Does Self-Hypnosis Work? The Evidence, and How to Actually Do It

“Self-hypnosis” sounds like something you do with a pocket watch and a candle.

The research tells a duller, more useful story.

Yes, self-hypnosis works — for specific things. A 2019 meta-analysis of randomized controlled trials (Eason & Parris) found a large overall effect size (Hedges’ g = .536) for self-hypnosis as a clinical treatment, strongest for pain, stress, and anxiety. The catch: it behaves like a skill, not a magic trick. It works best when you treat it as a self-directed technique and practice it a handful of times — not when you passively listen to a recording once and wait for something to happen.

Does self-hypnosis actually work?

The strongest direct evidence comes from a systematic review and meta-analysis by Eason and Parris, published in Psychology of Consciousness in 2019. They screened 576 studies and found 22 randomized controlled trials that met criteria, computing effect sizes from 17 of them.

Using a random-effects model, the mean effect size was Hedges’ g = .536 (95% CI .398 to .650), which is a large effect by Cohen’s conventions. The authors noted real variation between studies, and after correcting for possible publication bias the true effect is likely in the medium-to-large range — still a meaningful, replicated result, not a rounding error.

So the honest version isn’t “self-hypnosis cures everything.” It’s: for the outcomes it’s been tested on, the effect is real and not small.

This sits inside a broader evidence base for hypnosis generally. A 2024 umbrella review by Rosendahl and colleagues in Frontiers in Psychology synthesized 49 meta-analyses covering 261 distinct primary studies and found that 63.6% of the extracted effects were statistically significant, with the most robust evidence in pain and medical-procedure settings. Self-hypnosis is one delivery mechanism for those same effects — the version you can run yourself.

What does self-hypnosis help with?

The conditions where self-hypnosis has held up in randomized trials are fairly specific. In the Eason and Parris review, self-hypnosis was reported effective for pain, childbirth, pediatric applications, stress, and anxiety.

A pattern worth noting: the trials reporting no effect tended to be the ones where participants only listened to a recording of someone else’s hypnosis. The effects showed up when self-hypnosis was taught as an active, self-directed skill. That distinction matters more than which specific script you use.

Sleep is a more cautious “promising.” In a 2023 feasibility trial (Snyder et al.) of a self-administered hypnosis intervention for college students, 91% of participants completed the program and there were zero study-related adverse events. Self-reported measures moved — sleep onset latency dropped from about 28.8 minutes to 14.8 minutes over three weeks — but objective nightly sleep duration did not change significantly. Read that as: feasible, safe, and worth more research, not yet proven on hard sleep metrics.

The thread across all of it is that self-hypnosis is a way to deliberately shift your attention and your body’s stress response — which is exactly why it shows up for pain and anxiety, and why it’s low-risk to try. If you’re curious how that shift looks in the brain, see what happens in your brain during hypnosis and how neural repatterning rewires your brain.

Will it work for me?

Partly. Response to hypnosis varies between people, and that variation is measurable. Some people are highly responsive, most are moderate, and a minority respond weakly — a stable trait, not a reflection of intelligence or willpower. If you want the full picture, we’ve covered whether you’re hypnotizable and the science of who responds.

The practical takeaway from the trial data is more encouraging than the trait research alone: even moderate responders benefited, provided they practiced. The single most reliable predictor of whether self-hypnosis worked in these studies wasn’t natural talent — it was repetition.

How do you do self-hypnosis?

The Eason and Parris analysis flagged one concrete condition for effectiveness: self-hypnosis worked best when it was treated as an independent, self-directed skill and when participants did at least three practice sessions before it was tested. That’s the whole secret — it’s trained, not switched on.

A basic, evidence-aligned sequence looks like this:

Settle. Sit or lie somewhere quiet where you won’t be interrupted. Let your eyes close. Take a few slow breaths and let your body get heavy.
Narrow your focus. Pick one anchor — your breath, a point of warmth in your hands, a single calm image — and let everything else recede. This focused absorption is the core mechanism, not the relaxation per se.
Give yourself a specific suggestion. Phrase it concretely and in the present: “My jaw is loosening,” “This sensation is fading,” “I am calm and steady.” Repeat it slowly. Specific beats vague.
Stay a few minutes, then return. Sit with the state, then count yourself back up to full alertness.
Repeat. Do it a few times across several days before you judge whether it works. The trials that found effects gave people that runway; the ones that didn’t, often didn’t.

You don’t need an induction from someone else first — the research found prior experience of therapist-led hypnosis wasn’t necessary for self-hypnosis to produce an effect.

The takeaway

Self-hypnosis isn’t fringe and it isn’t a cure-all. The evidence supports a narrow, useful claim: practiced as a self-directed skill, it produces a medium-to-large effect on pain, stress, and anxiety, with a strong safety record. The failure mode is treating it as something passive that happens to you. Treat it as a skill, give it a few honest reps, and judge it on outcomes.

For the wider research on how and why this works, see the pillar overview on the science of clinical hypnosis.