The impact of action plans on habit and intention strength for physical activity in a web-based intervention: is it the thought that counts?

Abstract

Objective

Action planning is a common approach used in physical activity interventions. The aim of this study was to assess the association of frequency, consistency and content of action planning with physical activity behaviour, intention strength and habit strength.

Methods and Measures

Within a 3-month web-based, computer-tailored physical activity intervention, participants (N = 115; 68.7% female, M age =43.9; range = 22–73 years) could create 6 rounds of action plans for 4 activities each (24 total).

Results

Consistency of action planning during the intervention was associated with change in physical activity at 9-months, and intention and habit strength at 3-months and 9-months. Frequency of action planning was negatively associated with intention at 3-months and 9-months. The effect of action planning consistency on physical activity behaviour was no longer significant when accounting for change in intention and habit strength.

Conclusion

Consistency of how, where, when and with whom people plan their physical activity may translate into stronger physical activity habits. Interventions should avoid encouraging making many distinct action plans, but rather encourage stable contexts through consistent action planning.

Keywords:

• Habit formation
• motivation
• exercise
• behavioural maintenance
• implementation intentions

Physical activity can extend the quality and longevity of life (2018 Physical Activity Guidelines Advisory Committee, 2018; Warburton et al., 2006). Some benefits of physical activity are accrued immediately. For example, immediately after a bout of physical activity, people benefit from reduced feelings of anxiety, improved sleep and cognition. However, most health benefits of physical activity including lower risk of cardiovascular disease, hypertension, type 2 diabetes and cancers are accrued over the long-term with regular engagement in physical activity (Physical Activity Guidelines Committee, 2008). Unfortunately, as few as 1-in-4 adults engage in regular physical activity globally (World Health Organization, 2018). More is needed to understand how to increase adults’ long-term engagement in physical activity.

Physical activity habits

One mechanism for enhancing long-lasting physical activity behaviour change is through targeting habit formation (Rebar et al., 2018). Habit is the process by which behaviour is influenced from a prompt to act based on well-learned associations between cues and behaviours (Gardner, 2015; Wood & Neal, 2016). With consistent engagement in the same behaviour in the same context over time, we learn to associate cues within that context (e.g. time of day, mood, people around, part of routine) to the behaviour (Rebar, 2017). When behaviour change is intentional, a person initially engages in behaviour through willed implementation of strong intentions, and then habit formation occurs when these cue-behaviour associations strengthen, typically through repetition of experience (Carden & Wood, 2018; Gardner & Rebar, 2019). The strong mental context-behaviour associations that underpin habits lead to urges to enact the behaviour when we experience the context cue (Carden & Wood, 2018). If the cue is experienced regularly, then the habit makes it likely that the behaviour will be enacted frequently and consistently over time. So, while physical activity behaviour change may be initiated through conscious regulation of intentions, as habit forms and the mental cue-response associations become well-learned, the behaviour will become also driven by strength of habit (Verplanken & Melkevik, 2008; Wood & Neal, 2007). We tend to engage in our habitual behaviours even if we are stressed, busy, or not motivated (Gardner & Rebar, 2019; Rebar et al., 2014). There is strong evidence to suggest that if people form strong physical activity habits, they tend to engage in more physical activity than people with weaker habits (Gardner et al., 2011; Rebar et al., 2016). Theoretically, over longer terms, motivation for behaviour shifts from intentional to habitual regulation, inferring that habit should drive behaviour change maintenance (Gardner et al., 2020; Kwasnicka et al., 2016).

Action planning

Habits form with repeated cue-behaviour pairings, so physical activity programs aimed at habit formation should encourage people to engage in physical activity repeatedly in consistent contexts. One behaviour change technique used for linking cues to behaviours is action planning (Michie et al., 2013). Action plans are a technique of self-regulation defining specific plans about the behaviour, location, frequency, duration, or intensity for the planned activity (Leventhal et al., 1965; Sniehotta et al., 2005). Whereas goals describe a desired outcome, action plans describe how, when, where and with whom goals are planned to be achieved. For example, a goal may be to increase physical activity, but an action plan would be to start doing yoga for 45 min every morning on the front porch by yourself. Action plans have been shown to lead to small increases in physical activity behaviour, although notably the effect of action plans on behaviour is quite varied, with stronger effects for older people, rehabilitation patients and those who are inactive, but highly motivated (Carraro & Gaudreau, 2013). Action plans are a compelling intervention approach for physical activity, given that they are easy to implement at a large scale, for example through digital (e.g. web- or mobile-based) interventions (Rhodes et al., 2020; Vandelanotte et al., 2015, 2021).

Action plans specify both behaviours and cues, and so may assist in the cue-behaviour association learning that underpins habit formation (de Bruijn et al., 2014; Fleig et al., 2013; Rhodes, 2017). Theoretically, action plans may help people maintain strong intentions and form habits because the planned context-behaviour pairings become well-learned and automatic through repetition of engagement of the behaviour in response to the context cues (Sheeran & Orbell, 1999; Webb & Sheeran, 2007). Indeed, action planning has been shown to lead to habit formation for behaviours including flossing, nutrition and physical activity (Fleig et al., 2013; Judah et al., 2013; Lally et al., 2010; Orbell & Verplanken, 2010). Although there is some inconsistency amidst the evidence of the conditions and directionality of the effect, most evidence suggests action planning can lead to increased physical activity habit strength and physical activity behaviour (Fleig et al., 2013; Maltagliati et al., 2022; Schwarzer et al., 2018; van Bree et al. 2016).

What’s in an action plan?

Some evidence suggests characteristics of action planning impacts its translation into behaviour change. Studies show that people who make more specific action plans are more effective at implementing their plans than those who make less specific plans (de Vet et al., 2011; Dombrowski et al., 2016). A review found that people who make plans with four components (e.g. what, when, where and with whom) were more effective than those who made plans with fewer components (Carraro & Gaudreau, 2013). Evidence also shows that plans focused on places in a routine as opposed to specific times of day are more likely to be enacted successfully (Keller et al., 2017). Clearly, characteristics of action planning are important but most of this literature focusses on single time action planning. Little is known about how the use of repeated bouts of action planning impact its effectiveness in increasing physical activity.

Frequency of action planning

It may be that a single application of action planning is not effective, but some argue that more frequent action planning is needed to form strong habits (Hagger & Luszczynska, 2014). Generally, the more people tend to engage with physical activity interventions, the more their physical activity behaviour increases (Smith & Liu, 2020; Vandelanotte et al., 2007). Given that habits form through repeated cue-behaviour exposure (Carden & Wood, 2018; Gardner & Rebar, 2019), it is reasonable to suspect that more frequent use of action planning would lead to more impact on behaviour, intention strength and habit formation.

Consistency of action planning

Consistency of context when engaging in physical activity is a potentially important, but often overlooked, predictor of physical activity behaviour. Evidence shows that the more consistent people are in the time, place and people they are with when they engage in physical activity, the more physically active they tend to be (Kaushal et al., 2017; Maher et al., 2021; Schumacher et al., 2019). It may be that when people form multiple action plans that are more consistent in the planned activity, time, place and social surroundings, they will form stronger habits and engage in more physical activity behaviour over the long-term. Given that intentions are not reliant on specifics of how physical activity is done, it is not anticipated that consistency of action planning would have an impact on intention strength.

Content of action planning

Finally, certain activities, places, times of day, or social surroundings may be more or less beneficial for the formation of physical activity. Some planned circumstances may be more prone to barriers than others. For example, friends may be unavailable sometimes or weather may interfere with plans for outdoor exercise. By testing the effects of frequency, consistency and content of action planning, this study provides unique insight into this commonly applied behaviour change technique.

The present study

This study is a secondary analysis of a three-arm randomised controlled trial of a web-based computer-tailored physical activity intervention, TaylorActive (Vandelanotte et al., 2015, 2021). As part of the intervention, participants in the two intervention groups were encouraged to complete action plans for physical activity across 6 different intervention modules delivered across 3 months. Within each of the 6 modules, participants could create action plans for 4 activities, for a total of 24 possible action plans created per participant. Assessments of monitor-assessed physical activity behaviour and self-reported physical activity intention and habit strength were conducted at baseline, completion of the intervention (3-months post baseline) and at a 6-month follow-up (9-months post baseline). It was hypothesised that both intention strength and habit strength would be positively associated with change in physical activity behavioural outcomes. It was also hypothesised that more frequent action planning during the intervention would lead to more engagement in physical activity behaviour and stronger physical activity intention and habit strength. Additionally, it was hypothesised that more consistent action planning during the intervention would lead to more engagement in physical activity behaviour and stronger physical activity habit strength, but not intention strength. Given the lack of existing evidence on content of action planning, no hypotheses were formed regarding whether certain activities, places, times of day, or certain surroundings would translate into more or less behaviour change, intention, or habit strength.

Materials and methods

Study design

The protocol and main outcomes of the TaylorActive intervention have been previously published (Vandelanotte et al., 2015, 2021). Briefly, TaylorActive was a three-arm intervention in which participants were randomised into one of three groups: video-tailored, text-tailored and control. All participants had access to the same website which housed a library of text-based articles about physical activity including ‘Why be active’, ‘Get started walking’, and ‘Make time to be active’. The two intervention groups also had access to eight personally-tailored sessions delivered over three months. Assessments of monitor-assessed physical activity, habit strength and intention strength were conducted at baseline (before the intervention), 3-months following baseline (immediately after the intervention ceased) and 9-months following baseline (6 months after the intervention ceased). All participants received regular reminders to return to the website throughout the study, including three reminders to complete sessions. The trial was registered with the Australian New-Zealand Clinical Trial Registry (ACTRN12615000057583, Registered 22 January 2015). All study procedures were pre-approved by the local Human Research Ethics Committee (H14/07-163). The specific hypotheses for this study were not pre-registered.

The intervention: TaylorActive

The content of the intervention modules was tailored based on participants’ responses to self-reported assessments integrated throughout the sessions that included questions about individual, social and environmental determinants of physical activity. Depending on the responses, personally-relevant advice would be presented based on self-determination theory (Ryan & Deci, 2000), social cognitive theory (Bandura, 2001) and theory of planned behaviour (Ajzen, 1991). The difference between the intervention groups was that the personally-tailored session content was either delivered through videos or text. Participants could only access the next session of the intervention on completion of the previous one. The action planning aspect of the intervention was included at the end of sessions 2–7 (i.e. were not in the first or last session). Participants were asked to make action plans about how they would meet their physical activity goals (i.e. what, where, when and with whom).

The primary aim of the intervention was to increase physical activity. Outcomes of the trial were that no significant differences in physical activity behaviour were found between groups at either 3- or 9-month assessments and attrition was high (Vandelanotte et al., 2021). The following sections describe methods relevant to the present study with focus on the impact of action planning on physical activity behaviour, intention and habit strength.

Participants

Eligible participants for the TaylorActive intervention were adults with access to broadband internet, who spoke and read the English language, lived in Australia, were not currently engaging in 150 min of moderate-to-vigorous physical activity over five or more days of the week, as per the recommended national physical activity guidelines (Commonwealth Department of Health, 2014), and could safely increase their physical activity (Cardinal et al., 1996). Pregnant women, those with a body mass index less than 17.5 and people with self-reported impairment that prevented them from increasing physical activity safely were excluded from participating in the trial.

For this secondary data analysis, only participants in the intervention groups who provided data for the nine-month assessment were included in the analyses, resulting in a N of 115 with 68.7% female (n = 79) and 31.3% male (n = 36), with a M age of 43.9 (SD = 11.65; range = 22–73). Most (50%) participants were between the ages of 35 and 55. Most participants were married (52.7%), with some single and had never been married (24.8%), others in de facto relationships (11.7%), some divorced (6.3%), and a few separated but not divorced (2.4%), or widowed (2.2%). The sample mostly identified as Caucasian (89.6%), with a small proportion identifying as Aboriginal, Torres Strait Islander or Pacific Islander (1.3%), Asian (1.7%), African (0.4%), or other (6.9%). Just under half of participants reported being employed full-time (48.6%), with 18.8% employed part-time, 11.0% employed casually, 7.6% retired, 5.6% student, 3.3% unemployed, 2.6% home duties, 2.4% pension/carer and 0.2% did not respond. On average, participants reported completing M = 17.28 (SD = 2.00) years of school and worked M = 7.66 (SD = 1.91) hours during workdays.

Of the data used for this study, 55.7% (n = 64) participants were in the text-tailored condition, and the other 44.3% (n = 51) were from the video-tailored condition. Compared to the whole intervention study sample, these participants were not statistically different on baseline measures of gender (p = .65), employment status (p = .28), ethnicity (p = .39), marital status (p = .23), years of schooling (p = .32), working day hours (p = .70), physical activity behaviour (p = .12), intention strength (p = .80), or habit strength (p = .42), but the sample used for this study was slightly younger than the overall intervention study sample (Δ M = 0.43, SE = 1.41, p = .02). Participants who provided follow-up data at the 9-month assessment completed more action plans than those who did not provide 9-month follow-up data (Δ M = 1.89, SE = 0.22, p < .01).

Measures

Physical activity behaviour

Time spent in moderate-to-vigorous physical activity as average minutes per day was assessed by validated ActiGraph GT3X + activity monitors worn on the hip during all waking hours over seven consecutive days at each assessment (Sasaki et al., 2011). Triaxial acceleration data were sampled at a frequency 30 Hz and then accumulated into 1-s and 60-s epochs using Actilife software (V.6.13.1). Valid wear time was defined as more than 10 h on 5 days within the 7-day period (Troiano et al., 2014). Non-wear time was defined as 90 consecutive minutes of 0 counts per minute, allowing for a 2-min interruption (Choi et al., 2012). Moderate-to-vigorous physical activity was set as greater than or equal to the vector magnitude of 2690 counts per min (Sasaki et al., 2011).

Physical activity intention strength

Intention strength was assessed as a single item in which participants responded to the question, ‘To what extent do you agree or disagree with this statement: I intend to engage in physical activity at least every other day, over the next three months?’ with response options ranging from 1 (Strongly Disagree) to 7 (Strongly Agree). This measure is in line with Rhodes and Rebar (2017)’s conceptualisation of intention strength as the intensity of determination to engage in physical activity.

Physical activity habit strength

Physical activity habit strength was assessed using the 4-item Self-Reported Behavioural Automaticity Index (Gardner et al., 2012) of the Self-Reported Habit Index (Verplanken & Orbell, 2003). Participants were asked to what extent they agreed with the items, ‘Physical activity is something I do… (1) automatically, (2) without having to consciously remember, (3) without thinking, (4) before I realise I’m doing it’, using response scales from 1 (Strongly Disagree) to 7 (Strongly Agree). The habit strength scale was calculated as the mean of the 4 responses with interitem reliability from the present study at α = .95 (baseline), α = .92 (3-month assessment) and α = .91 (9-month assessment).

Action plans

During the completion of each module in the intervention participants used open-ended response boxes to enter what activity they would do, where and when they would do it, how often, for how long and with whom (Vandelanotte et al., 2015). Once participants completed an action plan for their first activity (e.g. walking), they were asked if they would like a plan for additional activities (e.g. swimming). If so, they were given the same set of prompts to complete action plans for up to 4 physical activities per session. As such, they could action plan for a total of 24 activities across 6 sessions, spread across 3 months. Action plan frequency was calculated as the sum of the completed action plans for all activities, with a possible range of 0–24.

For consistency and content variables, action plans were coded by RW and ALR into categories of related activities, locations, time of day, days of week and social surroundings. For each action plan category, consistency was calculated as the reversed entropy scores, with more positive scores interpretable as more consistent action plans. Reverse entropy scores provide consistency of categorical responses as per Maher et al. (2021). Further details of the entropy calculation are included in the data management and analyses section. Consistency was calculated as the mean of the reverse entropy scores across action plan categories. Individuals’ content variables were coded as their most frequent action planning activity, location, time of day, days of week and social support. In the case of a tie, the variable that was reported in the earliest reported action plan of the most frequent was used.

Data management and analyses

All data were analysed in R 4.2.1 (R Core Team, 2022). Entropy was calculated using shrinkage entropy estimator of the entropy R function with the robust maximum likelihood estimator and James-Stein-type skrinkage intensity (Hausser et al., 2021; Hausser & Strimmer, 2009). Given that there was minimal level 2 data nesting present (ICC < 5%) and high risk of multicollinearity between 3- and 9-month assessments, the hypotheses were tested with separate linear regression models rather than generalised linear models. The pooled estimates of multivariate linear regression models from 5 datasets simulated with multiple imputation with chained equations (van Buuren & Groothuis-Oudshoorn, 2011) was used to test whether intention and habit strength was associated with physical activity behaviour at the 3-month and 9-month follow-ups, controlling for the covariates of baseline physical activity behaviour, intervention group, age and gender. The same approach was used to test whether action planning frequency and consistency led to short-term and long-term changes in physical activity behaviour, intention and habit strength, controlling for the same covariates. Model assumptions were tested with no violations present. To test the effects of action planning content on change in physical activity behaviour, intention and habit strength, linear regression models were estimated with estimated marginal means using the emmeans R package with Bonferroni significance adjustments for multiple comparisons (Lenth, 2020).

Results

Descriptive statistics

Descriptive statistics and bivariate correlations of physical activity behaviour, intention strength, habit strength and action planning frequency and consistency are shown in Table 1. Physical activity behaviour at the 3-month follow-up was positively associated with concurrent habit strength but not intention strength and was not significantly associated with either intention or habit strength at the 9-month follow-up. At the 9-month follow-up, physical activity behaviour, intention and habit strength were positively associated with one another. How many action plans people made across the 3-month intervention (i.e. action planning frequency) was positively associated with habit strength at the 9-month follow-up but not with any other study variable. How consistent people’s action plans were was positively associated with intention strength at all time assessments, habit strength at the 3-month and 9-month follow-up, and physical activity behaviour at the 3-month and 9-month follow-ups.

Table 1. Descriptive statistics and bivariate correlations (r) of physical activity behaviour, habit strength and action planning.

	M (SD)	Baseline			3-Month follow-up			9-Month follow-up			Action planning
		1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.
Baseline
1. Physical activity (min/day)	16.38 (16.86)	–	.09	.05	.46*	.09	.25*	.62*	.11	.12	.03	.10
2. Intention strength	6.05 (1.30)		–	−0.01	.08	.14	.02	−0.02	.33*	.10	.07	.27*
3. Habit strength	3.38 (1.75)			–	.07	−0.02	.48*	.21*	.14	.58*	.06	.11
3-Month assessment
4. Physical activity (min/day)	18.50 (14.33)				–	.17	.22*	.72*	.15	.14	.12	.18*
5. Intention strength	5.90 (1.50)					–	.27*	.11	.37	.20*	.05	.63*
6. Habit strength	4.06 (1.61)						–	.27*	.32*	.27*	.17	.39*
9-Month assessment
7. Physical activity (min/day)	17.49 (14.70)							–	.21*	.24*	.11	.21*
8. Intention strength	5.43 (1.78)								–	.34*	.04	.69*
9. Habit strength	4.05 (1.74)									–	.19*	.37*
Action planning
10. Action planning frequency	3.87 (2.15)										–	.61*
11. Action planning consistency	2.62 (0.68)											–

Note. *p < .05.

The content of all 710 action plan categories and their frequencies and percentages are shown in Supplemental Table 1. The most common reported action planning activity was walking, and the most common location in action planning was outdoors. People tended to make action plans to engage in physical activity during a general time of day (e.g. morning, afternoon, after work) vs a specific time (e.g. 7:00 am). People were more likely to make action plans for physical activity on weekdays compared to weekend days. People reported action plans to engage in physical activity by themselves the most often, compared to with others.

Intention and habit strength predicting change in physical activity behaviour

The results of the models testing whether intention and habit strength at baseline predicted changes in physical activity at the 3-month and 9-month follow-up assessments are shown in Table 2. It was hypothesised that both intention strength and habit strength would be positively associated with both prospective physical activity behavioural outcomes at both 3-month and 9-month follow-ups, neither intention strength nor habit strength significantly predicted change in physical activity behaviour at 3 months (directly following conclusion of the intervention). Baseline physical activity was the only significant predictor of 3-month follow-up physical activity behaviour. For physical activity behaviour at the 9-month follow-up (6 months following conclusion of the intervention), habit strength but not intention strength significantly predicted change in physical activity behaviour. The models with intention and habit strength predicting physical activity behaviour explained 19% and 41% of variability for the 3-month and 9-month follow-ups, respectively.

Table 2. Results of multiple imputation with chained equations pooled results of regression models testing whether intention and habit strength predicted physical activity behaviour at 3-month and 9-month follow-ups.

Term	3-Month follow-up		9-Month follow-up
	b	95% CI	b	95% CI
Intercept	3.07	−14.47 to 20.61	7.50	−9.84 to 24.84
Baseline physical activity behaviour	0.36*	0.20 to 0.52	0.51*	0.37 to 0.65
Intention strength	0.92	−0.90 to 2.74	1.07	0.93 to 1.21
Habit strength	0.61	−1.19 to 2.41	1.33*	0.01 to 2.65
Intervention group	−1.20	−6.12 to 3.72	−0.13	−4.97 to 4.71
Age	0.02	−0.20 to 0.24	−0.19	−0.41 to 0.03
Gender	0.70	−5.12 to 6.52	−0.54	−5.98 to 4.90

Note. *p < .05; Adj. R²: 3-Month Follow-Up Model = 0.19*; 9-Month Follow-Up Model = 0.41*.

Action planning frequency, consistency and content

Table 3 shows the results of the models testing whether action planning frequency and consistency predicted change in physical activity behaviour, intention strength and habit strength at the 3-month and 9-month follow-ups. It was hypothesised that more frequent action planning would lead to more engagement in physical activity behaviour and stronger physical activity intention and habit strength. Additionally, it was hypothesised that more consistent action planning would lead to more engagement in physical activity behaviour and stronger physical activity habit strength, but not intention strength. The results revealed that neither action planning frequency nor action planning consistency predicted short-term change in physical activity at the 3-month follow-up assessment (immediately following completion of intervention). At the 9-month follow-up assessment (6-months following completion of intervention), however, people with more consistent action plans had a more positive increase in physical activity behaviour than those with less consistent action plans. There was no significant impact of how many action plans people made on long-term physical activity. The models predicted 19% and 40% of change in physical activity behaviour at the 3-month and 9-month follow-ups, respectively.

Table 3. Results of multiple imputation with chained equations pooled results of regression models testing whether action planning frequency and consistency predicted physical activity behaviour, intention strength and habit strength at 3-Month and 9-Month Follow-Ups.

Predicting physical activity behaviour	3-Month follow-up		9-Month follow-up
	b	95% CI	b	95% CI
Intercept	4.56	−12.40 to 21.52	8.85	−9.13 to 26.83
Baseline physical activity behaviour	0.38*	0.22 to 0.54	0.53*	0.39 to 0.67
Action planning frequency	0.29	−1.21 to 1.79	−0.46	−1.78 to 0.86
Action planning consistency	2.14	−2.46 to 6.74	4.55*	0.11 to 8.99
Intervention group	−1.43	−6.37 to 3.51	−0.16	−5.06 to 4.74
Age	0.00	−0.22 to 0.22	−0.20	−0.44 to 0.04
Gender	0.86	−4.84 to 6.56	−0.53	−6.07 to 5.01
Predicting intention strength	3-Month Follow-Up		9-Month Follow-Up
Intercept	1.91*	0.57 to 3.25	−0.52	−1.86 to 0.82
Baseline intention strength	−0.13	−0.27 to 0.01	0.13	−0.01 to 0.27
Action planning frequency	−0.37*	−0.47 to −0.27	−0.49*	−0.59 to −0.39
Action planning consistency	2.14*	1.80 to 2.48	2.68*	2.32 to 3.04
Intervention group	−0.16	−0.52 to 0.20	0.14	−0.22 to 0.50
Age	0.00	−0.02 to 0.02	0.00	−0.02 to 0.02
Gender	0.37	−0.03 to 0.77	−0.10	−0.50 to 0.30
Predicting habit strength	3-Month Follow-Up		9-Month Follow-Up
Intercept	0.69	−1.09 to 2.47	0.34	−1.44 to 2.12
Baseline habit strength	0.41*	0.27 to 0.55	0.55*	0.41 to 0.69
Action planning frequency	−0.07	−0.21 to 0.07	−0.02 (0.07)	−0.16 to 0.12
Action planning consistency	0.93*	0.45 to 1.41	0.85*	0.37 to 1.33
Intervention group	−0.13	−0.65 to 0.39	−0.13	−0.63 to 0.37
Age	0.00	−0.02 to 0.02	−0.01	−0.03 to 0.01
Gender	−0.18	−0.74 to 0.38	0.17	−0.39 to 0.73

Note. *p < .05; Physical activity behaviour: Adj. R²: 3-Month Follow-Up Model = 0.19*; 9-Month Follow-Up Model = 0.40* Predicting intention strength: Adj. R²: 3-Month Follow-Up Model = 0.58*; 9-Month Follow-Up Model = 0.70*; Predicting habit strength: Adj. R²: 3-Month Follow-Up Model = 0.32*; 9-Month Follow-Up Model = 0.41*.

The models predicting intention strength revealed that people who made more action plans reported weaker intentions to engage in regular physical activity than those who made fewer action plans. Additionally, people who made more consistent action plans reported stronger intentions to engage in regular physical activity than those who made less consistent action plans. The models predicted 58% and 70% of change in intention strength at the 3-month and 9-month follow-ups, respectively. The models predicting habit strength revealed that action planning frequency was not statistically significantly associated with habit strength; however, those who made more consistent action plans reported stronger growth in physical activity habit strength at both the 3-month and 9-month follow-ups compared to those who made less consistent action plans. The models predicted 32% and 41% of habit strength at the 3-month and 9-month follow-ups, respectively.

Table 4 shows the results of the models testing whether the content of action planning impacted physical activity behaviour, intention or habit strength at both the 3-month and 9-month follow-ups. Given the lack of existing evidence on content of action planning, no hypotheses were formed regarding whether certain activities, places, times of day, or certain surroundings would translate into more or less behaviour change, intention, or habit strength. The only significant effect was that people who made action plans of engaging in fitness classes had a larger long-term increase in physical activity behaviour than those who made action plans for other activities. Other than that, no significant effects were found.

Table 4. Results of linear regression with estimated marginal means to test whether action planning content predicted change in physical activity behaviour change at 3-month and 9-month follow-ups.

Term	3-Month follow-up			9-Month follow-up
	Physical activity behaviour	Intention strength	Habit strength	Physical activity behaviour	Intention strength	Habit strength
	EMM (SE)	EMM (SE)	EMM (SE)	EMM (SE)	EMM (SE)	EMM (SE)
Activity content	−4.91 (4.29)	−0.45 (0.43)	−0.43 (0.40)	13.25 (4.31)*Fitness classes	−0.59 (0.48)	−0.71 (0.40)
Location content	−2.99 (3.81)	−0.37 (0.37)	−0.21 (0.35)	−3.69 (3.60)	−0.51 (0.41)	−0.55 (0.35)
Time content	−1.62 (0.82)	−0.35 (0.32)	−0.31 (0.31)	−2.52 (3.20)	−0.35 (0.37)	−0.28 (0.31)
Day content	−4.74 (0.41)	−0.23 (0.38)	−0.47 (0.36)	−5.30 (3.71)	−0.64 (0.43)	−0.66 (0.37)
Social content	−3.01 (4.45)	−0.20 (0.42)	−0.28 (0.39)	−3.93 (4.21)	−0.58 (0.49)	−0.61 (0.41)

Note. *p < .05; Adj. R²: 0.00 to 0.42*; EMM: Estimated marginal means.

Post-hoc exploratory analyses

To test effects of action planning frequency and consistency on physical activity behaviour were present above and beyond those of intention and habit strength, additional post hoc models were tested with physical activity at the 3-month and 9-month follow-ups regressed onto action planning frequency, consistency, intention strength and habit strength (at baseline as well as the respective follow-up assessment), controlling for the covariates of baseline physical activity behaviour, intervention group, age and gender. The results of these models are shown in Table 5. When all predictors were included in the model, no variable significantly predicted physical activity beyond baseline habit strength predicting 9-month physical activity behaviour, suggesting the association of action planning consistency on prospective physical activity behaviour was not unique to, but rather overlapped with, that of intention or habit strength.

Table 5. Results of post hoc results of regression models testing whether action planning frequency and consistency were associated with physical activity behaviour above and beyond intention and habit strength at the 3-month and 9-month follow-ups controlling for baseline physical activity behaviour, intervention group, age and gender.

Predicting physical activity behaviour	3-Month follow-up		9-Month follow-up
	b	95% CI	b	95% CI
Intercept	−6.97	−30.61 to 16.66	3.26	−18.42 to 24.94
Baseline physical activity behaviour	0.36*	0.20 to 0.53	0.52*	0.38 to 0.67
Baseline intention strength	0.76	−1.45 to 2.97	−0.00	−1.98 to 1.98
Follow-up intention strength	0.77	−1.19 to 2.73	0.65	−1.14 to 2.44
Baseline habit strength	0.45	−1.29 to 2.19	2.03*	0.48 to 3.57
Follow-up habit strength	0.68	−1.33 to 2.70	−0.03	−1.84 to 1.78
Action planning frequency	0.48	−0.89 to 1.84	0.73	−0.52 to 1.98
Action planning consistency	−6.57	−16.90 to 3.76	0.63	−8.06 to 9.33
Intervention group	−1.84	−7.30 to 3.62	−0.79	−5.65 to 4.08
Age	−0.08	−0.32 to 0.17	−0.24*	−0.45 to −0.02
Gender	0.92	−5.53 to 7.38	1.86	−4.27 to 7.98

Note. *p < .05; Adj. R²: 3-Month Follow-Up Model = 0.19*; 9-Month Follow-Up Model = 0.43*.

Discussion

The aim of this study was to evaluate whether the frequency, consistency, or content of action planning impacted change in physical activity behaviour, intention and habit strength, both immediately following a physical activity intervention (short-term change) and 6-months following the end of the intervention (long-term change). Overall, the results revealed the consistency of action planning rather than the frequency or content of action planning predicted intention, habit strength and engagement in physical activity both immediately and 6-months following a web-based physical activity intervention. Post-hoc analyses results revealed the effect of action planning consistency was not independent of change in intention and habit strength, suggesting these may be the mechanisms through which consistent action planning changed physical activity behaviour.

Intention, habit strength and physical activity behaviour change

As is commonly found when predicting future physical activity behaviour (e.g. Hagger et al., 2002), there was a strong effect of past physical activity behaviour on prospective physical activity behaviour, such that people who were more active prior to the intervention tended also to be the more active at the end of the intervention and 6-months following its completion. After accounting for this effect of past behaviour, change in physical activity immediately post-intervention was not associated with intention or habit strength, but change in physical activity 6 months following the intervention was predicted by change in habit strength for physical activity, but not intention strength. Notably, the commonly theorised habit moderation effect on intention—behaviour associations (Gardner et al., 2011; Rebar et al., 2016) was not present in this study.

It was unexpected that physical activity behaviour and intention strength were not strongly associated, concurrently or prospectively, as is typical amongst the literature (McEachan et al., 2011; Rhodes & de Bruijn, 2013). Most of what is known about intention-behaviour processes in physical activity research is from observational studies, so it may be that the intervention interfered with typical motivation-behavioural processes. For example, people could have been particularly motivated with enrolment in the intervention, despite not currently engaging in high levels of activity, leading to a temporary misalignment of intention to behaviour. More work is needed with alternative assessments of intention strength before conclusions can be made about the role of intentions in digital interventions such as the one in the present study.

Habit strength was associated with change in physical activity behaviour 6 months following the intervention but not immediately following the intervention, which aligns with the theoretical notion that habits can underpin behaviour change maintenance (Kwasnicka et al., 2016; Rebar et al., 2018). Throughout the intervention, people set activity goals, engaged in self-monitoring, and were trained in various self-regulation strategies to increase physical activity (Vandelanotte et al., 2015), but for the 6-months following the intervention, it seems physical activity behaviour was more reliant on people’s habits. These findings suggest that for longer lasting behaviour change, physical activity interventions should integrate strategies to encourage habit formation (Gardner & Rebar, 2019). The present study provides further insight into whether the intervention technique of action planning can be used in this regard.

Action planning frequency

Although given ample opportunity across regular intervals, people in this intervention chose to not make many action plans. At most, participants completed 6 of the possible 24 action plans, with most participants completing fewer than 5. Unexpectedly, there was no evidence that those who made more action plans had a larger increase in physical activity behaviour or stronger habit formation for physical activity than those who made less. In fact, those who made many action plans tended to have resultant decreases in physical activity intention strength compared to those who made fewer action plans. Contrary to speculations (e.g. Hagger & Luszczynska, 2014), evidence from this study suggests that encouraging people to make many action plans is not an effective approach to intervene with physical activity behaviour, and instead may have detrimental effects on motivation. Whereas generally, the more people engage with online intervention content, the more effective the behaviour change (Smith & Liu, 2020; Vandelanotte et al., 2007), it seems that action planning does not have a dose-response effect.

It should not be ruled out, however, that high frequency of action planning represents purposeful shifts in planning, which may be a byproduct of shifts in intention strength. For example, it may be that people make initial plans that do not work when executed in real-life, so then they adjust their plans to find something that better suits them. If this were the case, it would indeed be reasonable that someone makes fewer action plans as their intentions become stronger. Given the importance of having plans for activity that fit well within day-to-day life for long-term maintenance (Gardner & Rebar, 2019), it may be that interventions could consider a ‘trial and error’ period to first find effective plans and then consider focusing on consistency of the most effective plans.

Action planning consistency

Rather than encouraging people to make many different action plans, evidence from the present study suggests it is more effective to encourage people to make the same or similar action plans repeatedly. The consistency with which action planning was made predicted physical activity behaviour change 6 months following the intervention, but not immediate increases in physical activity behaviour. Additionally, action planning consistency predicted increases in intention and habit strength for physical activity at both short- and long-term follow-ups. Post hoc analyses revealed the effects of action planning consistency on change in long-term physical activity was not independent of, but rather overlapped with, change in intention and habit strength. Although further work is needed to replicate these findings, these findings suggest the mechanisms through which action planning consistency changes behaviour is through enhancement of intention and habit strength.

These findings align with habit theory’s proposition that behavioural maintenance is facilitated with engaging in behaviour regularly within stable contexts (Carden & Wood, 2018; Gardner et al., 2022; Rhodes & Rebar, 2018). Although unexpected a priori, the finding that action planning consistency was aligned with growth in intention strength for physical activity is sensible, given past evidence that the stability of planning and intentions is an important indicator of the quality of intentions for behaviour change (Conner & Godin, 2007; Sheeran & Abraham, 2003). In an important advancement for the field, this study provides the first evidence that encouraging people to make consistent action plans over time can increase motivation, habit and long-term behaviour change for physical activity.

Action planning content

When prompted to make action plans, people tended to make plans for the activity of walking, within the location of an outdoor setting, during a general time of day such as ‘morning’ or ‘after work’, on weekdays vs weekend days, and by themselves as opposed to with others. The content of action planning did not have impact on changes in intention or habit strength; however, those who made action plans for fitness classes tended to have a larger increase in physical activity behaviour at the 9-month follow-up than those who made action plans for other types of activities. Notably, this effect was not present immediately following the intervention, but rather was present 6 months following the intervention. It may be that the social support, routineness, or accountability of fitness classes lends itself to a longer-term behavioural maintenance than other types of activities (Burke et al., 2006).

That there were no differences in behaviour change for those who planned to do physical activity by themselves compared to with others is somewhat surprising, given the body of evidence on the benefits of social support and connectedness for physical activity motivation (Scarapicchia et al., 2017). However, social support can take many forms and evidence suggests having added support from a person sharing in the planning of the activity could have benefits (Burkert et al., 2011; Hagger & Luszczynska, 2014; Keller et al., 2017). Future work should consider whether social support is beneficial for making and/or implementing planned physical activity. The null finding regarding the importance of time of day for implementing planned physical activity behaviour is also counter to some evidence suggesting people are more likely to engage in the structured planned activity during morning compared to later in the day (Domke et al., 2019). Although given the idiosyncrasy of schedules, chronotypes and preferences, it is likely that the impact of the content of plans is not as straight-forward as an overall group-level effect.

Study strengths and limitations

Overall, these findings provide important insight into how action planning may be utilised to effectively increase physical activity long-term. However, these findings must be considered in the context of the physical activity intervention the participants were engaging in, of which action planning was one element (Vandelanotte et al., 2015). Although, the intervention did not show significant changes in behaviour (Vandelanotte et al., 2021), it is likely that the other intervention elements impacted motivation and behaviours. Future research is needed to consider whether these findings replicate if action planning is the only intervention component. Additionally, the sample participants used for these analyses were younger than that of the overall intervention sample, so further work is needed to test whether these findings would differ amongst samples with a wider age range. Given that those included in the analyses completed more action plans than those not included as a result of attrition, more is needed to rule out the potential for selection bias to impact these findings.

Notably, within the constraints of this study, the effects of action planning content could only be looked at across immediately following intervention and 6 months later. Although this is a strength of the study to consider both short- and long-term change in motivation and behaviour, it means that we cannot establish if there were any immediate, acute effects of action planning on behaviour on a more micro time scale. Potential moderating factors should also be considered. For example, evidence suggests less specific goals may be more beneficial for people starting to be physically active from inactivity (Swann et al., 2023), so it is important the impact of action planning consistency is tested in a variety of populations. Evidence also suggests that when dyads make plans together, it can enhance their likelihood of success (Keller et al., 2017), so how and with whom plans are made are important to consider as well. Such future research should also consider the important part of the action planning process that was not considered in this study—that of plan enactment (de Vries et al., 2013; Sniehotta, 2009). Whether people engaged in the physical activity in the contexts specified by their action plans has been shown to be important for habit formation and long-term behaviour change maintenance and should be a focus in future intervention approaches and assessments (Fleig et al., 2017). Importantly, the focus of this study was action planning but there are many other related forms of self-regulation such as goals, coping plans and a more specific form of action planning—implementation intentions, that may also have relevancy for habit formation and long-term behavioural maintenance (Carraro & Gaudreau, 2013; Gollwitzer, 1999; Swann et al., 2023). Given their prominence as intervention approaches, it is essential research continues to help establish how to utilise goals and plans in the most effective ways for the betterment of people’s health and wellbeing.

Conclusions

Overall, the findings of this study suggest that for the most impact on habit strength and physical activity behaviour change, people should be encouraged to make consistent action plans over time. The findings of the present study provide an important insight into the potential impact that consistency of action planning can have on motivation and behaviour change that should lead to future research including more various timing of assessments, experimental manipulation of consistency of action planning, and more heterogenous samples to establish if these findings generalise beyond the constraints of the present study.

Disclosure statement

The authors report there are no competing interests to declare.

Data availability statement

Data are available on reasonable request. The dataset supporting the conclusions of this article will not be shared at present as it is still being used for analysis of other outcomes of the TaylorActive randomised controlled trial. However, any reasonable request to access the data will be considered and must be made to Corneel Vandelanotte ([email protected]).

Additional information

Funding

The work was supported by National Health and Medical Research Council and National Heart Foundation of Australia.