Earned Secure Attachment: What It Means and What the Evidence Actually Shows
Author: Ukrainian Psychological Hub · Published: September 22, 2026 · Editorial Policy
Earned secure attachment is a research concept for adult attachment security that appears despite a history described as difficult, inadequate, or insecure. The term is often presented online as the end point of “healing your attachment style.” The scientific literature is more specific. Earned security has historically been studied mainly through the Adult Attachment Interview (AAI), where researchers examine how coherently an adult organizes and discusses attachment-related childhood experiences. A major 2026 scoping review found only 24 empirical studies through 2023 and concluded that the field shows more formal than substantive agreement about what earned security is and how it should be measured; because the studies are heterogeneous, many conclusions remain preliminary. Filosa et al., 2026
That distinction matters. “Earned secure” is not a diagnosis, not a recognized personality type, and not simply a synonym for moving from an anxious or avoidant romantic attachment score to a secure one. Adult attachment research contains several measurement traditions that overlap conceptually but do not measure the same thing. The AAI assesses a person’s current state of mind regarding childhood attachment experiences; common self-report instruments assess anxiety and avoidance in current close relationships. A meta-analytic study found only a trivial-to-small average association between AAI security and self-reported attachment dimensions, which is one reason the two traditions should not be treated as interchangeable. Roisman et al., 2007
What does earned secure attachment mean?
In its classic research usage, earned security describes a discrepancy: an adult shows a secure or autonomous organization of attachment-related discourse while also reporting notably unfavorable early caregiving experiences. Early work distinguished “earned-secure” adults from “continuous-secure” adults, who also showed a secure current state of mind but described comparatively favorable childhood caregiving. Pearson et al., 1994
The crucial part of that definition is current organization, not the idea that a person has erased the past. In AAI-based work, security is inferred from qualities such as coherence, balance, and the ability to discuss attachment experiences in a reasonably integrated way. An earned-secure classification therefore does not mean that childhood adversity was harmless, that painful memories disappeared, or that the person never becomes distressed in relationships.
The 2026 scoping review uses a broad formulation: earned-secure attachment concerns people who move beyond adverse or insecure childhood attachment experiences toward secure relationship functioning in adulthood. At the same time, the review emphasizes that individual studies have operationalized this idea in different ways. Some rely on AAI classifications, some combine current security with retrospective reports of poor care, and newer studies may use self-report measures rather than the original interview-based approach. Filosa et al., 2026
Earned security grew out of the Adult Attachment Interview tradition
The Adult Attachment Interview and romantic attachment questionnaires belong to related but distinct research traditions. A broad review of adult attachment measurement identified the AAI alongside many self-report instruments, including the Experiences in Close Relationships measures and relationship-specific scales. The instruments differ in relationship focus, constructs, administration, scoring, and the kind of attachment information they are designed to capture. Ravitz et al., 2010
The AAI asks adults about childhood relationships with caregivers and evaluates the organization of the resulting discourse. Self-report romantic attachment research typically locates people on dimensions of attachment anxiety and attachment avoidance. Those dimensions are often summarized with labels such as secure, anxious/preoccupied, dismissive-avoidant, and fearful-avoidant, but the dimensions are the stronger scientific model for many adult romantic attachment questions. For the broader scientific model and measurement framework, see Adult Attachment Theory: How Anxiety and Avoidance Shape Relationships.
This is why an internet statement such as “I used to be anxiously attached, did therapy, and now I am earned secure” may describe a meaningful personal change without reproducing the original research definition. It could refer to lower self-reported attachment anxiety, a different relationship-specific pattern, improved emotion regulation, a secure AAI classification, or some combination of these. Those are related possibilities, yet they are not measurement equivalents.
Earned secure versus continuous secure attachment
The earned-secure versus continuous-secure distinction was developed to ask an important developmental question: can adults who report difficult early relationships nevertheless show an organized, secure attachment state of mind later? Early studies suggested that some could. Pearson and colleagues compared adults categorized as earned secure and continuous secure and examined depressive symptoms and parenting-related variables. Pearson et al., 1994
The distinction is useful, but it is not as simple as “bad childhood plus good adulthood equals earned secure.” A retrospective classification depends partly on how childhood experiences are remembered and rated in adulthood. It does not automatically prove that the person was insecurely attached as an infant, nor does it prove a particular developmental pathway from insecurity to security.
The most important challenge: retrospective earned security may not prove developmental change
A 23-year longitudinal study by Roisman and colleagues directly tested a key assumption behind retrospectively defined earned security. The researchers asked whether adults classified as earned secure because they coherently described negative childhood experiences actually had documented histories of insecure infant attachment or consistently harsh and ineffective parenting. The answer was more complicated than the construct’s popular interpretation suggests. Roisman et al., 2002
In that high-risk longitudinal sample, retrospectively defined earned-secure adults were not more likely than continuous-secure adults to have been anxiously attached in infancy. They had also experienced relatively supportive and structured maternal parenting across childhood and adolescence. By contrast, participants who could be identified prospectively as changing from insecure infant attachment toward later security did show evidence consistent with genuine developmental change and later success in close relationships. Roisman et al., 2002
That finding does not make earned security meaningless. It changes what can be inferred from a retrospective label. A secure adult who remembers difficult caregiving may indeed have undergone major relational development, but the classification alone cannot reconstruct the person’s infant attachment history with certainty.
A later experiment added another caution. Among adults who were secure on the AAI, experimentally induced sadness made an earned-secure classification more likely, whereas a happy mood made a continuous-secure classification more likely; the mood manipulation did not determine whether people were secure versus insecure overall. The study raised the possibility that current mood can influence the valence of autobiographical reports used to distinguish earned from continuous security. Roisman, 2006
What the 2026 scoping review actually found
The strongest current synthesis is the scoping review by Filosa, Sharp, Gori, and Musetti, published in the June 2026 issue of Psychological Reports. The review was preregistered, searched Scopus, PsycInfo, PubMed, and Web of Science, covered research published through 2023, included 24 studies, and appraised their quality with the Mixed Methods Appraisal Tool. Filosa et al., 2026
Its central conclusion is unusually important because it pushes against the certainty of popular attachment content: the field agrees on the broad idea of earned security more than it agrees on a single substantive definition, measurement strategy, or developmental pathway. The included findings were heterogeneous enough that the authors characterized current conclusions as preliminary. Filosa et al., 2026
The review identified secondary attachment figures and reflective functioning as promising contributors to earned security. At the same time, several areas remained poorly defined, including developmental factors, the relation between earned security and depressive symptoms, and the role of fathers. That combination—promising mechanisms plus substantial unresolved questions—is a better summary of the evidence than a fixed recipe for “earning security.” Filosa et al., 2026
What studies suggest about supportive relationships
One influential study examined 121 first-time expectant mothers using the AAI. Sixteen were classified as earned secure. Compared with insecure and continuous-secure participants, the earned-secure group reported more emotional support from alternative support figures, though not more instrumental support. They had also spent more time in therapy, and they were more likely than insecure participants to have infants classified as secure in the Strange Situation at 12 to 15 months. Saunders et al., 2011
These findings are consistent with the idea that emotionally supportive relationships outside the original caregiver relationship may matter. They do not show that one supportive person causes earned security, and they do not establish that therapy caused the group difference. The study was observational, the earned-secure group was small, and the classification was retrospective.
A more recent study used a different operationalization. Jańczak studied 272 women and identified an “earned secure” group as adults reporting secure attachment to their mothers despite low reported maternal care. The earned-secure and continuously secure groups showed better emotion regulation and some aspects of mentalization than insecure and “lost secure” groups. Alternative support figures were not more common in the earned-secure group, but relationships with those figures were described as more secure. The groups did not differ in time spent in psychotherapy. Jańczak, 2024
The difference between the Saunders and Jańczak findings illustrates the field’s central problem: studies can define earned security differently and therefore may not be studying exactly the same phenomenon. This is one reason a single checklist of causes or signs outruns the evidence.
Reflective functioning, narrative integration, and emotion regulation
Reflection is a recurring theme in the earned-security literature. The 2026 review highlighted reflective functioning as a potentially important contributor, while Jańczak found group differences in some mentalization and emotion-regulation measures. Filosa et al., 2026 Jańczak, 2024
In plain language, the evidence is compatible with a process in which a person becomes better able to represent attachment experiences in a coherent, differentiated way: recognizing what happened, understanding that relationships can contain mixed motives and emotions, regulating activation without erasing it, and revising expectations in light of new experience. That is a plausible interpretation of the research pattern. It is not a validated step-by-step intervention protocol for earned security.
A small qualitative study of 20 adults who reported histories of insecurity and evidence of later security generated a grounded-theory model involving broader conditions for change, intrapersonal changes, and interpersonal changes. The study is useful for understanding how people narrate positive attachment change, but its design was qualitative and its sample was small; it cannot establish a universal causal sequence. Dansby Olufowote et al., 2020
Does therapy create earned secure attachment?
Therapy is often presented online as the primary route to earned security. The evidence supports a more measured statement: psychotherapy may be one context in which attachment-related change occurs, and some earned-security studies report associations with therapy, but the classic earned-security literature does not establish one therapy as a proven method for producing the construct.
In Saunders and colleagues’ sample, earned-secure women reported more time in therapy than insecure and continuous-secure women. That association is notable but does not demonstrate that therapy caused the classification. In Jańczak’s study, time in psychotherapy did not differ between the earned-secure and comparison groups. Saunders et al., 2011 Jańczak, 2024
A 2024 psychotherapy study used the phrase “earned secure therapeutic attachment,” but it measured the development of a secure attachment to the therapist during treatment and associations with therapy process and interpersonal outcomes. That is a therapist-specific construct and should not be treated as identical to the classic AAI-based earned-secure classification. Jacobsen et al., 2024
For readers considering therapy, the practical implication is straightforward: therapy can be used to work on relationship expectations, emotion regulation, trauma-related responses, communication, and ways of making sense of past relationships. Whether those changes should be called “earned secure attachment” depends on the construct and measurement being used. A person does not need to acquire a research label for therapy to be useful. For practical work specifically on high attachment avoidance and deactivating patterns, see How to Heal Avoidant Attachment: Evidence-Based Ways to Build More Security.
Earned security is not the same question as whether attachment style can change
The broader question of whether attachment style can change belongs to a larger literature. Longitudinal research shows both continuity and change. In a study following more than 4,000 people across repeated assessments, many life events were associated with immediate shifts in adult attachment, but enduring average changes were less common; there were also substantial individual differences, with some people becoming more secure and others less secure after the same kinds of events. Fraley et al., 2021
Across development, attachment also shows meaningful but incomplete stability. A meta-analysis spanning infancy to early adulthood found a medium overall stability association and much weaker evidence of stability across the longest intervals. Pinquart et al., 2013
These findings support a model in which attachment-related tendencies have continuity without being immutable. They do not turn every shift in self-reported anxiety or avoidance into “earned security.” A person can become less anxious or less avoidant in romantic relationships without ever being assessed with the AAI, and a person classified as secure-autonomous on the AAI can still experience insecurity in a particular current relationship.
Can someone be earned secure and still feel anxious or avoidant sometimes?
Yes. Security is not the absence of all attachment activation. Close relationships naturally involve separation, uncertainty, conflict, dependence, and loss, all of which can activate attachment concerns. The scientific question is how those concerns are organized and regulated, how flexible a person can be, and which measurement system is being used.
This is especially important because the AAI and self-report attachment measures show limited empirical overlap. In Roisman and colleagues’ meta-analytic study, the mean association between AAI security and self-report attachment dimensions was about r = .09. The measures therefore capture substantially distinct aspects of attachment organization and relationship experience. Roisman et al., 2007
Can you identify earned secure attachment from “signs”?
There is no authoritative diagnostic checklist for earned secure attachment. The term is not a DSM or ICD diagnosis, and the research construct was not created as a consumer personality label. Online lists of “earned secure signs” may describe behaviors associated with healthy relationship functioning—tolerating closeness, communicating needs, recovering after conflict, setting boundaries, accepting support, or reflecting on the past—but those behaviors alone do not establish an earned-secure research classification.
The strongest reason for caution is measurement. Adult attachment instruments differ substantially in what they assess, and even measures that use the word “security” may operationalize it differently. Ravitz et al., 2010 A person should therefore treat a quiz result as information about the construct that quiz actually measures, not as proof of a complete developmental history.
What earned secure attachment does not mean
Earned security does not mean that a difficult childhood has been disproved, forgiven, or made irrelevant. It does not require reconciliation with caregivers. It does not imply that every painful response has disappeared. It does not establish that someone is “fully healed,” and it is not a guarantee of a conflict-free relationship.
It also does not mean that insecure adult attachment is an attachment disorder. Adult romantic attachment anxiety and avoidance are research constructs. Reactive Attachment Disorder and Disinhibited Social Engagement Disorder are separate clinical diagnoses with different criteria and developmental contexts. Earned security should not be used as an informal diagnostic opposite of those disorders.
Finally, earned security does not come with a scientifically established timeline. Claims that people can become earned secure in 30 days, 90 days, or another standardized period go beyond the evidence. The literature contains different definitions, designs, populations, and time frames; the 2026 synthesis explicitly describes current conclusions as preliminary. Filosa et al., 2026
What may support movement toward greater security?
The evidence is most useful when it is translated into processes rather than promises. Supportive relationships, the quality of emotional support from alternative attachment figures, reflective functioning, mentalization, emotion regulation, and psychotherapy all appear in parts of the literature. Their evidentiary status differs, and none constitutes a guaranteed route to an earned-secure classification. Filosa et al., 2026 Saunders et al., 2011 Jańczak, 2024
In current relationships, feeling understood, valued, and cared for is studied under the broader construct of perceived partner responsiveness. That literature is relevant to how close relationships become emotionally supportive, although it should not be treated as direct evidence that partner responsiveness causes earned security.
Similarly, learning to disclose, respond, repair, and stay emotionally engaged can strengthen emotional intimacy in marriage. These relationship processes can matter in their own right. They do not need to be reframed as a test of whether someone has “achieved” a particular attachment label.
A more accurate way to think about earned security
The most defensible interpretation is that earned security names a scientifically interesting form of developmental discontinuity: present attachment security in the context of an unfavorable early attachment history. The concept asks how secure organization can emerge when early experience would seem to predict something else.
Its scientific value lies precisely in that question. The answer remains incomplete. Research suggests that later supportive relationships and reflective capacities may matter; some studies find differences in emotion regulation, mentalization, therapy exposure, parenting, or offspring attachment. Yet definitions vary, retrospective classification has known interpretive problems, and the evidence base remains small enough that causal claims should be restrained. Filosa et al., 2026
For an individual reader, the constructive conclusion is not “find your earned-secure status.” It is that attachment-related patterns can show meaningful development across adulthood, while the specific label earned secure refers to a narrower and more contested research construct than popular psychology often implies.
Frequently Asked Questions
Is earned secure attachment a real research concept?
Yes. Earned security has been studied for decades and was the subject of a dedicated 2026 scoping review of 24 empirical studies. The concept is established enough to have a recognizable research literature, while its definition, measurement, and causal interpretation remain heterogeneous. Filosa et al., 2026
Is earned secure attachment a diagnosis?
No. Earned secure attachment is a research construct, not a DSM or ICD diagnosis. It should not be used to diagnose a person, and insecure adult romantic attachment should not be called an attachment disorder.
How is earned secure attachment measured?
Historically, much of the construct comes from the Adult Attachment Interview tradition: researchers identify a secure-autonomous state of mind while also coding reports of notably unfavorable early caregiving. Later studies have used other combinations of interview and self-report measures. The 2026 review specifically identifies this variation in definition and assessment as a major issue in the literature. Filosa et al., 2026
Is earned secure attachment the same as becoming securely attached in romantic relationships?
Not necessarily. Romantic attachment is commonly measured with self-reported anxiety and avoidance. Classic earned-security research is largely tied to the AAI and representations of childhood attachment experiences. Those measurement traditions are related but only weakly correlated on average. Roisman et al., 2007
Can someone with anxious or avoidant attachment become more secure?
Adult attachment anxiety and avoidance can change. Longitudinal studies show both stability and change, including enduring change for some people following life events. That broader evidence supports plasticity, but it is not the same as demonstrating the classic earned-secure construct. Fraley et al., 2021
Does therapy cause earned secure attachment?
The evidence does not justify that causal claim. Some studies find more therapy exposure among retrospectively defined earned-secure participants, while others do not find such differences. Therapy can support many clinically meaningful changes, but no single therapy has been established as a universal procedure for producing earned-secure status. Saunders et al., 2011 Jańczak, 2024
How long does it take to develop earned security?
There is no validated universal timeline. The literature does not support a fixed number of days, weeks, sessions, or months. Attachment-related change depends on what construct is being measured, the person, the relationship context, life events, and the time frame of assessment.
Can a person be secure in one relationship and insecure in another?
Yes. Adult attachment can be measured globally or in relation to specific people. Relationship-specific anxiety and avoidance can differ across partners, parents, friends, and other attachment figures. That variability is another reason to avoid treating attachment style as a fixed personality type.
Related Articles
References
Dansby Olufowote, R. A., Fife, S. T., Schleiden, C., & Whiting, J. B. (2020). How can I become more secure?: A grounded theory of earning secure attachment. Journal of Marital & Family Therapy, 46(3), 489–506. https://doi.org/10.1111/jmft.12409
Filosa, M., Sharp, C., Gori, A., & Musetti, A. (2026). A comprehensive scoping review of empirical studies on earned secure attachment. Psychological Reports, 129(3), 1807–1832. https://doi.org/10.1177/00332941241277495
Fraley, R. C., Gillath, O., & Deboeck, P. R. (2021). Do life events lead to enduring changes in adult attachment styles? A naturalistic longitudinal investigation. Journal of Personality and Social Psychology, 120(6), 1567–1606. https://doi.org/10.1037/pspi0000326
Jańczak, M. O. (2024). Mentalization, emotional dysregulation and attachment to alternative attachment figures in retrospectively defined earned secure adults. Current Issues in Personality Psychology, 12(1), 30–40. https://doi.org/10.5114/cipp/172328
Jacobsen, C. F., Falkenström, F., Castonguay, L., Nielsen, J., Lunn, S., Lauritzen, L., & Poulsen, S. (2024). The relationship between attachment needs, earned secure therapeutic attachment and outcome in adult psychotherapy. Journal of Consulting and Clinical Psychology, 92(7), 410–421. https://doi.org/10.1037/ccp0000900
Pearson, J. L., Cohn, D. A., Cowan, P. A., & Cowan, C. P. (1994). Earned- and continuous-security in adult attachment: Relation to depressive symptomatology and parenting style. Development and Psychopathology, 6(2), 359–373. https://doi.org/10.1017/S0954579400004636
Pinquart, M., Feußner, C., & Ahnert, L. (2013). Meta-analytic evidence for stability in attachments from infancy to early adulthood. Attachment & Human Development, 15(2), 189–218. https://doi.org/10.1080/14616734.2013.746257
Ravitz, P., Maunder, R., Hunter, J., Sthankiya, B., & Lancee, W. (2010). Adult attachment measures: A 25-year review. Journal of Psychosomatic Research, 69(4), 419–432. https://doi.org/10.1016/j.jpsychores.2009.08.006
Roisman, G. I., Padrón, E., Sroufe, L. A., & Egeland, B. (2002). Earned-secure attachment status in retrospect and prospect. Child Development, 73(4), 1204–1219. https://doi.org/10.1111/1467-8624.00467
Roisman, G. I., Fortuna, K., & Holland, A. (2006). An experimental manipulation of retrospectively defined earned and continuous attachment security. Child Development, 77(1), 59–71. https://doi.org/10.1111/j.1467-8624.2006.00856.x
Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., & Clarke, A. (2007). The Adult Attachment Interview and self-reports of attachment style: An empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–697. https://doi.org/10.1037/0022-3514.92.4.678
Saunders, R., Jacobvitz, D., Zaccagnino, M., Beverung, L. M., & Hazen, N. (2011). Pathways to earned-security: The role of alternative support figures. Attachment & Human Development, 13(4), 403–420. https://doi.org/10.1080/14616734.2011.584405