Documented Q&A
Yes—but with an important qualification. Adults can receive an earned-secure classification when their Adult Attachment Interview is coherent despite an inferred negative childhood history. However, the evidence does not establish that everyone so classified was insecure in childhood and later “rewrote” their attachment style. Some findings support meaningful change; others suggest that negative retrospective accounts, including depression-linked recall, may partly create the category.
ISSPR.org editorial team: What does “earned secure attachment” actually mean in attachment research?
Documented answer: In its technical sense, earned security is an Adult Attachment Interview, or AAI, classification. A transcript is classified secure/autonomous when the speaker discusses attachment experiences in a coherent, balanced and internally consistent way—even when the developmental history inferred from the interview is negative.
The crucial point is that classification turns on the coherence of the narrative, not whether the childhood described was happy. A person may recount rejection, inconsistency or other difficult experiences while remaining relevant, credible and reflective. AAI coherence is generally described as drawing on Grice’s conversational maxims of quality, quantity, relation and manner, although that specific attribution is better established in secondary descriptions than in the primary manual.
The underlying AAI coding manual by Main and Goldwyn was an unpublished manuscript circulated from the 1980s and commonly cited across the 1984–1998 period; it was never formally published in a journal. That history matters because popular definitions often make earned security sound like a universally standardized clinical diagnosis. It is more accurately understood as a specialized research classification.
Readers new to the broader framework can begin with our overview of attachment theory. Neither a checklist nor a person’s preferred attachment label is equivalent to an AAI classification.
ISSPR.org editorial team: Does an earned-secure classification prove that somebody overcame an insecure childhood?
Documented answer: No. This is the central scientific controversy.
Roisman, Padrón, Sroufe and Egeland’s 2002 study in Child Development followed participants prospectively over 23 years. Unlike studies relying only on adults’ memories, it included caregiving and attachment information collected during childhood. The researchers found that adults defined retrospectively as earned secure were not more likely than continuous-secure adults to have been anxiously attached in infancy. Moreover, they had received some of the sample’s most supportive and structured maternal parenting during childhood and adolescence.
That finding directly complicates the familiar recovery story. If an adult gives a coherent AAI account but describes childhood negatively, the discrepancy may not always represent genuine early insecurity followed by transformation. It may partly reflect negative recall bias, potentially associated with depressive symptoms.
Yet the study does not prove that earned security is fictitious. Participants identified as earned secure using prospective developmental information showed successful close relationships in young adulthood. The dispute concerns what the retrospective category measures and how often it captures real adversity overcome.
Taxometric research by Roisman and colleagues adds another complication: inferred developmental experience among secure adults appears to be continuously distributed, with little evidence that earned-secure and continuous-secure adults form two sharply distinct kinds of people.
| Research question | What the evidence supports | What it does not establish |
|---|---|---|
| Can a difficult childhood coexist with a coherent adult attachment narrative? | Yes; that is the conceptual basis of earned security in AAI coding. | That every negative retrospective account accurately identifies childhood insecurity. |
| Are earned-secure adults capable of effective relationships or parenting? | Yes; several findings indicate effective functioning. | That their histories and emotional burdens are identical to those of continuous-secure adults. |
| Can attachment-related responses change? | Yes, especially in particular relationships or temporarily activated states. | That global attachment organization is always permanently overwritten. |
| Are earned-secure adults a distinct category? | Taxometric research suggests the boundary may be continuous rather than categorical. | That the label has no descriptive or research value. |
ISSPR.org editorial team: What evidence most strongly supports a genuine pathway toward earned security?
Documented answer: One of the clearest findings concerns sustained emotional support from people other than parents.
Saunders, Jacobvitz, Zaccagnino, Beverung and Hazen’s 2011 study in Attachment & Human Development examined 121 mothers. Sixteen were classified as earned secure: 25% of the secure/autonomous participants and 13% of the total sample. This denominator is important. The study does not show that 25% of insecure adults become earned secure.
Earned-secure mothers reported substantially more emotional support from alternative, non-parental figures than both insecure and continuous-secure mothers. The same pattern did not appear for instrumental support, and the emotional-support effects remained after depressive symptoms were statistically controlled. This specificity suggests that feeling understood, cared for or emotionally supported may matter differently from receiving practical assistance.
Pearson, Cohn, Cowan and Cowan’s 1994 study in Development and Psychopathology also found a meaningful combination of strength and continuing vulnerability. Earned-secure women parented as effectively as women classified as continuously secure, but they reported significantly more depressive symptoms on the CES-D scale.
Together, these findings fit a nuanced account: adults with difficult histories may develop coherent, effective ways of relating, potentially with help from alternative attachment figures, while still carrying emotional costs. Earned security should not be interpreted as total erasure of distress or developmental history.
ISSPR.org editorial team: Is adult attachment fixed, or can it actually change?
This question connects directly to broader work on attachment styles beyond romance, where the same malleability debate applies to friendships and professional relationships.
Documented answer: The evidence supports neither absolute permanence nor unrestricted reinvention.
Fraley’s 2002 meta-analysis and dynamic modeling in Personality and Social Psychology Review found moderate attachment stability across the first 19 years of life. The pattern fit a prototype model better than a purely revisionist model. In a prototype account, an underlying organization remains relatively stable while observable attachment responses fluctuate with later circumstances. Early representations are not simply deleted and replaced.
Research on adult attachment stability by Fraley and colleagues, published in 2011, also indicates that test–retest stability is higher in adulthood than in childhood when equivalent time intervals are compared. Change remains possible, but continuity exerts a measurable pull.
Mikulincer and Shaver likewise describe adult attachment as dynamic rather than completely fixed. Security-priming experiments activate security-related mental representations and produce short-term changes in cognition or behavior—for example, increased compassion toward a suffering person in research reported by Mikulincer and colleagues in 2005.
But a temporary attachment state is not a lasting attachment style. A successful security prime demonstrates momentary malleability, not permanent reorganization. The most defensible conclusion is therefore that attachment can change partially, contextually and sometimes durably, while earlier patterns may remain influential.
ISSPR.org editorial team: Does Emotionally Focused Couple Therapy turn insecure adults into secure adults?
Documented answer: That claim goes beyond the available evidence.
Burgess Moser, Johnson, Dalgleish, Lafontaine, Wiebe and Tasca’s 2016 process study in the Journal of Marital and Family Therapy followed 32 couples receiving Emotionally Focused Couple Therapy, or EFT. Relationship-specific attachment avoidance decreased significantly. Among couples completing the therapeutic event called “blamer softening,” relationship-specific attachment anxiety also decreased.
These are relevant findings, but the measured construct was attachment toward that particular partner. It was not global or dispositional attachment style, and it was not an AAI state-of-mind classification. Conflating relationship-specific change with a wholesale transformation of adult attachment is one of the most common errors in popular descriptions of EFT.
The analysis also had no control group and was a process study, not a randomized trial demonstrating global attachment change. It can show that attachment-related experience within a couple changed during treatment; it cannot by itself prove that participants became earned secure.
A 2016 review by Wiebe and Johnson reports the widely cited estimate that 70–75% of distressed couples move to recovery with EFT, with gains maintained at two-year follow-up. That supports EFT as a relationship intervention, but treatment recovery and earned-secure AAI classification remain different outcomes. Improvement with a partner is valuable without being relabeled as complete dispositional change.
ISSPR.org editorial team: How should adults evaluate claims that they can “become securely attached”?
Documented answer: First identify what kind of security is being discussed. Popular language often merges several non-equivalent outcomes: a coherent AAI narrative, lower anxiety with one partner, a short-term security state, effective parenting and a stable global attachment style.
A practical evidence-checking sequence is:
- Name the measure. Was security assessed with an AAI, a relationship-specific measure or a temporary experimental prime?
- Check the time scale. Did the effect last minutes, through treatment or across years?
- Separate functioning from classification. Better communication or parenting can matter even if no AAI reclassification was measured.
- Examine how childhood was assessed. Prospective observations and retrospective adult memories can produce different conclusions.
- Check the comparison group. Earned-secure adults may resemble continuous-secure adults in one domain but differ in depressive symptoms or support histories.
- Verify the denominator. In Saunders et al. (2011), 25% referred to the secure/autonomous participants, not the proportion of insecure adults who later became secure.
- Avoid all-or-nothing promises. Prototype research supports meaningful change alongside persistent continuity.
Someone who recognizes themselves in descriptions of an anxious attachment style should therefore treat that label as a framework, not an irreversible identity. Likewise, lists of signs of secure attachment cannot substitute for longitudinal evidence or formal AAI coding.
The scientifically careful answer is hopeful but bounded: adults may build safer relationships, more coherent narratives and more effective patterns. Research does not justify promising complete erasure of earlier attachment organization.
Frequently asked questions
1. Is earned security an official mental-health diagnosis?
No. It is a research classification derived from the Adult Attachment Interview coding system. It refers to a secure/autonomous, coherent transcript combined with an inferred negative developmental history. The foundational coding manual circulated as an unpublished manuscript rather than a journal publication.
2. Can someone be earned secure and still experience depression?
Yes. Pearson and colleagues (1994) found that earned-secure women parented as effectively as continuous-secure women but reported significantly more depressive symptoms. Security of attachment discourse does not mean the absence of emotional distress.
3. Does supportive parenting rule out later earned security?
The issue is how earned security is defined. In Roisman et al.’s 2002 longitudinal study, some adults retrospectively classified as earned secure had actually received highly supportive and structured maternal parenting. This raises questions about whether retrospective negativity always reflects objectively adverse early care.
4. Can a supportive partner change attachment?
Relationship-specific attachment can change. The 2016 EFT process study found reductions in avoidance toward the specific partner, with anxiety also decreasing among couples completing blamer softening. It did not test permanent global attachment change or earned-secure AAI status.
5. Does security priming prove that attachment style can be permanently rewritten?
No. Security priming demonstrates short-term changes in attachment-related states, cognition or behavior. It shows that the attachment system is responsive, but it does not establish lasting change in dispositional style.
6. So, can adults truly become earned secure?
Possibly—but the label should not be treated as proof of a simple journey from insecure childhood to permanently secure adulthood. The strongest interpretation is that some adults develop coherent and effective attachment functioning despite adversity, while the category may also include people whose negative adult recollections do not match prospectively observed childhood care.