptive use of a specific caregiver as a launchpad for autonomous exploratory mastery.
- Q-Methodology: The broader psychometric family developed by William Stephenson utilizing forced-choice, ipsative card sorting to analyze subjectivity and intra-individual patterns. The AQS is a specialized developmental application of this broad psychometric paradigm.
15. Summary & Key Takeaways
The Attachment Q-Set represents a triumph of translational psychometrics in developmental science, establishing an empirically robust bridge between naturalistic ethological observation and quantitative measurement. By transforming 90 discrete descriptions of infant-caregiver domestic interaction into an ipsative profile compared against an international expert consensus of the ideal secure child, the AQS measures the child's ability to use their primary caregiver as a secure base for autonomous exploration and a safe haven during distress. Extensively validated by meta-analytic research encompassing tens of thousands of dyads, it provides an indispensable, ecologically valid complement to laboratory paradigms, illuminating how early relational attachments are constructed, sustained, and expressed in the fabric of everyday life.
References
- Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Lawrence Erlbaum Associates.
- Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment (2nd ed.). Basic Books. (Original work published 1969).
- Posada, G., Gao, Y., Wu, F., Posada, R., Tascon, M., Schöelmerich, A., Sagi, A., Kondo-Ikemura, K., Haaland, W., & Normandin, M. P. (1995). The secure-base phenomenon across cultures: Children's behaviors, mothers' preferences, and experts' concepts. Monographs of the Society for Research in Child Development, 60(2-3), 27–48. https://doi.org/10.1111/j.1540-5834.1995.tb00202.x
- van IJzendoorn, M. H., Vereijken, C. M., Bakermans-Kranenburg, M. J., & Riksen-Walraven, J. M. (2004). Assessing attachment security with the Attachment Q-Sort: Meta-analytic evidence for its validity. Child Development, 75(4), 1188–1213. https://doi.org/10.1111/j.1467-8624.2004.00733.x
- Waters, E. (1995). The Attachment Q-Set (Version 3.0). In E. Waters, B. E. Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural, and cognitive perspectives on secure-base behavior and working models: New growing points of attachment theory and research. Monographs of the Society for Research in Child Development, 60(2-3), 234–246.
- Waters, E., & Deane, K. E. (1985). Defining and assessing individual differences in attachment relationships: Q-methodology and the organization of behavior in infancy and early childhood. Monographs of the Society for Research in Child Development, 50(1-2), 41–65. https://doi.org/10.2307/3333826