IX Coach

Identity-Based Habits: What the Idea Is, What Backs It Up, and Where It Goes Wrong

What are identity-based habits?

Identity-based habits is the idea, popularised by James Clear in Atomic Habits (2018), that lasting habit change works best when each small action is treated as evidence of the kind of person you are becoming, rather than as a step toward an outcome. This page explains the idea in plain words, how it relates to the two-minute rule and habit stacking people search alongside it, what research it leans on and how strong that research actually is (including studies that did not replicate), where the idea can backfire, and how a person might practise it. Written to be useful whether or not you ever open IX Coach.

Identity-based habits is the idea that the most durable way to change what you do is to change what you take yourself to be, and to let small repeated actions serve as the evidence. Instead of aiming at a result such as running a marathon, you aim at being a person who runs, and treat each run, however short, as a piece of proof. James Clear popularised this framing in his 2018 book Atomic Habits.

This page tries to keep the idea at its real size. It is a useful way of thinking, and it has a research lineage, but the lineage is thinner than the popular retellings suggest, and one of its best-known studies has since failed to replicate. It also covers the two things people tend to search for in the same breath, the two-minute rule and habit stacking, because they are easy to confuse with the identity idea and they rest on different evidence.

Nothing here is a treatment for any condition. If a habit you are trying to change is tied up with something like an eating disorder, addiction or depression, a licensed clinician is the right place to start rather than a self-image exercise.

The three layers, in plain words

Clear describes behaviour change as having three layers. The outermost is outcomes: the results you want, such as losing weight or finishing a draft. The middle layer is processes, the habits and systems that produce those results, such as a training routine or a writing hour. The innermost layer is identity: your beliefs about yourself and the world. His argument is that most people begin at the outside and work in, while identity-based change begins at the inside and works out.

His account of how identity changes has two steps. First you decide what kind of person you want to be. Then you build evidence for that picture through small wins. He sums it up in a line that has become the book's shorthand, that every action is "a vote for the type of person you wish to become", and adds that no single vote settles the matter, but that votes accumulate into evidence. That is the whole mechanism as he presents it: the self-image is a running tally, and you are the one filling in the tally.

A fair way to read this is as a claim about where motivation comes from. If skipping a run is a failure to hit a number, it can be shrugged off. If it is a vote against the person you are trying to be, it is meant to matter more, and so is a small action taken on a bad day. Whether it works that way for a given person is an empirical question, and the next sections look at what is known.

Explore: atomic habits · self concept clarity

Where the research lineage comes from

The study most often placed behind this idea was published in 2011 by Christopher Bryan, Gregory Walton, Todd Rogers and Carol Dweck in the Proceedings of the National Academy of Sciences. In three experiments, people answered a short survey in which voting was described either as being a voter (a noun, an identity) or as voting (a verb, an action). In the first, 34 unregistered Californians who got the noun wording reported more interest in registering. In the second, 88 registered Californians were checked against official records after the 2008 election, and turnout was 95.5 per cent in the noun group against 81.8 per cent in the verb group. In the third, 214 people from a national panel were checked after the 2009 New Jersey election, with 89.9 per cent against 79.0 per cent.

Those are striking numbers from a tiny change in wording, and the paper is often offered as support for the claim that framing an action as part of who you are makes people more likely to do it. The authors also named limits. Turnout in both groups was higher than expected, possibly because people who answer a survey are likelier to vote, and there was no group that received no survey. They suggested the effect may depend on voting being something people already feel they should do, so it may not carry over to behaviours people see as undesirable.

The result has not held up well when tested again. A 2017 paper by Al Fang, Alan Gerber and Greg Huber in Political Psychology reported a new large-scale field experiment and found no evidence that the wording change increased turnout. A 2020 paper by Gerber, Huber and Fang in Behavioural Public Policy, a preregistered experiment during the 2016 US presidential election chosen because critics had argued earlier tests lacked a high-stakes setting, likewise found no evidence that noun wording increased turnout compared with verb wording, including in competitive states.

What does and does not follow from this deserves care. These later studies tested one specific thing, a wording change in a voting survey. They do not show that thinking of yourself as a runner or a writer is useless, and the original paper does not show that it works for habits. What they do show is that the single most-cited experiment behind the claim cannot be relied on as strong support for it. So identity-based habits is better held as a plausible practitioner framework than as an established finding.

Explore: identity priming · narrative identity

The two-minute rule and habit stacking are different things

People searching for the two-minute rule from Atomic Habits are usually after one of two ideas, and Clear's version is the first. He suggests scaling a new habit down until the first step takes two minutes or less, such as reading one page instead of an hour, and treating that small start as a gateway to the larger habit. A separate productivity rule, associated with David Allen, says that if a task would take under two minutes you should do it now instead of postponing it. Clear's page footnotes the second one, which is the source of a lot of mix-ups.

In relation to identity, the small start is the point: the action is meant to be easy enough to do on any day, so that you keep casting votes. We did not find a controlled study that tests two minutes as a threshold. It is best read as a practical rule of thumb for lowering the cost of starting, not as a measured number.

Habit stacking is attaching a new behaviour to something you already do, so the existing habit acts as the reminder. Clear writes the pattern as after or before a current habit, I will do the new habit, and credits BJ Fogg, who calls the approach anchoring, for the underlying idea, and SJ Scott for inspiring the name. He presents it as a special case of implementation intentions, the if-then plans studied by Peter Gollwitzer.

That research is of a different kind from the voting study, and it is considerably stronger. A 2006 meta-analysis by Gollwitzer and Sheeran, as summarised in a later review by Wieber, Thürmer and Gollwitzer, combined 94 independent studies with more than 8,000 participants and found a medium-to-large average effect of if-then planning on goal attainment (d = 0.65). Note what this supports: making a plan that ties an action to a specific situation. It does not test identity, and that evidence does not rest on how you think of yourself.

Explore: if then planning · tiny habits

How long it takes, and why one slip is not the story

A number often attached to this topic is how long a habit takes to form. A 2010 study by Phillippa Lally and colleagues followed 96 volunteers who each picked one small daily behaviour tied to a regular cue, such as eating a piece of fruit at lunch, for 12 weeks. Of those, 82 provided enough data, and for 39 of them the curve of rising automaticity fitted well enough to estimate when it levelled off. For those 39, the estimates ranged from 18 to 254 days, and about 66 days is the figure usually quoted as the central one. The study ran only 84 days, so the longer estimates are model projections rather than observed results, and automaticity was measured by self-report.

The honest reading is that the range matters more than the middle. There is no fixed number of days, and the 66 should not be taken as a promise for any individual. The same study found that missing a single day did not materially disrupt the process, though it said little about many missed days. That is relevant to identity because it supports a gentler stance on slips than a vote metaphor can sometimes carry: one missed day was not shown to undo anything.

Where the idea goes wrong

The first problem is identity as pressure. A label such as I am a runner can motivate, and it can also become a standard you feel you are failing. The studies above do not measure this, so it is reasoning rather than a finding, but it follows from how self-labels work: if your sense of who you are rests on never skipping, then a skipped day becomes a threat to the self rather than a missed day. The same authors who found the voting effect noted it may not extend to behaviours people see as undesirable, which is a reminder that identity framing depends on the person already valuing the thing.

The second is all-or-nothing labelling. Treating each action as a vote can slide into counting votes, and a bad week then reads as evidence for the wrong identity. A more useful reading of the metaphor is the one Clear himself stresses: no single instance decides the outcome. A person who keeps score harshly is not using the idea the way it was meant.

The third is choosing an identity that is borrowed. Becoming the kind of person who wakes at five is not a meaningful goal if you do not care about mornings. The idea assumes the direction is yours. If you are unsure what you value, spending time on that question first may do more than any habit system.

Explore: self concept clarity · narrative identity

How a person might practise it

Start with a direction rather than a title. Instead of I am a writer, try something like I am someone who gives my ideas a place to land. A direction is easier to move toward on an off day and harder to fail at.

Choose an action small enough to do when you are tired, and tie it to a specific moment, using the same if-then structure the research supports: After I [current habit], I will [new habit]. The identity belongs in how you interpret the action afterwards, not in how hard you make it.

Notice the evidence you actually collect. At the end of a week, ask what the actions you took say about what you care about, and write it in your own words. If a slip happens, ask what it tells you about the plan, since a plan that fails on every bad day needs to be smaller. And check the label occasionally: if holding it feels like pressure and not like direction, loosen the wording.

Explore: atomic habits · if then planning

Where IX Coach fits

This page is about the idea, and nearly everything on it applies without any product. For people who would like to think out loud about who they are trying to become and what small actions would stand for it, IX Coach is an AI coaching system designed around asking questions rather than handing out plans. It is not a clinician and does not diagnose or treat. One way it can be used here is to put the direction into your own words and look at what a week of actions says about it.

The tradeoff is the one this page has repeated: the identity framing is a plausible tool with a modest and partly unreplicated evidence base, and a conversation cannot change that. Whether talking it through is useful to you is best found out by trying it.

Research

Practice this with IX Coach

Try this practice

Keep reading