Earned Security: What the Concept Means

Where the idea of earned security comes from, why it is more specific and more controversial than social media suggests, and what current research can cautiously say about developing security after adverse early experience.

“Earned secure” sounds simpler than it is

Online attachment language often uses the phrase like this:

I used to be anxious.

I went to therapy.

Now I am secure.

Therefore:

earned secure.

That is understandable shorthand.

It is not the original research concept.

Earned security emerged from a specific developmental attachment tradition, especially research using the Adult Attachment Interview, or AAI.

The basic idea was striking:

Some adults described difficult or inadequate early caregiving but nevertheless discussed those experiences in the coherent, balanced way associated with secure autonomous attachment representations.

Researchers asked:

How can a person develop security when early experience appears not to have provided it?

That question produced the concept of earned security.

It also produced decades of debate about how the concept should be defined.

Earned security belongs to a different measurement tradition than romantic attachment anxiety

This distinction is essential.

The self report attachment scales popular in romantic relationship research usually measure dimensions such as:

• attachment anxiety;
• attachment avoidance.

The Adult Attachment Interview is different.

It does not ask only how secure someone feels with a current partner.

It evaluates how a person organizes and discusses attachment related experiences, especially early caregiving.

Coding focuses heavily on coherence of mind, not simply whether childhood was described as good or bad.

A major meta analytic comparison found that AAI security and self reported romantic attachment dimensions overlap only weakly.

So an adult saying:

“I used to score high on anxious attachment and now I score lower”

may describe meaningful change.

But that does not automatically establish earned secure status in the AAI research sense.

Different construct.

Different measure.

Different evidence.

The original retrospective definition

Historically, researchers often distinguished two secure groups.

Continuous secure

Adults who were classified secure on the AAI and who described relatively supportive childhood care.

Earned secure

Adults classified secure on the AAI who nevertheless described substantial adversity, rejection, low care or other negative childhood experiences.

The interpretation was that the second group had somehow earned a secure state of mind despite a difficult developmental history.

That interpretation became influential.

It was also difficult to verify.

Why?

Because the definition often depended on adults remembering their childhood and researchers inferring earlier insecurity from those memories.

Retrospective adversity is not the same as prospectively observed infant attachment.

A major longitudinal study challenged the simple story

In 2002, Glenn Roisman and colleagues used a 23 year longitudinal dataset to test a core assumption.

If retrospectively defined earned secure adults truly began with insecure attachment, they should have been more likely to show insecure attachment earlier in life.

That is not what the study found.

Retrospectively defined earned secure adults were not more likely than continuous secure adults to have been anxiously attached in infancy.

They also had experienced relatively supportive and structured maternal parenting in childhood and adolescence within the high risk sample.

This created a serious conceptual problem.

Some adults classified as retrospectively earned secure may not actually have moved from observed childhood insecurity to adult security.

The retrospective label could be mixing several different developmental pathways.

Mood can even influence the retrospective distinction

A 2006 experiment pushed the challenge further.

Secure young adults completed the Adult Attachment Interview after a mood induction.

Among secure participants, those induced into a sad mood were more likely to be classified as retrospectively earned secure.

Those in a happy mood were more likely to be classified as continuous secure.

Importantly, mood did not determine secure versus insecure classification.

It influenced the earned versus continuous distinction among secure adults.

That suggests part of the retrospective classification can be sensitive to how negative early experiences are narrated at the time.

This does not invalidate the concept.

It shows why the concept cannot be treated as a simple historical fact detector.

Prospective earned security is a stronger claim

The strongest version of earned security would identify insecurity earlier in life and later demonstrate security using prospective data.

That avoids relying entirely on adult retrospective description.

In the 2002 longitudinal work, prospectively defined earned secure individuals did show successful close relationships in adulthood.

This supports the general developmental possibility:

early insecurity does not make adult insecurity inevitable.

That may be the most important idea underneath earned security.

But the exact pathways remain less certain than popular psychology suggests.

The first comprehensive review is surprisingly recent

A preregistered scoping review published online in 2024 and in print in 2026 searched four major databases for empirical earned security research.

Only 24 studies met inclusion criteria.

The review found more agreement in the broad formal idea than in the actual details of how earned security had been defined, assessed and explained.

The authors described the conclusions as preliminary because of heterogeneity.

Several important questions remained poorly defined, including:

• developmental factors;
• links with depressive symptoms;
• the role of fathers.

The review highlighted possible roles for:

• secondary attachment figures;
• reflective functioning.

This is a crucial evidence boundary.

Earned security is a legitimate research concept.

It is not a settled developmental formula.

Alternative attachment figures may matter

A 2024 empirical study examined adults who were retrospectively classified into earned secure, continuously secure, insecure and “lost secure” groups.

The earned secure and continuously secure groups reported better aspects of emotion regulation and mentalization than insecure groups.

The study also found that adults other than parents who had been important during childhood were common across groups, but attachment to those alternative figures was more secure among earned secure and secure participants.

This fits a plausible developmental pathway.

Security may be supported not only by parents but by:

• grandparents;
• relatives;
• teachers;
• mentors;
• later close relationships;
• therapeutic relationships.

But the study was cross sectional, used a female adult sample and employed a retrospective operationalization.

So it adds a possible mechanism.

It does not prove a universal pathway.

Reflective functioning may be part of the process

One reason earned security is interesting is that security in the AAI tradition is not defined by having a perfect childhood.

It is strongly related to how experiences are represented.

A person can describe pain, rejection or inconsistency and still do so coherently.

They can recognize complexity.

They can avoid idealizing caregivers unrealistically.

They can acknowledge harm without becoming completely disorganized by the narrative.

This capacity is often discussed alongside mentalization or reflective functioning.

The ability to think about behavior in terms of mental states may help a person integrate adverse experience instead of simply repeating it.

Again, this is a developmental hypothesis with empirical support, not a guaranteed mechanism.

Therapy may contribute to security, but “therapy equals earned secure” is too strong

Psychotherapy can provide repeated experiences of:

• responsiveness;
• emotional exploration;
• rupture and repair;
• reflection;
• autonomy;
• reliable boundaries.

A meta analysis found that increases in attachment security during psychotherapy were associated with better outcomes.

A 2024 study specifically examined the development of secure therapeutic attachment and growing comfort with closeness and autonomy.

This is consistent with the possibility of attachment relevant change.

But therapy studies do not justify saying:

“If you complete enough therapy, you become earned secure.”

Different studies measure attachment differently.

Treatment effects vary.

And earned security is not a formal clinical certification.

A secure current relationship is not the entire concept either

A loving partner may provide corrective experience.

Repeated responsiveness can challenge expectations of rejection or unavailability.

That matters.

But romantic security and AAI earned security are not interchangeable.

Someone can feel secure with one partner while still having unresolved issues in other attachment domains.

Someone can show coherent AAI security while experiencing anxiety in a specific romantic relationship.

Adult attachment is multi layered.

The measurement literature repeatedly shows only partial convergence across methods.

That complexity should be preserved.

Earned security is not proof of trauma

Social media often starts with:

“An earned secure person had a traumatic childhood.”

The research concept is broader and messier.

Negative childhood care has been operationalized in different ways.

Retrospective reports can be influenced by current perspective.

Some prospectively secure individuals may later describe childhood negatively.

Some people exposed to adversity never meet criteria for insecure infant attachment.

Adversity and attachment are related concepts.

They are not synonyms.

You should not infer maltreatment, neglect or trauma from an attachment label.

Earned security is not perfection

Even in early earned security research, retrospectively earned secure adults sometimes showed vulnerabilities such as more depressive symptoms than continuous secure adults.

Other work has found both similarities and differences.

A secure state of mind does not mean:

• no anxiety;
• no conflict;
• no triggers;
• no grief;
• no relational mistakes;
• no need for support.

Security is about organization and flexibility, not emotional invulnerability.

This connects directly with the existing DarkBrain principle:

Secure attachment is not perfect calm.

The DarkBrain model: history, representation, relationship, flexibility

A cautious way to think about earned security is through four layers.

History

What attachment relevant experiences actually occurred?

Prefer prospective evidence where available.

Representation

How are those experiences understood and narrated now?

Can the person hold complexity coherently?

Relationship

What later attachment figures provided reliable support or corrective experience?

Flexibility

Can closeness and autonomy both be used without rigid hyperactivation or deactivation?

This model does not assign an earned secure label.

It identifies the developmental questions behind the concept.

Why the concept is still valuable

If earned security is scientifically messy, why keep it?

Because the underlying idea is important.

Human development is not a one way sentence.

Early attachment matters.

It influences expectation, emotion regulation and relationships.

But later experience matters too.

People can develop new capacities.

They can form relationships that function differently from earlier ones.

They can become more reflective.

They can change how attachment relevant threat is regulated.

The research challenge is describing that process without pretending we know one universal recipe.

What the evidence does not support

The evidence does not support these universal claims:

• earned security means you had childhood trauma and healed it;
• anyone can self diagnose earned security from an online quiz;
• lower romantic attachment anxiety automatically equals earned security;
• therapy guarantees earned security;
• a secure romantic partner creates earned security;
• earned secure adults have no attachment triggers;
• the concept has one standardized definition across research;
• retrospective reports perfectly reconstruct infant attachment.

The current literature is too heterogeneous for that certainty.

The Collection in one model

The four Articles in Attachment Patterns in Adult Relationships create a progression.

Anxious Attachment Without Stereotypes explains hyperactivation without character judgment.

Avoidant Attachment explains deactivation without equating distance with absence of feeling.

Can Attachment Patterns Change in Adulthood? establishes that adult attachment is stable but not fixed.

Earned Security examines one specific research tradition for understanding developmental movement toward security.

The Collection thesis is:

Attachment patterns are strategies with history, context and plasticity. They are not personality verdicts.