What Is a Habit? Behaviour That Stopped Needing a Reason

Behavioural research distinguishes actions performed because of their expected outcome from actions performed because a situation has previously been followed by them. The standard experimental test is devaluation: an animal or person is trained to perform an action for a reward, the reward is then made undesirable, and the question is whether the behaviour stops. Goal-directed behaviour stops immediately, since the outcome is no longer wanted. Habitual behaviour continues, since it was never about the outcome, and that persistence is the operational definition. Extensive repetition shifts behaviour from the first category to the second, which is efficient, since deliberating about every routine action would be exhausting, and which is exactly why habits are hard to change once formed. Different neural systems are implicated, with the shift associated with a transition in which parts of the striatum are involved, and the practical implication is that reasoning about consequences addresses only one of the two systems.

Formation depends on a small number of factors and the popular account gets several wrong:

• •Context, meaning a stable cue such as a time, place, preceding action or emotional state, which is what the behaviour becomes attached to
• •Repetition in that context, which is the mechanism, and which matters more than motivation once the process is under way
• •Reward during formation, which drives the initial learning and becomes unnecessary once the habit is established
• •Time, which varies enormously, with one frequently cited study finding a median around sixty-six days and a range from under three weeks to well over eight months depending on the behaviour
• •The widely repeated claim that twenty-one days suffices, which traces to a plastic surgeon's observation about patients adjusting to altered appearance and was never a finding about habit formation at all
• •Consistency of context mattering more than consistency of performance, since missing an occasional day does not disrupt formation measurably

The difficulty follows from the mechanism. A habit cannot be deleted, since the association between context and action persists, and what is learned instead is a competing response, which means the original remains available and reappears under stress, fatigue or when the competing response is not maintained. Contexts are numerous and frequently unnoticed, so a person may identify one trigger and remain exposed to several others. Willpower approaches address the deliberative system while the habitual one continues operating, which is why intentions predict behaviour poorly for well-established habits and predict it reasonably well for new ones. The interventions with better evidence work on the context rather than on the intention: changing the environment so the cue is absent or the action is inconvenient, substituting a different response to the same cue, and using implementation intentions that specify exactly when and where the alternative will be performed. Life transitions that disrupt context, including moving house or changing job, are documented windows in which habits change more readily.

Habit has become a substantial commercial category and several claims in it run ahead of the evidence. The cue, routine, reward loop is a reasonable simplification and is presented with more precision than the research supports. Keystone habits, the idea that one change cascades into others, has limited direct evidence. Habit stacking and similar techniques are plausible applications of context learning without being separately validated. Claims about identity change driving behaviour are supported in part by work on self-concept and are frequently stated far more strongly than that work permits. The twenty-one day figure is simply wrong and continues circulating. What is reasonably well supported is narrower: context cues drive established behaviour, repetition in stable contexts builds automaticity, environmental change is more effective than resolve, and specific if-then plans outperform general intentions. That is useful and considerably less dramatic than the genre requires.

A habit persists when the outcome is no longer wanted, which is the operational test and shows it is triggered by context rather than chosen. Formation takes a median of about sixty-six days with enormous variation, and the twenty-one day claim comes from a surgeon's note about patients adjusting to appearance. Habits cannot be deleted, only competed with, which is why changing the environment beats resolving to do better.