Background
Regular physical activity (PA) among adults has an enormous number of health benefits [
1], but very low participation rates (e.g., [
2]). Consequently, PA promotion initiatives are of high importance to public health. Regular walking is one of the most preferred PAs and thus a key target for intervention [
3]. One correlate of regular walking that has seen considerable attention in public health is dog ownership. Specifically, dog owners report more walking during leisure-time than non-owners [
4‐
6]. While this is an interesting descriptive finding, its direct application to PA promotion is less practical, as dog ownership is an enormous responsibility with cost implications. Still, approximately 30% of the population in developed countries own dogs [
7], and it is estimated that half of all dog owners do not walk their dogs regularly [
4]. While dog walking is merely one of many types of PAs that dog owners could potentially enact, it seems a logical way to engender both human and canine health benefits simultaneously among this large potential target population [
8,
9]. Understanding the correlates of walking would thus help identify key intervention targets to promote owners to walk their dogs more.
A recent review of 31 studies on the correlates of walking among dog owners found that an attachment to the dog in the form of responsibility/obligation/support and environmental access to suitable walking areas with dog supportive features (e.g., off-leash exercise) were reliable factors [
10]. More recently, Richards et al. [
11] have shown that social cognitive theory constructs of dog outcome expectations, social support and measures of the walking environment were key predictors of regular walking. Other studies have also shown some evidence that the theory of planned behavior can predict regular dog walking [
12,
13]. Still, the Westgarth et al. review noted that only a handful of the studies used a theoretical framework to explain walking and this is a noteworthy limitation.
An additional potential limitation to prior work is the positioning of intention as the proximal determinant of walking behavior. Indeed, even Westgarth and colleagues [
10] suggested that all correlates of dog walking should affect behavior through intention. This positioning makes the assumption that once a dog owner has positive intentions to walk, it will be sufficient to enact the behavior. Still, there is only modest support for the relationship between intention and behavior in general PA research [
14]. Interestingly, the relationship between intention and behavior also shows that almost all discordance occurs from those who intend but fail to perform the behavior, and not from those with low intentions who enact the behavior [
14]. The sizeable proportion of intenders who subsequently fail to follow-through and enact behavior has prompted the term ‘intention-behavior gap’, because approaches that feature intention as the proximal determinant of behavior have limited theoretical explanation for this finding [
15]. It seems a worthy line of inquiry to examine the intention-behavior gap in dog owners’ dog walking behavior and this has not been formally explored at present.
Several models have attempted to understand the translation of intention into behavior, also known as action control [
16]. One of the most frequently applied of these in the PA domain is the multi-process action control framework (M-PAC; [
16‐
19]). In this framework, intention (i.e., intend/do not intend) and behavior are divided into quadrants, which creates four possible profiles, but only three of substantive value:
non-intenders who are subsequently not active;
successful intenders who are subsequently active, and
unsuccessful intenders who failed to enact their positive intentions. By contrast, the fourth profile of
disinclined actors, who despite lack of intention are subsequently active, is hypothesized as empty because intention is viewed as a necessary but insufficient process to achieve PA within M-PAC [
16,
17].
The intention-behavior profiles in M-PAC have similarities to the stages of change in the transtheoretical model [
20] as intention-behavior hybrid constructs, but M-PAC profiles are not stages, as someone can move from non-intender to successful intender without ever falling into the “unsuccessful intender” profile. M-PAC suggests that intenders may be predicted by motivational processes of instrumental attitude/outcome expectations (utility of the behavior), affective judgments (enjoyment of the behavior) and perceived control (ability and opportunity to perform the behavior) which supports the tenets of most intention-based theories [
21]. In M-PAC, however, affective judgments and opportunity are also considered predictors during the intention-formation to action control process, where higher values are considered necessary for successful translation of intentions into behavior than for intention formation. Furthermore, action control is thought to be dependent on regulation behaviors (e.g., planning, self-monitoring), as people begin to use volitional tactics to help translate positive intentions into action. Regulation behaviors are conceptually similar to action/coping planning in the health action process approach [
22] or the behavioral processes of change in the transtheoretical model [
20].
Continuance of action control is thought to also add reflexive processes such as automaticity/habit (i.e., behavior performed from stimulus–response bonds) and identity (self-categorisation) as one begins to perform the behavior more regularly. Specifically, as a behavior has become more routine, Rhodes and de Bruijn [
17] suggest that intention-driven behavior is executed partially from environmental cues [
23], selective processing of information congruent with one’s self-categorization [
24] and the dissonance that arises from any discrepancy between self-categorization and behavior [
25]. To date, M-PAC has not been applied to understand dog walking behavior but it may show utility, given the likelihood of the intention-behavior gap. Furthermore, the regulatory and reflexive processes in M-PAC offer constructs that have not yet been examined within this domain. As much of this walking behavior has the potential to be ritualized and routine, habit also seems like a worthy concept in this domain. Further, as dog-owner attachment constructs have prior validation with owner dog walking [
10], an exploration of dog-walking identity in action control may be important.
Thus, the purpose of this paper was to apply the M-PAC framework in a sample of dog owners to understand the translation of daily walking intention and behavior. It was hypothesized that participants would group into three (i.e., non-intenders, unsuccessful intenders, successful intenders) of the four possible intention-behavior profiles based on prior research in general physical activity contexts [
14]. It was further hypothesized that affective judgments (enjoyment of owner dog walking), opportunity (availability of time and environment to walk), regulation behaviors (detailed plans to dog walk) and reflexive processes of habit (learned responses to walking cues) and identity (personal standards of dog walking behavior) would explain successful, compared to unsuccessful, intenders based on prior research with this model [
17].
Discussion
This study was the first to examine the intention-behavior gap in daily walking behavior among dog owners and predict this gap using the M-PAC framework designed for this purpose [
16,
17]. We hypothesized that three of four possible intention-behavior profiles would emerge (i.e., non-intenders, unsuccessful intenders, successful intenders), commensurate with prior research in general PA [
14]. This hypothesis was supported. Only two participants in the sample were classified as having low intentions and engaged in daily dog walking. By contrast, 73% of the sample was comprised of intenders, yet 45% of these intenders were not walking congruent with their intentions and most of these participants walked <50% below what they intended.
The finding has both theoretical and applied implications. From a theoretical perspective, the results demonstrate that intention formation is a necessary process but it may be insufficient for walking enactment among many dog owners. Thus, frameworks that propose intention as the bridge to behavior, which comprise many of our most popular health behavior models and contemporary models for dog walking [
10], may not be as useful as models that separate intention translation from intention formation [
16]. From an applied perspective, these results also help explain that some dog owners have yet to form daily walking intentions, while even more participants intend to walk their dog daily but fail to follow-through. Thus, dog walking promotion may benefit from both intention formation and action control interventions, depending on the readiness of the population.
In light of this aim, the second purpose of the study was to predict these intention-behavior profiles using the M-PAC framework. The findings support almost all of the tenets of that model. M-PAC suggests that instrumental outcome expectations about walking may not be as important to action control because they do not reflect the experience of the action itself. Perceptions of capability are also generalized to the act (i.e., am I physically able to walk or not) and not specific to each action but could be important to dog owners given the additional demands of controlling dog behavior. Congruent with theory, outcome expectations and perceived capability to walk did not contribute to the intention-behavior profiles while controlling for other M-PAC variables. Thus, educational/informational campaigns based on the benefits of walking and interventions to improve one’s ability to walk are not recommended as standalone interventions for closing the intention-behavior gap, even though these constructs have been shown to be general correlates of owner dog walking [
11].
By contrast, affective judgements did contribute to the discriminant function, predicting all three intention-behavior profiles. While affective judgments have been shown to predict owner intentions to dog walk in prior research [
13], this finding supports the approach taken in M-PAC [
17], where higher affect is also needed to enact a behavior than form the intention. From a theoretical perspective, the importance of affective over instrumental outcome expectations during action control supports hedonic theories of behavior. The practical aspect of this finding suggests that consideration of pleasure in walking interventions may facilitate closing the intention-behavior gap. M-PAC constructs are considered the consequence of individual, social, and environmental/policy factors [
46]. For affective judgments, this could involve several factors that require future research, such as pleasant walking conditions (e.g., environmental design and dog-friendly amenities), social aspects of walking (e.g., walks with friends/ other dog owners), and dog-related variables that could affect the walking experience (e.g., level of training and responsiveness, dog’s sociability towards other dogs and people, owner enjoyment related to bonding with their dog).
The hallmark of most action control models, including M-PAC, is the premise that volitional self-regulation tactics are needed to tie good intentions to behavior [
16]. In support of this hypothesis, behavioral regulation correlated with the discriminant function and predicted all three intention-behavior profiles. The inclusion of behavioral regulation variables in understanding general PA has had considerable support [
47] but this is the first study in dog owners to apply this variable to our knowledge. The results suggest that having dog owners make action (when, where, how, and with whom) and coping (details about how to overcome potential set-backs) plans and subsequently track these plans (e.g., with mobile phone apps, diaries) may be very useful to close the intention-behavior gap.
The reflexive M-PAC constructs of habit/automaticity and identity also contributed to the discriminant function. Habit predicted all three intention-behavior profiles while identity was considerably higher for successful intenders compared to unsuccessful intenders. These are new constructs to the dog owner and walking correlates literature, and represents potentially important considerations for future intervention. Both constructs are considered for the maintenance of action control, as performance experience with the behavior is necessary for their formation. From a theoretical standpoint, habits are thought to be formed from consistent repetitions of action and exposure to similar cues [
23], which highlight the importance of environmental and social context. Preliminary research suggest that the key to habit formation is consistency of practice [
48]. Dog owners with a regular walk routine may thus be more likely to acquire automaticity over time and facilitate action control partially independent of motivation. Given the long period of owning a dog, this may be a critical factor to behavior maintenance over time.
Our measure of dog walking identity in this study may overlap with prior research on dog support/obligation (e.g., [
12,
13,
49]). Identity is a self-categorization of oneself into a particular role [
25] and a sense of obligation and responsibility for one’s dog would seemingly be a part of a dog walking identity. Our findings with identity and action control support this prior work. Qualitative interviews with dog owners highlight how dog walking is a duty or a role described with similarities to parenting children [
50]. Further, Brown and Rhodes [
13] found that a sense of dog responsibility/obligation was able to predict walking behavior independent of walking intention, which is similar to the results of the current study. In M-PAC, identity is expected to impact action control via selective processing of information, thus shielding from other intentions (i.e., staying on course) and by imposing dissonance (i.e., negative affect) when behavior is not congruent with the identity [
16]. It seems worthy to explore whether dog walking identity can be modified, presumably through pre-set rank-ordering (e.g., priority lists, I will walk the dog before I do household chores), commitment affirmations (realizations and statements about the value of dog walking to companionship), and social activations (purposeful statements about dog walking when describing one’s self to others) [
51].
The only variable discrepant with our hypotheses was perceived opportunity, which did not contribute to the discriminant function with a meaningful effect size. Our assessment of this construct used Williams and Rhodes’ [
37] suggestion to include a motivational qualifier in the assessment (i.e., if I wanted to), which may have explained the null result as opportunity was circumscribed from motivation. Westgarth and colleagues [
10] point out that opportunities to dog walk is not a consistent correlate within the dog walking literature, and it may also not be a critical predictor of action control.
Interestingly, dog breed size also predicted action control independent of the M-PAC variables. Those owners with larger dogs were more likely to enact their walking intentions compared to those with smaller dogs, yet dog size did not distinguish between non-intenders and unsuccessful intenders. The finding is interesting and may highlight how the dog influences walking independent of initial owner-related walking motivations, although assessment with a stronger design (i.e., longitudinal, experimental) is needed to advance this conjecture. Large dogs have higher energy expenditure needs than smaller dogs [
9] so this finding may represent the dog’s influence on action control. Regular dog walking has been characterised as a unique physical activity due to the symbiotic inter-dependency between the canine and owner [
26] and dog characteristics have been correlated with regular walking [
10]. This is the first study to examine action control of dog walking, but future research into the role the dog plays in facilitating or inhibiting owner intentions seems warranted.
Despite the novel findings of this study, the results need to be considered within the context of its limitations and these prompt areas for future research. First, the survey was cross-sectional, making interpretations limited to the assumption that past walking is a good predictor of future dog walking. Second, the assessment of walking is subject to self-report bias. It would stand to reason that an objective assessment of walking would be ideal and a more stringent test of the M-PAC model. Third, the outcome expectation, perceived capability, and perceived opportunity items had limited items that might be compromising the reliability and validity of these constructs. Replication with measures employing multi-item measures with test-retest validation, frames other than daily walking (e.g., 5 times per week) and more dog-related and environmental characteristics is warranted. Finally, the sampling frame was limited to those who visit dog-related social media, posters, and advertisements around Victoria, Canada, and included mainly affluent middle-aged non-Hispanic white women. Recent research suggests that the effects of dog ownership on walking may not extend across racial/ethnic groups [
6], so a broader sampling approach is needed in future research in order to examine whether these findings replicate to other regions and participants with different socio-demographic profiles.