What Are Habits and Habitual Behaviors?

Everyday behaviors shape human health. Many of the dominant causes of death, including heart disease, diabetes, cancer, chronic lower respiratory diseases, and stroke, are preventable (World Health Organization, 2017). Adopting health-promoting behaviors such as eating more healthily or increasing physical activity may improve quality of life, physical and mental health, and extend lives (Aune et al., 2017; Centers for Disease Control and Prevention, 2014; Rebar et al., 2015; World Health Organization, 2015). For some behaviors, one performance is sufficient to attain desired health outcomes; a single vaccination, for example, can yield immunity to disease (e.g., Harper et al., 2004). For many behaviors, however, achieving meaningful health outcomes depends on repeated performance: Going for a run once, for example, will not achieve the same health benefits as regular activity over a prolonged period (Erikssen et al., 1998). In such instances, behavior change must be viewed as a long-term process, which can be conceptually separated into stages of initiation and maintenance (Prochaska & DiClemente, 1986; Rothman, 2000). This distinction is important from a practical perspective because while people may possess the capability, opportunity, and motivation to initiate behavior change (Michie, van Stralen, & West, 2011), they often fail to maintain it over time, lapsing back into old patterns of behavior (Dombrowski, Knittle, Avenell, Araujo-Soares, & Sniehotta, 2014). Some have attributed this to changes in motivation after initial experiences of action (Armitage, 2005; Rothman, 2000). People may overestimate the likelihood of positive outcomes or the valence of such outcomes, or they may fail to anticipate negative outcomes (Rothman, 2000). Alternatively, a newly adopted behavior may lose value and so become deprioritized over time. Motivation losses threaten to derail initially successful behavior change attempts.

Habit formation has attracted special attention as a potential mechanism for behavior change maintenance (Rothman, Sheeran, & Wood, 2009; Verplanken & Wood, 2006) because habitual behaviors are thought to be protected against any dips in conscious motivation. Viewing habit as a means to maintenance may seem truistic; in everyday discourse, a habit is an action done repetitively and frequently, and so making action habitual will necessarily entail maintenance. Within psychology, however, the term habit denotes a process whereby exposure to a cue automatically triggers a non-conscious impulse to act due to the activation of a learned association between the cue and the action (Gardner, 2015). Habit is learned through “context-dependent repetition” (Lally, van Jaarsveld, Potts, & Wardle, 2010): Repeated performance following exposure to a reliably co-occurring cue reinforces mental cue-action associations. As these associations develop, the habitual response gradually becomes the default, with alternative actions becoming less cognitively accessible (Danner, Aarts, & de Vries, 2008). Habit is formed when exposure to the cue is sufficient to arouse the impulse to enact the associated behavior without conscious oversight (Gardner, 2015; Neal, Wood, Labrecque, & Lally, 2012; Wood, Labrecque, Lin, & Rünger, 2014). In the absence of stronger influences favoring alternative actions, the habit impulse will translate smoothly and non-consciously into action, and the actor will experience behavior as directly cued by the context (Wood & Neal, 2007).

Defining habit as a process that generates behavior breaks with earlier definitions, which depicted habit as a form of behavior (see Gardner, 2015). This definition of habit as a process resolves a logical inconsistency that arises from portraying habit as a determinant of behavior (e.g., Hall & Fong, 2007; Triandis, 1980); as Maddux (1997, pp. 335–336) noted, “a habit cannot be both the behavior and the cause of the behavior.” It also allows for the habit process to manifest in multiple ways for any behavior. A distinction has been drawn between habitually instigated and habitually executed behavior (Gardner, Phillips, & Judah, 2016; Phillips & Gardner, 2016). Habitual instigation refers to habitual triggering of the selection of an action and a non-conscious commitment to performing it upon encountering a cue that has consistently been paired with the action in the past. Habitual execution refers to habit facilitating completion of the sub-actions that comprise any given action such that the cessation of one action in a sequence automatically triggers the next. Take, for example, “eating a bag of chips.” While people typically mentally represent this activity as a single unit of action (Wegner, Connally, Shearer, & Vallacher, 1983, cited in Vallacher & Wegner, 1987), it can be deconstructed into a series of discrete sub-actions (e.g., “opening bag,” “putting hand in bag,” “putting food in mouth,” “chewing,” “swallowing”; Cooper & Shallice, 2000). “Eating a bag of chips” is habitually instigated to the extent that the actor is automatically cued to select “eating chips” from available behavioral options. This may also activate the first sub-action in the sequence (“opening bag”). “Eating a bag of chips” is habitually executed to the extent that the cessation of, for example, “putting my hand in the bag” habitually cues “putting food in mouth,” the cessation of which habitually cues “chewing,” and so on, until the perceptually unitary action (“eating a bag of chips”) is complete.1 The term habitual behavior describes any action that is either instigated or executed habitually. This includes actions that are habitually instigated but non-habitually executed (e.g., habitually triggered to begin eating a bag of chips, but deliberates about how many chips to put in mouth), non-habitually instigated but habitually executed (e.g., consciously decides to eat a bag of chips, but habitually puts the chips in mouth, chews, and swallows), or both habitually instigated and habitually executed (e.g., habitually starts eating chips, and habitually puts them in mouth, chews, and swallows; Gardner, 2015). This description allows for a behavior to be habitual, yet not fully automated (see Aarts, Paulussen, & Schaalma, 1997; Marien, Custers, & Aarts, 2023) and better resonates with everyday experiences of complex health behaviors such as physical activity, which may be partly habit-driven, yet also require conscious oversight to be successfully completed (Rhodes & Rebar, 2023).

Habit has been implicated in behaviors across a range of domains, including media consumption (LaRose, 2010), purchasing patterns (Ji & Wood, 2007), environmentally relevant actions (Kurz, Gardner, Verplanken, & Abraham, 2014), and health behaviors. Studies have pointed to a multitude of health-related actions that may potentially be performed habitually, including dietary consumption (Adriaanse, Kroese, Gillebaart, & De Ridder, 2014), physical activity (Rebar, Elavsky, Maher, Doerksen, & Conroy, 2014), medication adherence (Hoo, Boote, Wildman, Campbell, & Gardner, 2017), handwashing (Aunger et al., 2010), and dental hygiene (Wind, Kremers, Thijs, & Brug, 2005). Habit strength is consistently found to correlate positively with behavioral frequency (Gardner, de Bruijn, & Lally, 2011; Rebar et al., 2016) and may bridge the “gap” between intention and behavior, though there are varying accounts regarding interplay between habits and intentions in regulating behavior. Some have argued that people are more likely to act on intentions when they have habits for doing so (Rhodes & de Bruijn, 2013). When motivation is momentarily low upon encountering associated contexts, habit may translate into performance despite motivational lapses. In this way, habit has been proposed to represent a form of self-control, protecting regularly performed behaviors that are desired in the longer-term against shorter-term motivation losses (Galla & Duckworth, 2015). Other studies have suggested that habit can direct action despite intentions not to act (Neal, Wood, Wu, & Kurlander, 2011; Orbell & Verplanken, 2010; but see Rebar et al., 2014). For example, one study showed that United Kingdom smokers with habits for smoking while drinking alcohol reported “action slips” after the introduction of a smoking ban in public houses; despite intending to adhere to the ban, several reporting “finding themselves” beginning to light up cigarettes while consuming alcohol (Orbell & Verplanken, 2010). These two perspectives concur in highlighting the potential for habit to override conscious motivational tendencies. Such effects may be attributable to habitual instigation rather than execution (Gardner et al., 2016); someone who is habitually prompted to act is more likely to frequently perform those actions and to do so without relying on intention.

The effects of habit—or more specifically, instigation habit (Gardner et al., 2016)—have important implications for behavior maintenance. By virtue of their cue-dependent, automatic nature (Orbell & Verplanken, 2010), habitually instigated behaviors should, in theory, persist even when they no longer serve the goal that initially motivated performance, or where motivation has eroded (Wood & Neal, 2007). For example, a person starting a new job out of town may consistently decide to commute by bicycle, which will likely create a habit for bicycle commuting whereby the workday morning context automatically prompts bicycle use without any deliberation over available alternatives (Verplanken, Aarts, Knippenberg, & Moonen, 1998). This may, however, lead to instances whereby the commuter “accidentally” uses the bicycle out of habit, despite, for example, knowing of road closures that will slow the journey and which would render alternative transport modes preferable (see Verplanken, Aarts, & Van Knippenberg, 1997). This example demonstrates several key features of habitual responses: learning via consistent pairing of cues (e.g., 8 a.m. on a workday) and action (selecting the bicycle); cue-dependent automaticity (using the bicycle at 8 a.m. on a workday without deliberation); and goal-independence, persisting even where an actor no longer has the motivation to act or is motivated to act in another way (e.g., when roads are closed). It also demonstrates how habit formation can maintain behavior by “locking in” new behaviors, protecting them against losses in conscious motivation. Habit development may also play a useful role in cessation of unwanted behaviors. Many ingrained behaviors—for example, eating high-calorie snacks—persist because they have become habitual and so are difficult to change. The lack of reliance on conscious intentions that is characteristic of habitual behavior, and which is thought to protect new behaviors against motivation losses, makes it difficult to break unwanted habits despite strong intentions to do so (Webb & Sheeran, 2006). While habit formation per se is not a sufficient strategy for “giving up” an unwanted behavior, behavior change can be made easier by seeking to form a new (“good”) habit in place of the old (“bad”) habit, rather than attempting only to inhibit the unwanted action (Adriaanse, van Oosten, de Ridder, de Wit, & Evers, 2011). Indeed, in the real world, habit development often involves displacing existing actions with more desirable alternatives such as eating healthy snacks in place of higher-calorie foods (Lally, Wardle, & Gardner, 2011; McGowan et al., 2013). Such “habit substitution” can take one of two basic forms, involving either avoidance of cues to the unwanted action or the development of new responses that compete with the unwanted habitual response. The “habit discontinuity hypothesis” speaks to the former of these, arguing that naturally occurring disruption of contexts—such as a residential relocation, for example—discontinues exposure to old habit cues (Walker, Thomas, & Verplanken, 2015). This represents an opportunity for people to act on their conscious motivation in response to newly encountered cues, and so to develop new, potentially more desirable habitual responses such as using active travel modes in place of more sedentary travel options like driving (Verplanken & Roy, 2016). Bad habits offer established cue-response structures that can hasten learning of new, good habits. Thus, where discontinued cue exposure is not feasible, people may seek to develop new cue-behavior associations to compete with and ultimately override old associations (Bouton, 2000; Walker et al., 2015). For example, people wishing to reduce habitual unhealthy snacking may form plans that dictate that when they are watching television and wish to snack (cue), they will eat fruit (new, desired behavior) instead of high-calorie foods (undesired, habitual behavior; e.g., Adriaanse, Gollwitzer, De Ridder, De Wit, & Kroese, 2011). In both instances of discontinued cue exposure and the adoption of competing responses to existing cues, the development of new habit associations and the decaying (or deprioritizing) of old habit associations are thought to occur concurrently (Adriaanse et al., 2011; Walker et al., 2015; Wood & Neal, 2007).

How Does Habit Form?

There have been calls for habit formation, whether focused solely on establishing new actions or displacing unwanted actions, to be adopted as an explicit goal for behavior change interventions (Rothman et al., 2009; Verplanken & Wood, 2006). Developing effective habit formation interventions requires an understanding of how habit forms.

The concept of behavior as an automatic response to covarying contextual cues, directed by learned cue-action associations, is rooted in behaviorist principles and studies of animal learning (e.g., Hull, 1943; Skinner, 1938; Thorndike, 1911). For example, in his maze-learning studies, Tolman (1932) noted that his rats, having repeatedly run down the route at the end of which was a food reward, continued to pursue that route even when the reward was removed. Adams (1982) trained rats to press a lever in a cage so as to receive intermittently delivered sucrose pellets. After receiving a lithium chloride injection that caused ingestion of the sucrose to induce nausea, those rats that were more highly trained (i.e., had pressed and received the sucrose reward a greater number of times in the training phase) were likely to persist longer in pressing the lever. Of course, unlike rats, humans possess the cognitive capacity to anticipate and reflect on their actions, and health-related behaviors among humans are inherently more complex than selecting maze routes or pressing levers. Yet, homologous neural processes are implicated in the acquisition and practice of habitual responses in rats and humans (Balleine & O’Doherty, 2010), and, like rats, people can acquire habitual behavioral responses despite a lack of insight into those behaviors or the associations that govern their performance (Bayley, Frascino, & Squire, 2005).

The route to human habit formation is conceptually simple: A behavior must be repeatedly performed in the presence of a cue or set of cues (i.e., context) so that cue-behavior associations may develop. For behaviors that are initially purposeful and goal-directed, the habit-formation process represents a period of transition whereby behavioral regulation transfers from a reflective and deliberative processing system to an impulsive system, which generates action rapidly and automatically based solely on activation of associative stores of knowledge (Strack & Deutsch, 2004). While there has been much lab-based research into the learning of relatively simple habitual responses in humans (e.g., button pressing; Webb, Sheeran, & Luszczynska, 2009), only relatively recently have studies focused on formation of real-world health-related habits (Fournier et al., 2017; Judah, Gardner, & Aunger, 2013; Lally et al., 2010). This work has largely been facilitated by the development of the Self-Report Habit Index (SRHI; Verplanken & Orbell, 2003), which affords reflections on the “symptoms” of habit, such as repetitive performance, mental efficiency, and lack of awareness.

Lally et al.’s (2010) seminal habit formation study used an SRHI sub-scale to assess the trajectory of the relationship between repetition and habit development among 96 participants for a 12-week period. They were instructed to perform a self-chosen physical activity or diet-related behavior (e.g., “going for a walk”) in response to a naturally occurring once-daily cue (e.g., “after breakfast”). Each day, they reported whether they had performed the action on the previous day, and if so, rated the experienced automaticity of its performance. Habit development within individuals was found to be most accurately depicted by an asymptotic curve, with early repetitions achieving sharpest habit gains, which later slowed to a plateau. The level at which habit peaked differed across participants, with some reportedly attaining scores at the high end of the automaticity index and others peaking below the scale mean. This plateau was reached at a median of 66 days post-baseline, though there was considerable between-person variation in the time taken to reach the plateau (18–254 days, the latter a statistical forecast assuming continued performance beyond the study period). These findings were echoed in a study of adoption of a novel stretching behavior (Fournier et al., 2017). Once-daily performance was found to yield asymptotic increases in self-reported habit strength. Habit plateaued at a median of 106 days for a group that performed the stretch every morning upon waking, and 154 days for those who stretched in the evening before bed, which the authors interpreted as evidence of the role of cortisol (which naturally peaks in the morning) in habit learning.

These studies reveal that habit development is not linear; if this were so, the fourth repetition of a behavior would have the same reinforcing impact on habit as would, say, the 444th. Rather, the asymptotic growth curve demonstrates that initial repetitions have the greatest impact on habit development. This in turn demands that the habit formation process be broken down into discrete phases and that the early phase, characterized by the sharpest gains in automaticity, may be a critical period during which people require most support to sustain motivation before the action becomes automatic (Gardner, Lally, & Wardle, 2012). Lally and Gardner (2013) have proposed a framework that organizes habit formation (and substitution) into four interlinked phases (see also Gardner & Lally, 2023). It argues that, for new behaviors initially driven by conscious motivation, habit forms when a person (1) makes a decision to act and (2) acts on his or her decision (3) repeatedly, (4) in a manner conducive to the development of cue-behavior associations. Phases 1 and 2 may be taken together to represent pre-initiation, occurring before the first enactment of the new behavior, whereas phases 3 and 4 are post-initiation phases, addressing the motivational and volitional elements needed to sustain behavior after initial performance (phase 3) and the effect of repetition on habit associations (phase 4) (see also Kuhl, 1984; Rhodes & de Bruijn, 2013; Rothman, 2000). Phase 3 captures the critical period after initiation but before habit strength has peaked (Fournier et al., 2017; Lally et al., 2010).

The framework is not intended as a theory or model of the habit formation process, but rather as a means to conceptually organize the processes and mechanisms that underpin habit development. According to the framework, any variable can promote habit formation in one or more of four ways: It may enhance motivation (phase 1) or action control (i.e., the enactment of intentions into behavior; Kuhl, 1984; Rhodes & de Bruijn, 2013) (phase 2) so as to initiate the behavior; it may modify motivation and other action control processes to continue to perform the behavior (phase 3); or it may strengthen cue-behavior associations (phase 4). One variable may operate through multiple processes: For example, anticipating pleasure from action can motivate people to perform it for the first time (phase 1) and to continue to perform it (phase 3) (Radel, Pelletier, Pjevac, & Cheval, 2017; Rothman et al., 2009). The experience of pleasure can also quicken learning of cue-behavior associations (phase 4) (de Wit & Dickinson, 2009). By extension, Lally and Gardner’s (2013) framework categorizes techniques that promote habit formation according to their likely mechanism (or mechanisms) of action; techniques may enhance motivation (phase 1) or action control (phase 2) to initiate change, sustain motivation and action control over time (phase 3), or reinforce cue-behavior associations (phase 4).

Which Behavior Change Techniques Should Be Used to Form Habit?

The most comprehensive taxonomy of behavior change techniques currently available defines habit formation as a discrete technique, which it defines as any effort to “prompt rehearsal and repetition of the behavior in the same context repeatedly so that the context elicits the behaviour” (Michie et al., 2013, Suppl. Table 3, p. 10). Yet, this definition incorporates only context-dependent repetition and not any other technique that may promote habit by increasing the likelihood of context-dependent repetition (i.e., promoting motivation or action control; phases 1–3 of Lally and Gardner’s framework) or enhancing the contribution of each repetition to the learning of habit associations (phase 4). Although context-dependent repetition is necessary for habit to form, it realistically requires supplementation with techniques targeting pre- and post-initiation phases en route to habit formation (Gardner Lally, & Wardle, 2012). While Michie et al. (2013) treat habit formation as a unitary technique, habit formation may perhaps be more realistically seen as an intervention approach that comprises a broader suite of techniques, which marry context-dependent repetition with strategies that: reinforce motivation; boost action control capacity, opportunity, or skills; facilitate post-initiation repetition; or quicken the learning of associations arising from repetition.

Theory points to techniques that may facilitate progression through these phases. Intention formation (phase 1 of Lally & Gardner’s [2013] framework) is likely when people anticipate that the action or its likely consequences will be positive and believe that they have a realistic opportunity and capability to perform the behavior (Ajzen, 1991; Bandura, 2001; Michie et al., 2011; Rogers, 1983; Schwarzer, Lippke, & Luszczynska, 2011). Providing information on the likely positive consequences of action, or choosing to pursue actions that are already most highly valued, may therefore aid habit development by enhancing motivation. Action control skills are required to initiate intention enactment (phase 2) and to maintain the behavior by consistently prioritizing the intention over competing alternatives (phase 3). This will likely be facilitated by self-regulatory techniques such as planning, setting reminders, self-monitoring, and reviewing goals to ensure they remain realistic and attractive, and receiving (intrinsic) rewards contingent on successful performance (Gardner et al., 2012; Lally & Gardner, 2013). People are most likely to engage in context-dependent repetition in response to highly salient cues (e.g., event- rather than time-based cues, which likely require conscious monitoring; McDaniel & Einstein, 1993). Pairing the action with more frequently and consistently encountered cues may quicken habit learning at phase 4 (Gardner & Lally, 2023). Highly specific action plans detailing exactly what will be done and in exactly which situation (i.e., implementation intentions; Gollwitzer, 1999) should therefore be conducive to the acquisition of associations (but see Webb et al., 2009). Implementation intentions can also facilitate habit substitution: By consistently enacting new, pre-specified cue responses that directly compete with existing habitual responses, such as feeding children water instead of sugary drinks (McGowan et al., 2013), new responses may acquire the potential to override and erode old habitual responses (Adriaanse et al., 2011). The reinforcing value of repetition may also be strengthened where intrinsic reward is delivered or attention is drawn to an undervalued intrinsic reward arising from action (Radel et al., 2017).

Which Behavior Change Techniques Have Been Used to Form Habit, and with What Effect?

While theory can recommend techniques that should be used to promote habit formation, evaluations of habit-based interventions are needed to show which techniques have been used, and with what effect, in real-world behavior change contexts. To this end, a systematic literature search was run to identify habit-based health-promotion interventions and to document the behavior change methods used.

Four psychology and health databases (Embase, Medline, PsycInfo, Web of Science) were searched in March 2018 to identify sources that had cited one of nine key papers about habit and health. These sources were selected to capture topics of habit measurement (Gardner, Abraham, Lally, & de Bruijn, 2012; Ouellette & Wood, 1998; Verplanken & Orbell, 2003), principles and processes of habit formation (Gardner, Lally, & Wardle, 2012; Lally & Gardner, 2013; Lally et al., 2010; Lally et al., 2011), and conceptual commentaries (Gardner, 2015; Wood & Rünger, 2016). Papers were eligible for review if they (a) were published in English, (b) were peer-reviewed, (c) reported primary quantitative or qualitative data, (d) had tested efficacy or effectiveness for changing behavior or habit, (e) used interventions designed to promote habit formation for health behaviors, (f) targeted context-dependent repetition, and (g) were informed by theory or evidence around habit, operationalized as a learned automatic response to contextual cues or a process that generates such responses. Interventions adopted primarily to elucidate the habit formation process (rather than to develop or assess intervention effectiveness; e.g., Judah et al., 2013; Lally et al., 2010) and any that focused exclusively on breaking existing habits (e.g., Armitage, 2016) were excluded. For each eligible intervention, all available material was coded, including linked publications (e.g., protocols), to identify component techniques using the Behavior Change Technique Taxonomy v1 (Michie et al, 2013).

Twenty papers, reporting evaluations of 19 interventions, were identified. Four of the 19 interventions represented variants of interventions used elsewhere in the 20 papers. For example, one trial evaluated the same habit-based intervention component in two conditions, which varied only in the frequency of supplementary motivational interviews and booster phone calls (Simpson et al., 2015). Thus, the 19 could be reduced to 15 unique habit-based interventions, of which four focused on both dietary and physical activity habits, six on physical activity (or sedentary behavior) only, two on dietary consumption only, two on dental hygiene, and one on food safety. In all of the studies, habit measures were self-reported.

Diet and Physical Activity Interventions

One randomized controlled trial (RCT) compared, in overweight and obese adults, an intervention that included advice on forming and substituting healthy for unhealthy habits, with a non-habit-based intervention that emphasized relationships with food, body image, and weight biases (Carels et al., 2014; see also Carels et al., 2011). Those in the habit-based intervention received training on changing old routines and developing new ones, including advice on using cues and forming implementation intentions. Both intervention groups received weekly weight assessments and monitored their physical activity, calorie intake, and output. At a 6-month follow-up, both the habit-based (n = 30) and non-habit intervention groups (n = 29) were eating a healthier diet, exercising more regularly, and had lost weight. Physical activity habit strengthened and sitting habit weakened in both groups, though no between-group differences were found in weight loss or habit strength.

Lally et al.’s (2008) “Ten Top Tips” weight loss intervention centered on a leaflet outlining recommendations for forming healthy eating and physical activity habits, as supplemented by a daily adherence monitoring diary. The leaflet included advice on routinization, identifying effective cues, and habit substitution. A small non-randomized trial compared the intervention, augmented with monthly (n = 35) or weekly weighing (n = 34), against a no-treatment control. The intervention group lost more weight than the control group at 8 weeks and maintained weight loss at 32 weeks. Scores at 32 weeks suggested the tips had become habitual, and habit change correlated positively with weight loss (Lally et al., 2008; see also Lally et al., 2011). In a subsequent RCT (Beeken et al., 2012, 2017), intervention recipients (n = 267) lost more weight at 3 months than did a usual-care group (n = 270). At 24 months, the intervention group had maintained weight loss, though the usual care group had lost a similar amount of weight. Habit strength, measured only at baseline and 3 months, increased more in the intervention than in the control group (Beeken et al., 2017). Weight loss at 3 months was attributable to gains in both habit and self-regulatory skill (Kliemann et al., 2017).

Simpson et al.’s (2015) weight-loss intervention provided participants with motivational advice designed to prompt intention formation, with information about how to form dietary and activity habits, and social support. Two intervention variants, differing according to the frequency of sessions, were evaluated against a minimal-treatment control, which did not feature habit-based advice, in a feasibility RCT among obese patients. Recipients of the more intensive intervention variant (n = 55) showed greater BMI reduction at a 12-month follow-up than did the less intensive intervention (n = 55) or control groups (n = 60). There were no between-group differences at 12 months in physical activity or overall healthy eating, nor were there differences in activity or diet habit scores.

One RCT compared an 8-week computer-tailored intervention designed to reduce cardiovascular risk against a no-treatment control among cardiac and diabetes rehabilitation patients who already intended to increase their activity and fruit and vegetable consumption (Storm et al., 2016). The intervention provided information about health risks of inactivity and unhealthy diet and enhancing self-regulatory skills. Immediately following intervention cessation, fruit and vegetable consumption and physical activity habit and behavior scores were greater among the intervention (n = 403) than control group (n = 387), but no differences were observed 3 months post-baseline.

Physical Activity and Sedentary Behavior Interventions

An intervention for new gym members promoted habits for both physical activity and preparatory actions for gym attendance (e.g., packing a gym bag; Kaushal, Rhodes, Meldrum, & Spence, 2017). Members received advice on how to form habits, including selecting time cues, setting action plans, and using accessories to increase enjoyment and so support cue-consistent performance and foster intrinsic motivation, which theory suggests can strengthen the impact of repetition on habit development (Lally & Gardner, 2013). Moderate-to-vigorous physical activity gains, objectively observed at an 8-week follow-up, were greater among intervention recipients (n = 47) than the no-treatment control group (n = 47). Habit strength was not assessed.

All 49 participants in Fournier et al.’s (2017) RCT were given access to twice-weekly, 1-hour tailored physical activity sessions for 28 weeks, with one group (n = 23) also sent SMS reminders targeting intrinsic motivation and consistent performance to the intervention group to foster habitual attendance. Although physical activity habit strength (assessed using a subscale of the SRHI) increased for both groups immediately post-intervention, the SMS group experienced quicker habit gains. Marginally greater activity was observed in the SMS group at 12 months.

One 4-month intervention for middle- to older-aged adults comprised seven 2-hour group sessions and sought to create new balance and strength exercise habits by recommending small modifications to everyday routines (e.g., placing frequently used items on high shelves to promote stretching to reach them) (Fleig et al., 2016; see also Clemson et al., 2012). An uncontrolled trial among 13 participants showed that, while there were no apparent changes in objectively measured physical performance, there were considerable habit strength gains for the recommended actions over 6 months. Notably, participants reported in interviews that the exercises had become automatically triggered, yet they performed them consciously, suggesting that the intervention promoted habitual instigation rather than execution.

Another intervention promoting small activity changes in older adulthood was evaluated in two papers (Matei et al., 2015; White et al., 2017). Drawing on Lally et al.’s (2008) “Ten Top Tips,” it comprised a leaflet offering recommendations for integrating and substituting light-intensity physical activities into everyday routines, with supplementary self-monitoring record sheets (Gardner, Thune-Boyle, et al., 2014). An 8-week uncontrolled trial was undertaken among two discrete samples (Matei et al., 2015). No changes were found in sitting time, physical activity, or sitting or physical activity habit among one sample (n = 16), but a second sample (n = 27) reported decreased sitting time and increased walking. Qualitative data suggested both groups experienced automaticity gains and some health benefits. A subsequent pilot RCT showed that intervention recipients (n = 45) experienced no greater change than did a control group (n = 46) who received a pre-existing fact sheet promoting activity and reducing sitting, but with no habit-based advice (White et al., 2017). Both groups reduced sitting time and sitting habit and increased activity and activity habit.

Using an experience sampling design, Luo et al. (2018) tracked change in standing or moving breaks from sedentary behavior in office workers given 3 weeks of access to automated computer-based reminders to break up sitting, timed to occur based on daily self-selected work and break durations. Although sitting behavior was not monitored, habit strength and self-regulation for taking “moving breaks” during work hours both increased significantly across the study.

Similarly, Pedersen et al. (2014) evaluated a software package that automatically deactivated desk-based employees’ computer screens every 45 minutes to substitute new physical activity habits for existing prolonged sitting habits. Although all participants received information on the detrimental health impact of sitting and benefits of activity, self-report activity data suggested that those who used the software for 13 weeks (n = 17) expended greater energy per day than did those not given the software (n = 17).

Dietary Interventions

One intervention promoted habitual healthy child-feeding practices among parents of children aged 2–6 years (McGowan et al., 2013). On each of four occasions over 8 weeks, parents chose to pursue one of four families of habit formation targets (increased feeding of fruit, vegetables, water, and healthy snacks). They received advice on the importance of child dietary consumption and on self-regulatory strategies, including action planning, goal setting, and context-dependent repetition. An RCT showed that intervention parents (n = 58) reported greater child intake of vegetables, water, and healthy snacks but a waiting-list control group (n = 68) did not. Habit strength increased for all three behaviors, and a habit score averaged across behaviors correlated with behavior change (McGowan et al., 2013; see also Gardner, Sheals, Wardle, & McGowan, 2014).

In one RCT, fruit and vegetable consumption changes were compared between participants who received habit-based messages, and those receiving general, non-habit-based tips for increasing consumption or messages about healthy eating more broadly (Rompotis et al., 2014). Notably, habit-based messages focused on anticipating stimulus control and environmental modification and on eating the same fruits and vegetables at the same time each day, so targeting both habitual instigation and execution (see Phillips & Gardner, 2016). The intervention was delivered via SMS in one set of conditions and email in the other. At 8-weeks post-intervention, both intervention groups (SMS n = 26, email n = 30) had increased fruit consumption and fruit habit strength, but those in all other conditions had not (SMS fruit and vegetable tips, n = 24, SMS healthy eating tips, n = 23; email fruit and vegetable tips, n = 29, email healthy eating n = 29). No effects were found on vegetable consumption or habit.

Oral Hygiene

Two school-based interventions aimed to increase tooth brushing in primary school children. One involved weekly dental hygiene lessons and daily tooth brushing practice time (Gaeta, Cavazos, Cabrera, & Rosário, 2018). School visits were also made by health promoters, and a seminar was held for teachers. One control group (n = 52) received the visits and seminar only, and a second control group (n = 52) received the seminar only. A quasi-experiment showed that children in the habit-based intervention (n = 106) and visits-and-seminar control group had less dental plaque, and a stronger tooth brushing habit at 12-week follow-up than did the seminar-only control group. The habit-based intervention group had the lowest plaque.

Wind et al.’s (2005) intervention also involved allocation of a designated tooth brushing time during the school day and encouragement from teachers. Tooth brushing rates increased in the intervention group (n = 141) during treatment but not in the control group (the nature of which could not be identified from the published report; n = 155). There were no differences in behavior at 12-months post-intervention nor in habit at any follow-up.

Food Safety

An intervention promoted the microwaving of dishcloths or sponges, for hygiene reasons (Mullan, Allom, Fayn, & Johnston, 2014). Recipients received emails and a poster providing instructions on how and why to microwave the dishcloths and sponges, designed to be placed in kitchens to act as a cue to the action. In an RCT, one intervention group was instructed to self-monitor their action, for intervention purposes, every 3 days (n = 15) and another every 5 days (n = 17). Relative to those who received an unrelated control treatment (n = 13), frequency and habit strength increased in the two intervention groups at 3 weeks and was sustained to the final 6-week follow-up.

Behavior Change Techniques Used in Previous Interventions

A total of 32 discrete behavior change techniques were each identified in at least one of the 15 interventions (see Table 1 and Table 2). Aside from context-dependent repetition itself—which, as an inclusion criterion, was necessarily present in all interventions—the most commonly used were “use prompts and cues” (present in 11 interventions; 73%), “action planning” (8 interventions; 53%), “provide instruction on how to perform the behavior” (8 interventions; 53%), “set behavioral goals” (8 interventions; 53%), and “self-monitor behavior” (7 interventions; 47%). Also common were “behavioral practice or rehearsal” (6 interventions; 40%), “provide information on health consequences” (6 interventions; 40%), and “problem solving” (5 interventions; 33%). “Behavioral substitution” and habit substitution (labeled “habit reversal” in the taxonomy) were each used in 4 interventions (27%).

Table 1. Behavior Change Techniques Identified in 15 Habit Formation InterventionsOpen in new tab

Technique

Number of interventions (%)

Context-dependent repetition*

15 (100)

Use prompts and cues

11 (73)

Action planning

8 (53)

Provide instruction on how to perform the behavior

8 (53)

Set behavioral goals

8 (53)

Self-monitor behavior

7 (47)

Behavioral practice and rehearsal

6 (40)

Provide information on health consequences

6 (40)

Problem solving

5 (33)

Behavior substitution

4 (27)

Habit reversal

4 (27)

Restructure the physical environment

4 (27)

Self-monitor outcomes of behavior

4 (27)

Add objects to the environment

3 (20)

Social support (practical)

3 (20)

Review behavioral goals

3 (20)

Feedback on behavior

3 (20)

Demonstration of behavior

2 (13)

Graded tasks

2 (13)

Nonspecific reward

2 (13)

Reduce prompts and cues

2 (13)

Social comparison

2 (13)

Social support (unspecified)

2 (13)

Avoid exposure to cues to behavior

1 (7)

Discrepancy between current behavior and goal

1 (7)

Focus on past success

1 (7)

Framing and reframing

1 (7)

Identification of self as role model

1 (7)

Information on social consequences

1 (7)

Nonspecific incentive

1 (7)

Self-incentive

1 (7)

Social support (emotional)

1 (7)

Open in new tab

Note. With the exception of “context-dependent repetition,” all technique labels are taken from the BCT Taxonomy v1 (Michie et al., 2013).

* This technique is labeled “habit formation” in the BCT Taxonomy v1 (Michie et al., 2013). Rephrasing this as “context-dependent repetition” more clearly delineates the underlying technique (i.e., to consistently repeat behavior in an unvarying context) from the outcome that it is designed to serve (i.e., to form habit). It also better acknowledges the possibility that such repetition may not lead to the formation of habit. For example, Lally et al. (2010) observed some participants who failed to attain peak habit strength in an 84-day study period, and some who experienced gains that peaked at low levels, suggesting that while repetition had rendered the behavior more habitual, the action remained predominantly regulated by conscious motivation rather than habit.

Table 2. Behavior Change Techniques Documented in 15 Habit Formation InterventionsOpen in new tab

Behavior and Reference

Techniques Used

Diet and physical activity

Carels et al. (2014)

Problem solving, action planning, self-monitoring behavior, self-monitoring outcomes, use prompts and cues, reduce prompts and cues, behavior substitution, context-dependent repetition,* habit reversal, nonspecific reward, restructuring the physical environment, avoid exposure to cues to behavior

Lally et al. (2010, 2011); Beeken et al. (2017); Kliemann et al. (2017)

Goal setting (behavior), discrepancy between current behavior and goal, self-monitoring behavior, self-monitoring outcomes, information on health consequences, use prompts and cues, behavior substitution, context-dependent repetition,* habit reversal, restructuring the physical environment

Simpson et al. (2015)

Problem solving, action planning, feedback on behavior, self-monitoring behavior, self-monitoring outcomes, social support (unspecified), information on health consequences, social comparison, context-dependent repetition*

Storm et al. (2016)

Goal setting (behavior), problem solving, action planning, review behavioral goals, feedback on behavior, social support (unspecified), instruction on how to perform behavior, information on health consequences, social comparison, context-dependent repetition*

Physical activity and sedentary behavior

Kaushal et al. (2017)

Goal setting (behavior), action planning, use prompts and cues, context-dependent repetition,* nonspecific reward, nonspecific incentive, adding objects to the environment

Fournier et al. (2017)

Use prompts and cues, context-dependent repetition*

Fleig et al. (2016)

Goal setting (behavior), problem solving, action planning, review behavioral goals, feedback on behavior, self-monitoring behavior, social support (practical), social support (emotional), instruction on how to perform behavior, demonstration of behavior, use prompts and cues, behavioral practice and rehearsal, context-dependent repetition,* graded tasks, focus on past success

Matei et al. (2015); White et al. (2017)

Goal setting (behavior), action planning, self-monitoring behavior, self-monitoring outcomes, instruction on how to perform behavior, information on health consequences, demonstration of behavior, behavior substitution, context-dependent repetition,* habit reversal, graded tasks, restructuring the physical environment, adding objects to the environment, framing and reframing

Luo et al. (2018)

Goal setting (behavior), action planning, self-monitoring behavior, instruction on how to perform behavior, use prompts and cues, behavioral practice and rehearsal, context-dependent repetition*

Pedersen et al. (2014)

Instruction on how to perform behavior, use prompts and cues, behavioral practice and rehearsal, context-dependent repetition*

Diet only

McGowan et al. (2013); Gardner, Sheals, et al. (2014)

Goal setting (behavior), problem solving, action planning, instruction on how to perform behavior, information on health consequences, information on social consequences, use prompts and cues, reduce prompts and cues, behavioral practice and rehearsal, behavior substitution, context-dependent repetition,* habit reversal

Rompotis et al. (2014)

Use prompts and cues, context-dependent repetition,* self-incentive, restructuring the physical environment, adding objects to the environment

Dental hygiene

Gaeta et al. (2018)

Goal setting (behavior), review behavioral goals, self-monitoring behavior, social support (practical), behavioral practice and rehearsal, context-dependent repetition*

Wind et al. (2005)

Social support (practical), instruction on how to perform behavior, use prompts and cues, behavioral practice and rehearsal, context-dependent repetition*

Food safety

Mullan et al. (2014)

Instruction on how to perform behavior, information on health consequences, use prompts and cues, context-dependent repetition*

Open in new tab

Note. All technique labels are taken from the BCT Taxonomy v1 (Michie et al., 2013).

* This technique is labeled “habit formation” in the BCT Taxonomy v1 (Michie et al., 2013). Rephrasing this as “context-dependent repetition” more clearly delineates the underlying technique (i.e., to consistently repeat behavior in an unvarying context) from the outcome that it is designed to serve (i.e., to form habit). It also better acknowledges the possibility that such repetition may not lead to the formation of habit. For example, Lally et al. (2010) observed some participants who failed to attain peak habit strength in an 84-day study period, and some who experienced gains that peaked at low levels, suggesting that while repetition had rendered the behavior more habitual, the action remained predominantly regulated by conscious motivation rather than habit.

While all 15 interventions were based on the principle of habit formation, none used context-dependent repetition as a standalone technique.2 The use of techniques additional to repetition echoes the view that in the real world, habit is best promoted by embedding context-dependent repetition into a broader package of techniques that also target motivation and action control, which are prerequisites for repetition (Lally & Gardner, 2013). Techniques most commonly adopted in past interventions have focused predominantly on action control (e.g., planning, goal-setting, identifying cues, rehearsing action, problem solving). The relative paucity of techniques targeting motivation may reflect an assumption that, for most of the behaviors targeted, intervention recipients generally recognize the value of behavior change, but lack the volitional skills, opportunities, or resources to change. Whether motivation should be targeted as part of a habit-formation intervention will depend on whether target populations understand the need for change and prioritize the target behavior above alternatives.

Fewer than half of the 15 interventions appear to have addressed factors that may moderate the relationship between repetition and habit development. Theory and evidence suggest that the mental associations that underlie habit will develop most strongly or quickly where actions are more simple or intrinsically rewarding and in response to cues that are salient and consistently encountered (Lally & Gardner, 2013; McDaniel & Einstein, 1993; Radel et al., 2017). Several of the reviewed interventions purposively promoted habit formation for simple behaviors (Beeken et al., 2017; Fleig et al., 2016; Lally et al., 2010, 2011; Matei et al., 2015; Mullan et al., 2014; White et al., 2017). Kaushal et al. (2017) emphasized the importance of intrinsic reward in their physical activity promotion intervention, and Fournier et al. (2017) targeted intrinsic motivation. These studies highlight how interventions may move beyond simply promoting repetition toward targeting factors that may reduce the number of repetitions required for a target behavior to become habitual.

How Should Habit-Based Interventions Be Evaluated?

Previous interventions attest to the potential for habit-based approaches to change behavior. Although many intervention studies were not designed to test effectiveness, 13 of the 15 interventions were associated with positive change on at least one index of behavior or behavior-contingent outcomes (e.g., weight loss) at one or more follow-ups. Process evaluations pointed to the strengthening of habit as a key mechanism underpinning behavioral change based on increases in self-reported automaticity scores or qualitative reflections on the subjective experience of automaticity (Fleig et al., 2016; Gardner, Sheals, et al., 2014; Kliemann et al., 2017; Lally et al., 2011; Matei et al., 2015). Additionally, acceptability studies have suggested that recipients find the concept of context-dependent repetition—which distinguishes habit-based and non-habit-based interventions—easy to understand and follow (Fleig et al., 2016; Gardner, Sheals, et al., 2014; Lally et al., 2011; Matei et al., 2015).

Limitations of evaluation methods preclude understanding of how best to support habit formation. It is not yet clear whether promotion of context-dependent repetition is necessary for habit to develop or, indeed, whether it represents the most “active” ingredient of a habit formation intervention. One study found that a control group that did not receive habit-based advice reported similar physical activity habit gains to those among a group that received habit guidance (White et al., 2017). Conversely, another study showed that intervention recipients deviated from habit-based advice (e.g., by setting goals that were not specific, measurable, or achievable), yet habit strengthened (Gardner, Sheals, et al., 2014). Habit formation may therefore arise as a byproduct of interventions that do not explicitly target habit development. The unique contribution of context-dependent repetition to behavior change remains unclear because none of the reviewed studies compared a habit-based intervention with an otherwise identical non-habit-based equivalent. Indeed, most studies have evaluated habit formation interventions against minimal-treatment control groups or used uncontrolled designs. Future research should seek to compare matched habit- and non-habit-based interventions or otherwise use factorial designs, which allow testing for isolated effects within a multicomponent intervention, or mediation analyses, which can assess whether habit change underpins intervention effects.

Intervention evaluations have also been limited by short follow-up periods, which is ironic given that the key purported benefit of incorporating habit formation into interventions is the potential to increase longevity of behavior change. Few studies evaluated outcomes over 12 months or longer, with the longest observed follow-up being 24 months (Beeken et al., 2017). Beeken et al.’s (2017) “Ten Top Tips” intervention showed greater impact than did a non-habit-based usual-care treatment on dietary and physical activity habits, and weight loss, at the 3-month follow-up, which the authors found to be due in part to habit development (Kliemann et al., 2017). Yet, while weight loss was maintained at 24 months, the advantage conferred by the habit-based intervention over usual care was lost, suggesting that any habit gains may have dissipated, or alternatively, that for those who were successful in maintaining the behaviors over the 2-year period, habit formation had occurred regardless of condition. These possibilities cannot be investigated because habit strength was not evaluated at 24 months. Elsewhere, however, a small exploratory (non-intervention) study suggested that habit gains may erode over time: Among a group of participants forming dental flossing habits over 8 weeks, habit strength had considerably eroded in the subgroup of participants who provided data at a 6-month follow-up (Judah et al., 2013). Until more is done to assess the longevity of habit-based intervention effects, the hypothesis that habit persists over time, and so supports behavior maintenance, remains insufficiently tested.

Conclusion

Theory proposes that, through consistent performance, behaviors become habitual such that they are initiated automatically upon encountering cues via the activation of learned context-behavior associations. Habitual behaviors are thought to be self-sustaining, and so forming a habit has been proposed as a means to promote long-term maintenance of behavior. Interventions that seek to promote habit formation should include not only advice on context-dependent repetition, but also techniques that support the motivation and action control needed to repeat the action and that may enhance the reinforcing value of repetition on habit development. Fifteen interventions were found to have used habit formation principles to encourage engagement in health-promoting behaviors, and these have tended to supplement advice on repetition with action control techniques. Previous research suggests a habit-based approach has much to offer to behavior change initiatives; habit formation offers an acceptable, easily understood intervention strategy, with the potential to change behavior and yield favorable health outcomes. Yet, the unique effects of habit-specific techniques, and the longevity of effects, have not been adequately explored. The central assumption of the habit-based approach—that habit gains translate into long-term behavior maintenance—remains largely untested.