Research

The research behind Habit Lab

Free — no account needed to read

Every paper here was checked against its DOI record. Each entry says, in plain and hedged words, what the study found and — just as important — what it does not show.

How habits form — and how long it takes

How are habits formed: Modelling habit formation in the real world (2010)

Lally, van Jaarsveld, Potts & Wardle. European Journal of Social Psychology. Longitudinal study. doi:10.1002/ejsp.674 — https://doi.org/10.1002/ejsp.674

What it found: Volunteers each chose one eating, drinking or activity behaviour to repeat daily in the same situation for 12 weeks and rated how automatic it felt. For those whose data could be modelled, reaching near their own plateau took a median of 66 days — ranging from 18 to 254 days. Missing a single opportunity did not noticeably harm the process.
What it does not show: A small sample, simple behaviours and self-rated automaticity; many participants’ data did not fit the model. The median is a description of this group, not a deadline for anyone.

Time to Form a Habit: A Systematic Review and Meta-Analysis of Health Behaviour Habit Formation and Its Determinants (2024)

Singh, Murphy, Maher & Smith. Healthcare. Meta analysis. doi:10.3390/healthcare12232488 — https://doi.org/10.3390/healthcare12232488

What it found: Pooled 20 studies of health habits. Where time to form a habit was reported, medians sat around two months and averages were longer, with individual times ranging from a few days to most of a year. Morning practices and habits people chose for themselves tended to be stronger.
What it does not show: Few studies, many with a high risk of bias, and mostly self-reported habit strength. It shows wide variation, not a rule.

What can machine learning teach us about habit formation? Evidence from exercise and hygiene (2023)

Buyalskaya, Ho, Milkman, Li, Duckworth & Camerer. Proceedings of the National Academy of Sciences. Longitudinal study. doi:10.1073/pnas.2216115120 — https://doi.org/10.1073/pnas.2216115120

What it found: Used machine learning on millions of gym check-ins and hospital hand-washing records. Contrary to a “magic number” of days, gym-going typically took months to become predictable, while hand-washing at work took weeks — and the cues that predicted behaviour differed from person to person.
What it does not show: Predictability of behaviour is not the same thing as how automatic it feels, and these are two specific behaviours in specific settings.

Exercise habit formation in new gym members: a longitudinal study (2015)

Kaushal & Rhodes. Journal of Behavioral Medicine. Longitudinal study. doi:10.1007/s10865-015-9640-7 — https://doi.org/10.1007/s10865-015-9640-7

What it found: Followed people who had just joined a gym. How often they exercised in the early weeks predicted how strong the habit became; in this sample, roughly four sessions a week kept up for about six weeks was the pattern linked with a habit forming.
What it does not show: One group of new gym members and self-reported habit strength. A description of this sample, not a prescription.

Experiences of habit formation: A qualitative study (2011)

Lally, Wardle & Gardner. Psychology, Health & Medicine. Qualitative study. doi:10.1080/13548506.2011.555774 — https://doi.org/10.1080/13548506.2011.555774

What it found: Interviews with people forming a new habit: many described the behaviour becoming easier and needing less thought over time, and valued it no longer relying on remembering or motivation.
What it does not show: A small interview study describing experiences, not measuring effects.

Forming a flossing habit: An exploratory study of the psychological determinants of habit formation (2013)

Judah, Gardner & Aunger. British Journal of Health Psychology. Field experiment. doi:10.1111/j.2044-8287.2012.02086.x — https://doi.org/10.1111/j.2044-8287.2012.02086.x

What it found: People asked to floss after brushing (rather than before) tended to form stronger flossing habits and, eight months later, still flossed more often — suggesting the end of an existing routine is a good place to attach a new behaviour.
What it does not show: A small, exploratory study of one behaviour; the authors call for more rigorous tests.

Habit formation following routine-based versus time-based cue planning: A randomized controlled trial (2021)

Keller, Kwasnicka, Klaiber, Sichert, Lally & Fleig. British Journal of Health Psychology. Randomised trial. doi:10.1111/bjhp.12504 — https://doi.org/10.1111/bjhp.12504

What it found: People linked a daily nutrition behaviour either to an existing routine or to a set time. Both kinds of plan increased automaticity, with no difference between them; for those who formed a habit, peak automaticity came after a median of 59 days. Repeatedly carrying out the plan was the key predictor.
What it does not show: One kind of behaviour over 84 days. It suggests the cue type matters less than repeating the plan in response to it.

Making health habitual: the psychology of ‘habit-formation’ and general practice (2012)

Gardner, Lally & Wardle. British Journal of General Practice. Review. doi:10.3399/bjgp12X659466 — https://doi.org/10.3399/bjgp12X659466

What it found: A short guide for doctors: advise people to choose a simple action, do it in response to the same cue in the same situation, and expect it to feel easier over weeks to months — and not to worry about the occasional miss.
What it does not show: A practical summary built on a small evidence base, not a new study.

Promoting habit formation (2013)

Lally & Gardner. Health Psychology Review. Review. doi:10.1080/17437199.2011.603640 — https://doi.org/10.1080/17437199.2011.603640

What it found: Reviewed how habits form in three phases — deciding to act, repeating the action in a consistent context, and the action becoming automatic — and which techniques support each phase.
What it does not show: A narrative review; the phases are a useful model rather than a measured sequence.

What a habit is

Habits in everyday life: Thought, emotion, and action (2002)

Wood, Quinn & Kashy. Journal of Personality and Social Psychology. Diary study. doi:10.1037/0022-3514.83.6.1281 — https://doi.org/10.1037/0022-3514.83.6.1281

What it found: People recorded what they were doing through the day. Roughly four in ten everyday actions were repeated almost daily in the same place, often while their thoughts were elsewhere.
What it does not show: Mostly student samples, and the share depends on how “habitual” was coded.

Habits—A Repeat Performance (2006)

Neal, Wood & Quinn. Current Directions in Psychological Science. Review. doi:10.1111/j.1467-8721.2006.00435.x — https://doi.org/10.1111/j.1467-8721.2006.00435.x

What it found: Habits are behaviours repeated in stable contexts until the context itself triggers them with little thought — which is why they can continue even after intentions change.
What it does not show: A short review summarising the research programme.

A new look at habits and the habit-goal interface (2007)

Wood & Neal. Psychological Review. Theory. doi:10.1037/0033-295X.114.4.843 — https://doi.org/10.1037/0033-295X.114.4.843

What it found: Proposed that habits are learned links between contexts and responses, built through repeated reward, which can run independently of what a person currently wants.
What it does not show: A theoretical account; parts of it are still debated.

Psychology of Habit (2016)

Wood & Rünger. Annual Review of Psychology. Review. doi:10.1146/annurev-psych-122414-033417 — https://doi.org/10.1146/annurev-psych-122414-033417

What it found: Summarised habit research: habits form through repetition in stable contexts, are cued directly by those contexts, and are not easily changed by information or good intentions alone.
What it does not show: A review; the field is still working out how habits and goals interact in everyday life.

Habit in Personality and Social Psychology (2017)

Wood. Personality and Social Psychology Review. Review. doi:10.1177/1088868317720362 — https://doi.org/10.1177/1088868317720362

What it found: Argued that habit is its own mechanism, distinct from goals and attitudes, and explains much of the everyday consistency people attribute to personality.
What it does not show: A position piece arguing for a research agenda.

Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior (1998)

Ouellette & Wood. Psychological Bulletin. Meta analysis. doi:10.1037/0033-2909.124.1.54 — https://doi.org/10.1037/0033-2909.124.1.54

What it found: Past behaviour predicted future behaviour through habit when the behaviour was frequent and the setting stable, and through conscious intentions when it was infrequent or the setting changed.
What it does not show: Correlational studies pooled together; it describes when habit matters, not how to change it.

Habit formation and change (2018)

Carden & Wood. Current Opinion in Behavioral Sciences. Review. doi:10.1016/j.cobeha.2017.12.009 — https://doi.org/10.1016/j.cobeha.2017.12.009

What it found: A short review of how habits form (repetition, stable context, reward) and how they change: disrupting cues, adding friction, and moving to new contexts.
What it does not show: A brief review of the research, not a new study.

Healthy through Habit: Interventions for Initiating & Maintaining Health Behavior Change (2016)

Wood & Neal. Behavioral Science & Policy. Review. doi:10.1353/bsp.2016.0008 — https://doi.org/10.1353/bsp.2016.0008

What it found: Argued that health programmes should build habit mechanisms — stable cues, repetition, and less friction — rather than relying on motivation, which tends to fade.
What it does not show: A policy-oriented perspective drawing on existing research.

Changing circumstances, disrupting habits (2005)

Wood, Tam & Witt. Journal of Personality and Social Psychology. Longitudinal study. doi:10.1037/0022-3514.88.6.918 — https://doi.org/10.1037/0022-3514.88.6.918

What it found: Students moving to a new university: when the context of an old habit (like exercising or watching television) changed, the habit was disrupted and their intentions guided what they did instead.
What it does not show: One kind of life transition, self-reported behaviour.

Empowering interventions to promote sustainable lifestyles: Testing the habit discontinuity hypothesis in a field experiment (2016)

Verplanken & Roy. Journal of Environmental Psychology. Field experiment. doi:10.1016/j.jenvp.2015.11.008 — https://doi.org/10.1016/j.jenvp.2015.11.008

What it found: People who had recently moved home responded more strongly to an intervention promoting sustainable behaviour than people who had not moved — a “window of opportunity” when old habits are disrupted.
What it does not show: One field experiment about environmental behaviours.

A review and analysis of the use of ‘habit’ in understanding, predicting and influencing health-related behaviour (2015)

Gardner. Health Psychology Review. Review. doi:10.1080/17437199.2013.876238 — https://doi.org/10.1080/17437199.2013.876238

What it found: Reviewed how “habit” is defined and measured in health research, arguing it should mean a process in which a context triggers an impulse to act, learned through repetition.
What it does not show: A conceptual review of the literature.

Habitual instigation and habitual execution: Definition, measurement, and effects on behaviour frequency (2016)

Gardner, Phillips & Judah. British Journal of Health Psychology. Measurement study. doi:10.1111/bjhp.12189 — https://doi.org/10.1111/bjhp.12189

What it found: Separated two kinds of habit: automatically deciding to start an action, and automatically carrying out its steps. In their data, the habit of starting was what predicted how often people did the behaviour.
What it does not show: Self-report measures in specific samples.

Plans, goals and monitoring

Implementation intentions: Strong effects of simple plans (1999)

Gollwitzer. American Psychologist. Review. doi:10.1037/0003-066X.54.7.493 — https://doi.org/10.1037/0003-066X.54.7.493

What it found: Introduced “implementation intentions”: linking a specific situation to a specific response (“when X happens, I will do Y”) helps people start, stay on track and resist distraction, because the situation itself comes to prompt the action.
What it does not show: A review of early studies by the idea’s originator.

Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes (2006)

Gollwitzer & Sheeran. Advances in Experimental Social Psychology. Meta analysis. doi:10.1016/S0065-2601(06)38002-1 — https://doi.org/10.1016/S0065-2601(06)38002-1

What it found: Pooled 94 independent tests and found that forming if–then plans had a medium-to-large effect on reaching goals, compared with intending to reach them alone.
What it does not show: Effects vary by goal, person and study quality, and plans help most when the goal is genuinely wanted. Some later, larger studies found smaller effects.

Do implementation intentions help to eat a healthy diet? A systematic review and meta-analysis of the empirical evidence (2011)

Adriaanse, Vinkers, De Ridder, Hox & De Wit. Appetite. Meta analysis. doi:10.1016/j.appet.2010.10.012 — https://doi.org/10.1016/j.appet.2010.10.012

What it found: If–then plans helped people change what they ate, with effects generally larger for eating more healthy food than for eating less unhealthy food.
What it does not show: Diet studies only, many of them short.

Breaking Habits With Implementation Intentions: A Test of Underlying Processes (2011)

Adriaanse, Gollwitzer, De Ridder, de Wit & Kroese. Personality and Social Psychology Bulletin. Lab experiment. doi:10.1177/0146167211399102 — https://doi.org/10.1177/0146167211399102

What it found: If–then plans that named the cue of an unwanted snacking habit and a replacement response helped break the habit, by building a new link from the same cue to the new response.
What it does not show: Laboratory and short-term studies of snacking.

Bridging the intention–behaviour gap: Planning, self-efficacy, and action control in the adoption and maintenance of physical exercise (2005)

Sniehotta, Scholz & Schwarzer. Psychology & Health. Longitudinal study. doi:10.1080/08870440512331317670 — https://doi.org/10.1080/08870440512331317670

What it found: In people starting exercise after cardiac rehabilitation, action plans (when, where, how) and coping plans (what to do when obstacles arise) helped turn intentions into exercise.
What it does not show: One clinical group; plans were one part of a wider process.

Modeling Health Behavior Change: How to Predict and Modify the Adoption and Maintenance of Health Behaviors (2008)

Schwarzer. Applied Psychology. Theory. doi:10.1111/j.1464-0597.2007.00325.x — https://doi.org/10.1111/j.1464-0597.2007.00325.x

What it found: Set out the Health Action Process Approach: a motivational phase (deciding) followed by a volitional phase (planning, starting and maintaining), with action planning, coping planning and confidence in one’s ability as key ingredients.
What it does not show: A model supported by many studies, but a model, not a law.

Thought contents and cognitive functioning in motivational versus volitional states of mind (1987)

Heckhausen & Gollwitzer. Motivation and Emotion. Lab experiment. doi:10.1007/BF00992338 — https://doi.org/10.1007/BF00992338

What it found: Before deciding, people weigh options open-mindedly; once they have decided, their thinking shifts to how and when to act. Deciding and doing are different mental modes.
What it does not show: Laboratory studies of the “Rubicon” model.

Future thought and behaviour change (2012)

Oettingen. European Review of Social Psychology. Review. doi:10.1080/10463283.2011.643698 — https://doi.org/10.1080/10463283.2011.643698

What it found: Reviewed research on mental contrasting: imagining a wished-for future and then the obstacle in the way energises action when success is feasible — more than positive fantasy alone. Combined with if–then plans it becomes WOOP.
What it does not show: Largely from one research group; effects depend on whether the goal is realistic.

Building a practically useful theory of goal setting and task motivation: A 35-year odyssey (2002)

Locke & Latham. American Psychologist. Review. doi:10.1037/0003-066X.57.9.705 — https://doi.org/10.1037/0003-066X.57.9.705

What it found: Specific, challenging goals, combined with feedback and real commitment, led to better performance than vague “do your best” goals across many studies.
What it does not show: Mostly task performance at work and in the lab; very hard goals can backfire on complex or unfamiliar tasks.

Unique effects of setting goals on behavior change: Systematic review and meta-analysis (2017)

Epton, Currie & Armitage. Journal of Consulting and Clinical Psychology. Meta analysis. doi:10.1037/ccp0000260 — https://doi.org/10.1037/ccp0000260

What it found: Setting goals had a small-to-medium effect on behaviour change on its own, with larger effects when goals were difficult, public, or set as a group.
What it does not show: Studies varied widely, and goal setting is usually combined with other techniques in practice.

Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence (2016)

Harkin, Webb, Chang, Prestwich, Conner, Kellar, Benn & Sheeran. Psychological Bulletin. Meta analysis. doi:10.1037/bul0000025 — https://doi.org/10.1037/bul0000025

What it found: Pooled 138 experiments with close to 20,000 people. Interventions that increased how often people monitored their progress also improved goal attainment on average, and effects were larger when progress was physically recorded or shared with others.
What it does not show: An average across very different goals; monitoring can feel like a chore, and what you choose to monitor matters.

Effective techniques in healthy eating and physical activity interventions: A meta-regression (2009)

Michie, Abraham, Whittington, McAteer & Gupta. Health Psychology. Meta analysis. doi:10.1037/a0016136 — https://doi.org/10.1037/a0016136

What it found: Programmes that combined self-monitoring with at least one other technique from control theory (such as goal setting, feedback or reviewing goals) were more effective than others.
What it does not show: A comparison across programmes, not a direct test of each technique.

Control theory: A useful conceptual framework for personality–social, clinical, and health psychology (1982)

Carver & Scheier. Psychological Bulletin. Theory. doi:10.1037/0033-2909.92.1.111 — https://doi.org/10.1037/0033-2909.92.1.111

What it found: Described self-regulation as a feedback loop: compare where you are with where you want to be, act to reduce the gap, and check again — the logic behind monitoring and review.
What it does not show: A conceptual framework rather than a test of any one technique.

Reminders Through Association (2016)

Rogers & Milkman. Psychological Science. Field experiment. doi:10.1177/0956797616643071 — https://doi.org/10.1177/0956797616643071

What it found: When people linked an intention to a distinctive cue they would meet at the moment of action, they followed through more often than people who had only intended to.
What it does not show: Mostly one-off tasks rather than repeated habits.

Changing behavior by memory aids: A social psychological model of prospective memory and habit development tested with dynamic field data (2009)

Tobias. Psychological Review. Theory. doi:10.1037/a0015512 — https://doi.org/10.1037/a0015512

What it found: A model tested on field data, suggesting that external memory aids help people remember, but that the habit which would carry the behaviour later can develop more slowly while they rely on the aid.
What it does not show: A model fitted to one setting; the trade-off is a hypothesis worth testing on yourself.

Beyond Self-Tracking and Reminders: Designing Smartphone Apps That Support Habit Formation (2015)

Stawarz, Cox & Blandford. Proceedings of CHI 2015. Field experiment. doi:10.1145/2702123.2702230 — https://doi.org/10.1145/2702123.2702230

What it found: Reviewed habit apps and ran a field study: reminders helped people repeat an action, but leaning on them could keep it from becoming automatic, while cues tied to an existing daily event supported habit formation better.
What it does not show: A small field study of one app and one kind of task.

Intentions, motivation and confidence

Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence (2006)

Webb & Sheeran. Psychological Bulletin. Meta analysis. doi:10.1037/0033-2909.132.2.249 — https://doi.org/10.1037/0033-2909.132.2.249

What it found: Experiments that produced a medium-to-large change in intentions produced only a small-to-medium change in behaviour — wanting to do something more is not the same as doing it.
What it does not show: Across many behaviours; the gap was larger for some than others.

The Intention–Behavior Gap (2016)

Sheeran & Webb. Social and Personality Psychology Compass. Review. doi:10.1111/spc3.12265 — https://doi.org/10.1111/spc3.12265

What it found: Reviewed why intentions often fail to become actions — not getting started, getting derailed, and not bringing the effort to a close — and which strategies help at each point.
What it does not show: A review that organises existing studies rather than testing something new.

How big is the physical activity intention–behaviour gap? A meta-analysis using the action control framework (2013)

Rhodes & de Bruijn. British Journal of Health Psychology. Meta analysis. doi:10.1111/bjhp.12032 — https://doi.org/10.1111/bjhp.12032

What it found: Close to half of people who intended to be physically active did not follow through, while almost nobody who did not intend to be active became active.
What it does not show: Physical activity only; other behaviours may differ.

Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being (2000)

Ryan & Deci. American Psychologist. Theory. doi:10.1037/0003-066X.55.1.68 — https://doi.org/10.1037/0003-066X.55.1.68

What it found: Proposed that motivation lasts when three psychological needs are met — autonomy (it is your choice), competence (you can do it) and relatedness (connection to others) — and that motivation you endorse outlasts motivation from pressure.
What it does not show: A broad, well-researched theory; how strongly each need matters varies by person and culture.

A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation (1999)

Deci, Koestner & Ryan. Psychological Bulletin. Meta analysis. doi:10.1037/0033-2909.125.6.627 — https://doi.org/10.1037/0033-2909.125.6.627

What it found: Expected, tangible rewards for doing an already-interesting activity tended to reduce people’s interest in it afterwards, while positive feedback tended to increase it.
What it does not show: Applies to activities people already enjoy; other researchers dispute how large and general the effect is.

For the Fun of It: Harnessing Immediate Rewards to Increase Persistence in Long-Term Goals (2016)

Woolley & Fishbach. Journal of Consumer Research. Lab experiment. doi:10.1093/jcr/ucv098 — https://doi.org/10.1093/jcr/ucv098

What it found: How much people enjoyed an activity while doing it predicted how long they persisted better than how much they valued its delayed benefits — so making the activity itself more enjoyable helps.
What it does not show: Short-term studies of persistence.

Self-efficacy: Toward a unifying theory of behavioral change (1977)

Bandura. Psychological Review. Theory. doi:10.1037/0033-295X.84.2.191 — https://doi.org/10.1037/0033-295X.84.2.191

What it found: Proposed that belief in your ability to do something (self-efficacy) shapes whether you try and keep going — and that it grows most from actually succeeding at it, more than from encouragement.
What it does not show: A foundational theory; later research supports it but debates its exact role.

Possible selves and academic outcomes: How and when possible selves impel action (2006)

Oyserman, Bybee & Terry. Journal of Personality and Social Psychology. Randomised trial. doi:10.1037/0022-3514.91.1.188 — https://doi.org/10.1037/0022-3514.91.1.188

What it found: A school programme helping students link their “possible future selves” to concrete strategies improved school behaviour and grades — identity helped when it was tied to specific actions.
What it does not show: One programme with adolescents; identity without strategies was not what was tested.

Toward a theory-based analysis of behavioral maintenance (2000)

Rothman. Health Psychology. Theory. doi:10.1037/0278-6133.19.Suppl1.64 — https://doi.org/10.1037/0278-6133.19.Suppl1.64

What it found: Argued that starting a behaviour and keeping it going rest on different decisions: people start because they expect benefits, and continue if they are satisfied with what they actually got.
What it does not show: A theoretical paper that proposes rather than tests the idea.

Theoretical explanations for maintenance of behaviour change: a systematic review of behaviour theories (2016)

Kwasnicka, Dombrowski, White & Sniehotta. Health Psychology Review. Systematic review. doi:10.1080/17437199.2016.1151372 — https://doi.org/10.1080/17437199.2016.1151372

What it found: Across many behaviour theories, keeping a change going was explained by five themes: motives for continuing, self-regulation, resources (physical and mental), habits, and environmental and social influences.
What it does not show: A review of theories, not of intervention results.

Self-control, willpower and the debates

Situational Strategies for Self-Control (2016)

Duckworth, Gendler & Gross. Perspectives on Psychological Science. Review. doi:10.1177/1745691615623247 — https://doi.org/10.1177/1745691615623247

What it found: Argued that strategies acting early — choosing and changing the situation before temptation arrives — are generally more effective and less effortful than resisting an urge in the moment.
What it does not show: A review and framework; the authors note these strategies are under-studied compared with in-the-moment willpower.

Everyday temptations: An experience sampling study of desire, conflict, and self-control (2012)

Hofmann, Baumeister, Förster & Vohs. Journal of Personality and Social Psychology. Diary study. doi:10.1037/a0026545 — https://doi.org/10.1037/a0026545

What it found: People reported their desires through the day. Those high in self-control reported meeting fewer temptations and less inner conflict — not heroic resistance more often.
What it does not show: Self-reported moments sampled on phones, in one population.

More than resisting temptation: Beneficial habits mediate the relationship between self-control and positive life outcomes (2015)

Galla & Duckworth. Journal of Personality and Social Psychology. Longitudinal study. doi:10.1037/pspp0000026 — https://doi.org/10.1037/pspp0000026

What it found: Across several studies, people high in self-control had more beneficial habits, and those habits helped explain their better outcomes — suggesting self-control works partly by building routines that need less willpower.
What it does not show: Correlational and longitudinal data; habits are one route among several.

Ego depletion: Is the active self a limited resource? (1998)

Baumeister, Bratslavsky, Muraven & Tice. Journal of Personality and Social Psychology. Lab experiment. doi:10.1037/0022-3514.74.5.1252 — https://doi.org/10.1037/0022-3514.74.5.1252

What it found: The original “ego depletion” studies: after one act of self-control, people persisted less on a second task, which the authors took as evidence that willpower draws on a limited resource.
What it does not show: Large later replication efforts found the effect much smaller than claimed, or close to zero — see Hagger and colleagues, 2016.

A Multilab Preregistered Replication of the Ego-Depletion Effect (2016)

Hagger, Chatzisarantis, Alberts, Anggono, Batailler, Birt and colleagues. Perspectives on Psychological Science. Replication. doi:10.1177/1745691616652873 — https://doi.org/10.1177/1745691616652873

What it found: Twenty-three laboratories ran the same pre-registered ego-depletion experiment. The combined effect was close to zero.
What it does not show: It tested one version of the procedure; researchers still debate whether other versions show an effect.

Revisiting the Marshmallow Test: A Conceptual Replication Investigating Links Between Early Delay of Gratification and Later Outcomes (2018)

Watts, Duncan & Quan. Psychological Science. Replication. doi:10.1177/0956797618761661 — https://doi.org/10.1177/0956797618761661

What it found: With a larger, more varied sample, the link between how long a child waited for a treat and later achievement was much smaller than in the famous original studies, and largely explained by family background and early abilities.
What it does not show: Correlational; it says little about adults building habits, but a lot about over-reading “willpower” tests.

Can’t Control Yourself? Monitor Those Bad Habits (2010)

Quinn, Pascoe, Wood & Neal. Personality and Social Psychology Bulletin. Diary study. doi:10.1177/0146167209360665 — https://doi.org/10.1177/0146167209360665

What it found: In diary studies, strong unwanted habits were controlled best by vigilant monitoring — thinking “don’t do it” and watching carefully for slip-ups — while other strategies helped with temptations but not with strong habits.
What it does not show: Self-reported everyday episodes plus one experiment; vigilance takes effort and suits specific moments.

Rewards, incentives and fresh starts

Holding the Hunger Games Hostage at the Gym: An Evaluation of Temptation Bundling (2014)

Milkman, Minson & Volpp. Management Science. Field experiment. doi:10.1287/mnsc.2013.1784 — https://doi.org/10.1287/mnsc.2013.1784

What it found: People who could listen to gripping audiobooks only while at the gym went more often at first than a comparison group — the “temptation bundling” effect — though the boost faded after a holiday break.
What it does not show: One field experiment, and the effect decayed over time.

Megastudies improve the impact of applied behavioural science (2021)

Milkman, Gromet, Ho, Kay, Lee, Pandiloski and colleagues. Nature. Megastudy. doi:10.1038/s41586-021-04128-4 — https://doi.org/10.1038/s41586-021-04128-4

What it found: Tested 54 different four-week programmes on more than 61,000 gym members at once. Many raised gym visits during the programme; among the strongest were small rewards for returning after a missed workout. Neither experts nor the public predicted well which would work.
What it does not show: Most effects faded after the programme ended, and the incentives were small.

The Fresh Start Effect: Temporal Landmarks Motivate Aspirational Behavior (2014)

Dai, Milkman & Riis. Management Science. Field experiment. doi:10.1287/mnsc.2014.1901 — https://doi.org/10.1287/mnsc.2014.1901

What it found: Gym visits, goal-related searches and commitments rose after temporal landmarks — a new week, month or year, a birthday, a holiday — which the authors called the “fresh start effect”.
What it does not show: Mostly observational data; a fresh start helps people begin, not necessarily continue.

Commitment Devices (2010)

Bryan, Karlan & Nelson. Annual Review of Economics. Review. doi:10.1146/annurev.economics.102308.124324 — https://doi.org/10.1146/annurev.economics.102308.124324

What it found: Reviewed commitment devices — arrangements people choose now to restrict or penalise their own future choices — and found they can help, but that many people who would benefit do not choose them, and some over-commit.
What it does not show: A review; results depend heavily on design.

Put Your Money Where Your Butt Is: A Commitment Contract for Smoking Cessation (2010)

Giné, Karlan & Zinman. American Economic Journal: Applied Economics. Randomised trial. doi:10.1257/app.2.4.213 — https://doi.org/10.1257/app.2.4.213

What it found: Smokers offered a savings account they would forfeit if they failed a test were more likely to have quit at follow-up than those not offered it — though most people offered it did not sign up.
What it does not show: One country and one behaviour; money-based commitments do not suit everyone.

Environment, defaults and nudges

Altering micro-environments to change population health behaviour: towards an evidence base for choice architecture interventions (2013)

Hollands, Shemilt, Marteau, Jebb, Kelly, Nakamura and colleagues. BMC Public Health. Systematic review. doi:10.1186/1471-2458-13-1218 — https://doi.org/10.1186/1471-2458-13-1218

What it found: Mapped “choice architecture” — changes to what is available, where things are placed, how they are presented and what the default is — and found it a promising but patchily studied way to change behaviour.
What it does not show: A scoping review: it organises the evidence rather than measuring overall effects.

Do Defaults Save Lives? (2003)

Johnson & Goldstein. Science. Field experiment. doi:10.1126/science.1091721 — https://doi.org/10.1126/science.1091721

What it found: Countries where people were organ donors by default had far higher consent rates than countries where people had to opt in — defaults strongly shape what people end up doing.
What it does not show: Country comparisons plus an online experiment; consent is not the same as behaviour.

The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains (2022)

Mertens, Herberz, Hahnel & Brosch. Proceedings of the National Academy of Sciences. Meta analysis. doi:10.1073/pnas.2107346118 — https://doi.org/10.1073/pnas.2107346118

What it found: Pooled over two hundred studies and found a small-to-medium average effect of nudges, with the largest effects for changes to the structure of a decision, such as defaults.
What it does not show: See Maier and colleagues (2022): after adjusting for publication bias, the average effect may be much smaller.

No evidence for nudging after adjusting for publication bias (2022)

Maier, Bartoš, Stanley, Shanks, Harris & Wagenmakers. Proceedings of the National Academy of Sciences. Meta analysis. doi:10.1073/pnas.2200300119 — https://doi.org/10.1073/pnas.2200300119

What it found: Re-analysed the same nudging research and argued that once missing, unpublished results are accounted for, there is little evidence of an average effect — though some kinds of nudge may still work.
What it does not show: A statistical re-analysis: it questions the average, not every individual technique.

Influencing behaviour: The mindspace way (2012)

Dolan, Hallsworth, Halpern, King, Metcalfe & Vlaev. Journal of Economic Psychology. Review. doi:10.1016/j.joep.2011.10.009 — https://doi.org/10.1016/j.joep.2011.10.009

What it found: Summarised nine robust influences on behaviour used in public policy — messenger, incentives, norms, defaults, salience, priming, affect, commitments and ego (MINDSPACE).
What it does not show: A policy framework, not a test of each influence.

Breaking habits, lapses and procrastination

Habit-reversal: A method of eliminating nervous habits and tics (1973)

Azrin & Nunn. Behaviour Research and Therapy. Field experiment. doi:10.1016/0005-7967(73)90119-8 — https://doi.org/10.1016/0005-7967(73)90119-8

What it found: Introduced “habit reversal”: becoming aware of the early signs of an unwanted habit and practising a competing response that makes the habit physically hard to do at that moment.
What it does not show: Developed for nervous habits and tics and refined in later clinical research; for compulsive or distressing habits it is used with a professional.

A large-scale experiment on New Year’s resolutions: Approach-oriented goals are more successful than avoidance-oriented goals (2020)

Oscarsson, Carlbring, Andersson & Rozental. PLOS ONE. Randomised trial. doi:10.1371/journal.pone.0234097 — https://doi.org/10.1371/journal.pone.0234097

What it found: People with “approach” resolutions (start doing something) were more likely to report success after a year than people with “avoidance” resolutions (stop doing something).
What it does not show: Self-reported success, with a lot of drop-out.

Auld lang Syne: Success predictors, change processes, and self-reported outcomes of New Year’s resolvers and nonresolvers (2002)

Norcross, Mrykalo & Blagys. Journal of Clinical Psychology. Longitudinal study. doi:10.1002/jclp.1151 — https://doi.org/10.1002/jclp.1151

What it found: People who made explicit resolutions were far more likely to report success six months later than people with the same goals who did not resolve; many slipped along the way, and slips often strengthened their commitment.
What it does not show: Self-reported outcomes; people who resolve may differ from those who do not.

If at first you don’t succeed: False hopes of self-change (2002)

Polivy & Herman. American Psychologist. Review. doi:10.1037/0003-066X.57.9.677 — https://doi.org/10.1037/0003-066X.57.9.677

What it found: Described “false hope syndrome”: people set overly ambitious, fast and easy expectations for self-change, fall short, and try again the same way — instead of adjusting their expectations and methods.
What it does not show: A theoretical review drawing mainly on dieting research.

Relapse Prevention for Alcohol and Drug Problems: That Was Zen, This Is Tao (2004)

Witkiewitz & Marlatt. American Psychologist. Review. doi:10.1037/0003-066X.59.4.224 — https://doi.org/10.1037/0003-066X.59.4.224

What it found: Updated relapse-prevention thinking: a single lapse is not a relapse, and what people make of a lapse — “I’ve ruined everything” versus “this was one moment I can learn from” — shapes what happens next.
What it does not show: Built for addiction treatment; everyday habits borrow the idea, not the clinical programme.

The nature of procrastination: A meta-analytic and theoretical review of quintessential self-regulatory failure (2007)

Steel. Psychological Bulletin. Meta analysis. doi:10.1037/0033-2909.133.1.65 — https://doi.org/10.1037/0033-2909.133.1.65

What it found: Procrastination was most strongly linked with how unpleasant a task feels, how distant its rewards are, and impulsiveness — not with laziness.
What it does not show: Correlational research pooled together.

Procrastination and the Priority of Short-Term Mood Regulation: Consequences for Future Self (2013)

Sirois & Pychyl. Social and Personality Psychology Compass. Review. doi:10.1111/spc3.12011 — https://doi.org/10.1111/spc3.12011

What it found: Argued that procrastination is largely about repairing a bad mood now — escaping the unpleasant feeling a task brings — at the expense of your future self.
What it does not show: A theoretical review.

I forgive myself, now I can study: How self-forgiveness for procrastinating can reduce future procrastination (2010)

Wohl, Pychyl & Bennett. Personality and Individual Differences. Longitudinal study. doi:10.1016/j.paid.2010.01.029 — https://doi.org/10.1016/j.paid.2010.01.029

What it found: Students who forgave themselves for procrastinating before one exam procrastinated less before the next.
What it does not show: One student sample and self-report.

Stages and processes of self-change of smoking: Toward an integrative model of change (1983)

Prochaska & DiClemente. Journal of Consulting and Clinical Psychology. Longitudinal study. doi:10.1037/0022-006X.51.3.390 — https://doi.org/10.1037/0022-006X.51.3.390

What it found: Proposed that people move through stages when changing — not yet considering it, considering, preparing, acting and maintaining — and use different processes at each.
What it does not show: Widely used and widely criticised — see West (2005).

Time for a change: putting the Transtheoretical (Stages of Change) Model to rest (2005)

West. Addiction. Review. doi:10.1111/j.1360-0443.2005.01139.x — https://doi.org/10.1111/j.1360-0443.2005.01139.x

What it found: Argued that the stages-of-change model draws arbitrary lines, assumes people make stable plans, and has not held up as a guide for matching help to stages.
What it does not show: An editorial critique; others still find the model useful as a way to start a conversation.

Measuring habits, brains and frameworks

Reflections on Past Behavior: A Self-Report Index of Habit Strength (2003)

Verplanken & Orbell. Journal of Applied Social Psychology. Measurement study. doi:10.1111/j.1559-1816.2003.tb01951.x — https://doi.org/10.1111/j.1559-1816.2003.tb01951.x

What it found: Introduced the Self-Report Habit Index: twelve statements about a behaviour — done frequently, automatically, without thinking, as part of who one is — rated for agreement.
What it does not show: Self-report: it measures how automatic a behaviour feels, not what happens in the brain.

Towards parsimony in habit measurement: Testing the convergent and predictive validity of an automaticity subscale of the Self-Report Habit Index (2012)

Gardner, Abraham, Lally & de Bruijn. International Journal of Behavioral Nutrition and Physical Activity. Measurement study. doi:10.1186/1479-5868-9-102 — https://doi.org/10.1186/1479-5868-9-102

What it found: Showed that the four automaticity items of the Self-Report Habit Index work as a short, valid measure of habit strength — the Self-Report Behavioural Automaticity Index.
What it does not show: Still a self-report measure of how automatic a behaviour feels.

Actions and habits: the development of behavioural autonomy (1985)

Dickinson. Philosophical Transactions of the Royal Society B. Lab experiment. doi:10.1098/rstb.1985.0010 — https://doi.org/10.1098/rstb.1985.0010

What it found: In animal studies, heavily repeated actions kept going even after their reward was made worthless (“outcome devaluation”), while lightly trained actions stopped — a classic test separating habit from goal-directed action.
What it does not show: Animal research; how cleanly it transfers to human everyday habits is debated.

Habits, Rituals, and the Evaluative Brain (2008)

Graybiel. Annual Review of Neuroscience. Review. doi:10.1146/annurev.neuro.29.051605.112851 — https://doi.org/10.1146/annurev.neuro.29.051605.112851

What it found: Reviewed how circuits in the basal ganglia help “chunk” repeated sequences of action into units that can run with little attention.
What it does not show: Largely animal research; brain findings explain mechanisms, not how any one person should build a habit.

Shifting the balance between goals and habits: Five failures in experimental habit induction (2018)

de Wit, Kindt, Knot, Verhoeven, Robbins, Gasull-Camos and colleagues. Journal of Experimental Psychology: General. Replication. doi:10.1037/xge0000402 — https://doi.org/10.1037/xge0000402

What it found: Five experiments tried to create habits in people through extended training in the lab and did not show the expected habit pattern — a caution about how easily lab habits are made and measured.
What it does not show: Laboratory tasks, which may differ from habits built over months of real life.

The behaviour change wheel: A new method for characterising and designing behaviour change interventions (2011)

Michie, van Stralen & West. Implementation Science. Theory. doi:10.1186/1748-5908-6-42 — https://doi.org/10.1186/1748-5908-6-42

What it found: Introduced COM-B — behaviour happens when there is Capability, Opportunity and Motivation — and the Behaviour Change Wheel for choosing interventions that target whichever is missing.
What it does not show: A framework for designing interventions: it organises rather than predicts.

The Behavior Change Technique Taxonomy (v1) of 93 Hierarchically Clustered Techniques (2013)

Michie, Richardson, Johnston, Abraham, Francis, Hardeman and colleagues. Annals of Behavioral Medicine. Taxonomy. doi:10.1007/s12160-013-9486-6 — https://doi.org/10.1007/s12160-013-9486-6

What it found: An international consensus listing 93 distinct behaviour-change techniques in 16 groups, so researchers can describe exactly what an intervention does. Habit Lab’s technique library is built on it.
What it does not show: A reporting language, not a ranking of which techniques work best.

A behavior model for persuasive design (2009)

Fogg. Proceedings of the 4th International Conference on Persuasive Technology. Theory. doi:10.1145/1541948.1541999 — https://doi.org/10.1145/1541948.1541999

What it found: Proposed that a behaviour happens when motivation, ability and a prompt come together at the same moment; when it is not happening, one of the three is missing.
What it does not show: A design model from a conference paper, later developed into the practitioner method Tiny Habits.

The role of deliberate practice in the acquisition of expert performance (1993)

Ericsson, Krampe & Tesch-Römer. Psychological Review. Longitudinal study. doi:10.1037/0033-295X.100.3.363 — https://doi.org/10.1037/0033-295X.100.3.363

What it found: Expert musicians had built up far more deliberate practice — focused, effortful work on specific weaknesses, with feedback — than less accomplished peers.
What it does not show: Later work found deliberate practice explains less of the difference between people than first claimed; other factors matter too.

Habit and physical activity: Theoretical advances, practical implications, and agenda for future research (2019)

Hagger. Psychology of Sport and Exercise. Review. doi:10.1016/j.psychsport.2018.12.007 — https://doi.org/10.1016/j.psychsport.2018.12.007

What it found: Reviewed habit research on physical activity, including the difference between the habit of deciding to exercise and the habit of doing each step, and what interventions can realistically build.
What it does not show: A review focused on one behaviour.