Crystal Healing: What Controlled Studies Can and Cannot Show

Two people receive a small pink stone.

Both are told how to use it.

Both spend two weeks with it.

One has rose quartz.

The other has a visually similar imitation.

Afterward, one person says they feel calmer.

So does the other.

What just happened?

This is the sort of question that separates a compelling personal story from a test of treatment.

And in 2025, researchers put that distinction directly into a crystal-healing experiment.

First, ask what 'works' means

A crystal can 'work' in several completely different senses.

It can work as jewelry.

It can work as a reminder to pause.

It can make someone feel hopeful.

It can be culturally meaningful.

And it could, in theory, have a direct measurable physiological effect through a specific mechanism.

Only the last proposition would validate claims about an intrinsic clinical effect of the crystal material.

If the promised outcome is anxiety relief, the test has to distinguish whether the relief is caused by the particular stone rather than by the ritual and expectation shared by both conditions.

The study

The paper was published in CNS Spectrums in 2025 under the title Placebo effects in alternative medical treatments for anxiety: false hope or healing potential? [S07,S08]

The investigators recruited 138 adults from the general population.

Seventy were categorized as people who believed in crystal healing.

Sixty-eight were categorized as nonbelievers.

Participants were randomly assigned to receive either:

• real rose quartz, or
• a visually similar placebo crystal.

They followed a standardized 14-day protocol and completed anxiety measures before and after, including the Beck Anxiety Inventory and a version of the Kuwait University Anxiety Scale. [S07]

The key comparison was not whether anyone improved.

It was whether the real crystal did better than the imitation.

What did the researchers find?

In the reported results, reductions in anxiety were observed among participants who already believed in crystals.

Those reductions occurred regardless of whether the person had real rose quartz or the placebo.

Nonbelievers did not show reliable improvement.

There was no detected specific advantage of real rose quartz over the visually matched control on the primary outcomes. The paper also reports Bayesian results favoring the null hypothesis for mineral-specific treatment effects. [S07,S08]

This is the most important result.

If the mineral itself were making a clinically meaningful difference under those conditions, the real-crystal group should have separated from the imitation group in a consistent way.

It did not.

Does that mean the participants felt nothing?

No.

Some participants reported less anxiety.

The reduction can matter to the person experiencing it.

What the comparison challenges is the explanation 'rose quartz has a unique anti-anxiety power'.

In a controlled study, an improvement shared by the active and placebo groups is not evidence that the active ingredient added something specific.

The experience can be real.

The claimed cause can still be unsupported.

That is not a contradiction.

Why expectation matters

A person who believes a stone will calm them may notice sensations that fit that expectation.

They may enter the ritual in a hopeful state.

They may spend a few quiet moments focused on something concrete rather than a worry.

The act of participating can change self-report, attention and perceived symptoms.

NIH/NCCIH explains that placebo effects can arise from expectations and the therapeutic context, and that randomized placebo comparisons help distinguish specific benefits from other improvements. [S09]

These processes can have psychological and physiological correlates.

Calling something 'placebo-related' is not equivalent to calling the person a liar.

But did the study prove belief caused the benefit?

This is where a tempting conclusion needs restraint.

The researchers did not randomly assign people to believe or disbelieve in crystals.

They classified people by pre-existing belief and then randomized crystal type.

The study therefore shows a relationship between belief status and the observed outcome in that design.

It does not by itself establish that belief caused the improvement.

The authors discuss expectation and conditioning as plausible explanations, but their limitation section explicitly says that the proposed conditioning process was not experimentally manipulated. [S08]

It would be misleading to turn the paper into:

'Scientists proved conditioning causes crystal healing.'

That is a stronger statement than the experiment supports.

Why the placebo design is useful

Suppose a hundred people receive rose quartz and most say they like it.

That tells you they liked it.

It does not tell you whether:
- the mineral mattered;
- time itself changed symptoms;
- the researcher's instructions shaped expectations;
- people reported improvement because they wanted to please the researcher;
- attention or routine was helpful.

A visually matched control removes part of this ambiguity.

Both groups hold an apparently similar object.

Both groups follow the same time frame.

The only intended distinction is the material that the healing claim says should matter.

That is the experiment's strength.

The study had limits too

Strong science includes the parts that make a headline less easy.

Nonclinical participants

The sample came from people who reported no diagnosed mental disorder. The authors note that effects could look different in clinical populations with elevated symptoms. [S08]

The trial does not establish how crystals perform as an intervention for clinically diagnosed generalized anxiety disorder, panic disorder or PTSD.

Fourteen days

The observation period was two weeks.

It does not answer questions about years of use or every possible ritual setting.

Beliefs were observational

Participants were not randomly made believers or nonbelievers.

Blinding could be imperfect

Participants might potentially recognize whether an object was genuine, and masking integrity was not guaranteed. The authors describe this limitation themselves. [S08]

Outcomes were anxiety measures

The trial did not test whether crystals altered cancer outcomes, infection rates, bone healing or objectively measured 'chakra energy'.

A good study should not be asked to support claims it did not investigate.

Can one trial settle every possible crystal claim?

No.

It cannot prove that no crystal will ever be found to have any specific application under any conditions.

Different crystals can have different material properties; medicine and engineering use minerals in many unrelated ways.

But the study does directly challenge a familiar claim: that rose quartz possesses an anxiety-reducing effect above a visually similar placebo during the tested ritual.

If someone continues to claim that it does, they need comparably strong positive evidence, not merely testimonials.

What would a stronger future study look like?

Researchers might pre-register outcomes, recruit clinically characterized participants, measure beliefs carefully, verify masking and use an adequate sample size for realistic effects.

They could compare:

real stone;

identical-looking imitation;

standardized activity with no stone;

and where appropriate, evidence-based treatment conditions.

They might include objective physiological measures alongside self-reported symptoms.

A study should also separate the stone itself from the ritual instructions.

Otherwise, even a positive result could leave the active ingredient unclear.

Why testimonials are not enough

Personal stories are valuable for describing experience.

They are poor tools for identifying hidden causal mechanisms.

A person can report:

'My anxiety eased after I started carrying amethyst.'

They may be entirely sincere.

But the event could reflect:
- natural symptom fluctuation;
- hope;
- an accompanying behavioral change;
- attention shifting;
- a supportive social community;
- or the stone itself.

Only a fair comparison can begin separating those options.

The word 'placebo' should not be used as an insult

A useful analogy is ritual before a difficult conversation.

A person places both feet on the floor, breathes slowly and touches a familiar necklace.

They feel more stable.

It would be absurd to insist the necklace must emit a therapeutic radiation field for the routine to have been helpful.

It would also be unkind to say the experience was fake.

Meaning can change action.

Action can change experience.

That remains different from proving a mineral-specific health effect.

The real risk is substitution

Most ordinary crystals are not inherently dangerous to look at or carry, though some forms, fragments, dusts and preparations can introduce physical risks.

The biggest danger in healing claims is often what someone stops doing.

Someone with ongoing panic symptoms may postpone evaluation.

Someone with a serious illness may spend money on 'energy purification' instead of evidence-based treatment.

NCCIH explicitly advises against replacing or delaying indicated medical care with unproven complementary practices. [S10,S16]

The ethical boundary is simple.

A symbolic object can accompany care.

It should not be sold as a substitute for care.

DarkBrain's conclusion

The question is not:

Can holding a crystal feel good?

It can.

The question is:

Has the crystal itself been shown to produce the promised medical benefit beyond a suitable control?

For rose quartz and anxiety in the cited 2025 experiment, the answer was no.

That is a meaningful answer.

And it is precise enough not to become another myth.