Do Affirmations Work? Why They Can Make You Feel Worse

I'm in the middle of planning a three-part series on gaslighting. Somewhere along the way, a strange thought hit me: if I repeated certain words to myself in order to grow, could I actually change?

Watching news stories about people hurt by gaslighting, I kept wondering how mere words can make someone doubt even their own memory and judgment. And if words carry that much power, couldn't I put it to work for me?

I've been feeding myself lines like "I have no staying power" and "It's too late to start now." If I swapped in a different story, would my choices and actions change? Or would it buy me nothing more than a moment of comfort?

So the idea of gaslighting myself came down to one question: could I change what I believe about who I am, and actually get moving, without fooling myself?

A middle-aged man looking thoughtfully at his reflection in a bathroom mirror, questioning his self-image and inner beliefs.

Gaslighting, as the American Psychological Association defines it, means manipulating someone into doubting their own perceptions, experiences, and understanding of events. "Self-gaslighting" is only a metaphor here; the research below asks a plainer question: do affirmations and other self-talk actually work? If you've ever repeated a line to yourself after a bad week, you've already run a rough version of these experiments, and the first result you'll meet here points the wrong way.

Why would one affirmation, said sixteen times, leave some people feeling worse?

A man sitting alone at a desk at night, holding a blank sticky note while struggling with unconvincing positive affirmations.

Imagine four minutes with a pen. You're writing down whatever you think and feel. Every fifteen seconds you get a cue, and at each one you repeat the same sentence: "I am a lovable person." By the end, you've said it sixteen times.

That's the setup of a 2009 experiment by Wood, Perunovic, and Lee. The sentence is an affirmation, a positive statement about yourself meant to lift you. In a survey reported in the same paper, 85 percent of the college students who answered said they used statements like this before exams, and on the whole they rated them as helpful.

So you'd expect sixteen repetitions to help, or at worst to do nothing.

A comparison group did the same four minutes of writing, minus the sentence. People with high self-esteem who repeated it felt only slightly better than those who didn't. People with low self-esteem felt worse: they left in a lower mood than the ones who only wrote, and less eager to do anything enjoyable.

The line hurt exactly the people who seem to need it most.

You may know the feeling: you tell yourself something flattering, and a small voice answers, "sure you are."

One explanation the authors offered sounds airtight. A sentence that clashes sharply with your existing self-image may get rejected, or may summon the counterexamples. Your mind isn't a whiteboard, in other words: a new line gets held up against the old ones, and for someone with low self-esteem, the old ones may win.

If that's right, the fix is obvious: keep the counterexamples out. In a follow-up, the same team tried exactly that, and its opposite, side by side.

What happens when you let the doubts talk back?

An overhead view of a notebook, blank index cards, and a pen arranged on a desk, representing the process of examining thoughts and separating facts from assumptions.

Same sentence, new instructions. One group was told to focus only on the ways and times "I am a lovable person" was true. The other was told to think about the ways it might be true and the ways it might not.

If counterexamples were the poison, that second group was being handed a dose on purpose. They should have ended up worse.

The opposite happened. Among people lower in self-esteem, the both-sides group came out ahead on mood and on how happy they felt about themselves. A quick self-rating showed only a borderline gap, and the urge to do something enjoyable didn't differ.

Maybe the doubts were never the poison. Maybe the poison was the rule against them.

The authors had floated that very suspect. A positive self-statement, they suggested, may quietly set a standard: think only good thoughts about yourself. If so, every stray doubt counts as a small failure.

The cure for a ban is a seat at the table. Britain's National Health Service offers a worksheet along those lines, the thought record: you write down the harsh thought, then the evidence for it and against it, then a more realistic version.

It's homework, frankly. But it never asks you to erase the harsh thought. It puts the thought on trial.

The NHS's own example is a forgotten errand and the thought "I'm useless." The evidence against it: everything else got remembered. The rewrite doesn't praise anyone. It just counts: far more errands remembered than forgotten.

The failure stays on the page; only the verdict shrinks. Yet every result so far has measured how people felt. None of it tracked what anyone did next.

Does going easy on yourself after a failure make you slack off?

A man studying at his desk after a setback, reviewing his notes and preparing to try again through self-compassion and deliberate effort.

Now take a seat in a college student's chair. You've just struggled through a hard test of antonyms and been handed the correct answers. A second test is coming, and you can study for it as long as you like.

In this 2012 experiment by Breines and Chen, a short message arrived before the studying, assigned at random. One said trouble with a test like this is common, and ended, "try not to be too hard on yourself." Another said you must be intelligent if you got into Berkeley. A third group got nothing extra.

On paper, the first message is the risky one. Going easy looks a lot like permission to coast.

The go-easy students didn't coast. They studied the longest, just over five minutes, while the other two groups each stopped short of four minutes. Only the gap over no message was statistically clear.

Notice what the leading message left out: praise. It called the stumble ordinary and took the sting out. My reading: words that don't deny a failure leave you free to go back to it.

Then the second test. The go-easy group averaged 7.4 correct out of ten; the others, 6.9 and 7.0. The statistics couldn't tell them apart.

The gentle words bought minutes. The minutes didn't buy points, at least not on this test. Somewhere between trying harder and doing better, a piece is missing.

What if the self-talk is a plan instead of a pep talk?

One candidate for that missing piece is unglamorous: a plan. Say you want to start running, and what usually stops you is being wiped out after dinner. The plan becomes: if I'm too tired to run after dinner, I'll put on my shoes and walk around the block.

A 2021 meta-analysis by Wang, Wang, and Gai looked at a fuller version: picture the outcome you want, name what stands in the way, then make the if-then plan. Across 24 results from 21 studies, it found a small-to-moderate boost in reaching goals, possibly smaller than it looks, the authors note.

A pooled number can't show what that looks like up close. One recent study can.

In a preliminary 2025 study from the University of Leeds, by Divine and Astill, some people recruited online wrote if-then plans tying an everyday event to a workout. The three who got the format wrong were emailed and asked to fix it. All three did.

Nobody chanted anything. Someone checked a plan like a form.

Some people only wrote plans. Some only practiced picturing their workouts, about three minutes at a time. Some did both: their own plan was written into a script that walked them through getting ready, moving, and how it felt. A control group did neither.

If a plan were the missing piece, the plans-only group should have shown it.

It didn't. Its habit strength, meaning how automatic it felt to start exercising, didn't move at any point. Both picturing groups reported stronger habits, and the combined group got there first.

The plan told people when to move. Picturing the movement seems to be what made it feel automatic.

The workouts people reported were murkier. The combined group's activity rose week by week during the program, but across the study's three checkpoints, the groups didn't differ reliably.

None of this proves you can talk your way into a different self. I'd still say the small, checkable version is worth trying: a better first step, not a new identity.

A behavioral researcher examining scientific papers and statistical charts, evaluating the difference between personal effort and measurable results.

What were those first four minutes really testing?

Back to the pen and the cue. Remember the definition at the top: gaslighting works by getting someone to doubt their own perceptions.

Now set two instructions from this story side by side: repeat an affirmation on cue, and look only at what agrees with it. In effect, both asked people to talk over their own read of themselves.

That's the closest this research got to gaslighting yourself, and for people with low self-esteem, it lost both times. No study here tested this directly, so call it my bet: the line between persuading yourself and fooling yourself is whether the sentence is allowed to be cross-examined.

While digging through these sources, I started rewording my self-talk. It's too soon to tell how far that has taken me, or whether it will lead to growth. Still, today feels different from yesterday. That's not evidence of change, but I don't think I need to deny the feeling either.

"I can't be certain I've gotten any better, but I'm grateful to myself for still trying to change today."

At first I called this self-gaslighting. But what I want isn't to hide my shortcomings or insist that a failure was a success. It's to consider that the things I've been telling myself might not be all of me, and to try acting differently.

I hope this post gives you a little courage not to define your future self by who you are today.

I'll keep experimenting. But I mean to keep my reason intact: to tell the words I'd like to believe from what's real, and to stop and correct course when the two drift apart.

A middle-aged man walking alone along a quiet woodland trail at sunrise, symbolizing a realistic commitment to personal growth without guaranteed success.

Sources & References

This article is for educational and informational purposes. It is not professional advice. Readers should consult qualified professionals for decisions that affect their personal circumstances.

Search This Blog

History Meets Science