DARVO: What the Pattern Is and What It Is Not

How denial, counterattack and role reversal can redirect accountability, what experiments actually show, and why the label should not be used as a shortcut for deciding who is guilty.

Accountability can disappear without the original event being answered

A person raises a complaint.

The response is immediate:

That never happened.

Then:

You are unstable for accusing me.

Then:

Actually, you are the abusive one. Look what you are doing to me right now.

The conversation began with one allegation.

Minutes later, the person who raised it is defending their character.

That sequence has a name:

DARVO.

The acronym stands for Deny, Attack, and Reverse Victim and Offender.

The concept is useful because it describes a recognizable accountability response.

It is also easy to misuse.

DARVO is not a mind-reading device.

It does not prove that an accused person is guilty.

And ordinary disagreement is not automatically DARVO.

The value of the concept comes from identifying a communication pattern while keeping the underlying evidence separate.

What DARVO actually describes

The term was introduced by psychologist Jennifer Freyd in the context of responses to confrontation about wrongdoing.

The three components are straightforward.

Deny

The person rejects, minimizes or reframes the alleged wrongdoing.

Examples:

“That did not happen.”

“You are exaggerating.”

“It was not a big deal.”

“You misunderstood.”

Denial alone is not enough to establish DARVO.

An innocent person may deny something because they believe the allegation is false.

Attack

Attention shifts toward the credibility, motives or character of the person raising the concern.

Examples:

“You are crazy.”

“You always make drama.”

“You are just trying to destroy me.”

“No one believes you.”

Reverse Victim and Offender

The accused person positions themselves as the real harmed party and the accuser as the aggressor.

Examples:

“You are abusing me with these accusations.”

“You are the one who ruined everything.”

“I am the victim here.”

The full pattern redirects the conversation from:

What happened?

to:

Which person deserves sympathy, credibility and blame?

The first empirical study found a link with self blame

A 2017 study asked 138 undergraduate participants to report on a situation in which they confronted someone about wrongdoing.

The researchers developed a DARVO measure and found that denial, attack and reversal tended to cluster together.

Greater reported exposure to DARVO was associated with greater self blame among the people who had confronted the wrongdoer.

This study was important because it moved DARVO from a clinical and conceptual observation toward empirical testing.

But it had limitations.

It relied on retrospective self report.

The sample was small and consisted of university students.

It could not independently determine what actually occurred in each original conflict.

So the study supports the relevance of the pattern.

It does not turn DARVO into a forensic test.

Experiments show that DARVO can change how observers judge a conflict

The strongest evidence comes from controlled vignette experiments.

In 2020, Sarah Harsey and Jennifer Freyd published two experiments.

In the first, 316 university students read a fictional scenario involving interpersonal violence.

When the alleged perpetrator used DARVO, participants rated the victim as:

• less believable;
• more responsible;
• more abusive.

They also rated the perpetrator as:

• less abusive;
• less responsible.

The second experiment tested whether education about DARVO changed these judgments.

Participants who had learned about the pattern were more likely to rate the victim as believable and the perpetrator as less believable.

This suggests something important:

DARVO may work partly by reorganizing social perception.

The original allegation becomes harder to evaluate because attention has been redirected toward the accuser.

Later experiments found similar effects

A 2023 experiment tested DARVO in a fictional sexual assault scenario.

Participants exposed to perpetrator DARVO rated the perpetrator as less abusive, less responsible and more believable than participants who saw a response without DARVO.

They also rated the victim as more abusive and less believable.

Participants exposed to DARVO were less willing to punish the perpetrator and more willing to punish the victim.

Again, the study used a fictional scenario and an undergraduate sample.

That matters.

But across experiments, the same broad effect appears:

role reversal can influence third party judgments.

Why the pattern can be persuasive

DARVO does several things at once.

It creates competing allegations

Instead of one accusation, observers now have two.

The cognitive task becomes more difficult.

It changes the emotional frame

The accused person may appear injured, outraged or threatened.

Strong emotion can be mistaken for evidence.

It attacks credibility

If the accuser can be framed as unstable, dishonest, vindictive or abusive, the original allegation may receive less attention.

It exploits fairness instincts

Observers often want to avoid falsely blaming anyone.

A convincing role reversal can turn that caution into misplaced symmetry.

It shifts the burden

The accuser begins proving that they are not cruel, unstable or manipulative instead of discussing the original behavior.

None of these mechanisms requires a secret master plan.

A person may use parts of DARVO strategically, defensively or habitually.

The pattern can still redirect accountability.

DARVO is not the same as gaslighting

The terms overlap in online conversation, but they describe different processes.

Gaslighting refers to a pattern that undermines a person’s confidence in their perception, memory or judgment.

DARVO refers to a response to accusation or confrontation that involves denial, attack and role reversal.

A person could use both.

They could also use one without the other.

For example:

“You imagined the whole thing because your memory is unreliable” may function as gaslighting.

“You are the real abuser for accusing me” fits role reversal more directly.

Keeping the concepts separate prevents every harmful conversation from becoming one giant diagnostic label.

DARVO is not proof of guilt

This boundary is essential.

Suppose two people accuse each other of serious harm.

Both deny wrongdoing.

Both attack the other’s credibility.

Both claim victimization.

The pattern alone cannot tell an outside observer which account is true.

People who have actually been harmed can become angry.

People who are falsely accused can strongly deny allegations.

Victims can also counterattack.

Therefore, the presence of DARVO-like language is evidence about communication behavior, not proof about the underlying event.

A careful investigator must still ask:

• What happened?
• What records exist?
• What independent evidence exists?
• Are there witnesses?
• Is there a documented pattern?
• Are accounts internally consistent?
• What alternative explanations fit the evidence?

DARVO should sharpen investigation.

It should not replace it.

One conflict is not always a pattern

Online discussions often identify DARVO from a single screenshot.

That can be too confident.

The research concept was developed around responses to wrongdoing and confrontation, often in contexts of interpersonal violence.

Context matters.

A defensive reply in one argument may resemble one component without demonstrating a stable pattern.

The stronger case appears when:

1. a specific concern is raised;
2. the concern is denied or minimized;
3. the accuser’s character becomes the target;
4. roles are reversed;
5. the original concern remains unresolved.

Repeated structure is more informative than one phrase.

The label can itself be weaponized

Any useful psychological concept can become part of a conflict.

A person can say:

“You are DARVOing me.”

The other person replies:

“No, you are DARVOing me by saying I am DARVOing you.”

At that point, the label is no longer increasing clarity.

It has become another accusation.

DarkBrain’s rule is simple:

Return to observable behavior.

Instead of:

“You are using DARVO.”

Say:

“I raised one specific issue. You denied it, then shifted to my character, and now the conversation is about whether I am the bad person. The original issue still has not been addressed.”

Descriptions are harder to hide behind than labels.

Education appears to reduce some of the effect

The 2020 experimental work found that participants taught about DARVO judged the scenario differently from participants who had not received that education.

This does not mean a five minute explanation makes someone immune to manipulation.

It does suggest that naming a social pattern can improve defensive awareness.

The useful education is not:

“Anyone who does this is guilty.”

It is:

“When a confrontation suddenly becomes a credibility war and role reversal, pause before abandoning the original question.”

That is a much safer skill.

What the evidence does not support

The evidence does not support these claims:

• DARVO proves guilt;
• every denial is DARVO;
• every counteraccusation is false;
• only one gender uses DARVO;
• every difficult partner who becomes defensive is abusive;
• DARVO is a clinical diagnosis;
• identifying DARVO tells you what happened in the original event;
• one screenshot is enough to reconstruct an entire relationship.

The research base is growing but still smaller than the online popularity of the term suggests.

What remains uncertain

DARVO research has several limitations.

Many foundational studies use university samples.

Some use hypothetical scenarios.

Self report studies cannot independently verify original events.

Research has focused heavily on sexual violence and interpersonal abuse, which means generalization to workplaces, families, politics or ordinary disputes should be cautious.

The newest research is expanding the construct, but the scientific literature remains much smaller than the volume of public commentary.

That is precisely why boundaries matter.

The DarkBrain model: allegation, response, redirection, evidence

When a conversation resembles DARVO, separate four layers.

Allegation

What specific behavior was originally raised?

Response

Was it answered directly, denied, explained or minimized?

Redirection

Did the conversation shift toward the accuser’s character or toward a competing claim of victimhood?

Evidence

What independent information exists beyond either person’s emotional presentation?

This model protects against two opposite mistakes.

Mistake one:

Ignoring role reversal because the accused person appears emotionally convincing.

Mistake two:

Treating role reversal as automatic proof of guilt.

Both abandon evidence.

Why this matters

DARVO is most useful as a defensive awareness concept.

It teaches a person to notice when accountability has been replaced by a struggle over who gets to occupy the victim position.

The goal is not to win the label.

The goal is to keep the original question visible long enough to examine it properly.

Continue exploring

Next: Coercive Control: Why Patterns Matter More Than Single Incidents

The next article moves from confrontation tactics to a broader pattern of domination, examining how autonomy can be restricted through repeated surveillance, isolation, threats, economic control and other behaviors that may appear minor when viewed separately.