Deepening the measurement of motivation in the physical activity domain: Introducing behavioural resolve

https://doi.org/10.1016/j.psychsport.2012.12.010Get rights and content

Abstract

Objectives

Intention is theorized as the proximal determinant of behaviour in many leading theories and yet tests of its absolute predictive utility show discordance. While one line of research has been investigating constructs that may augment intention, another possibility may be to improve measurement. The purpose of this study was to compare a typical measure of exercise intention with a measure that attempts to deepen the motivational domain via considerations of contextual barriers and other competing goals (named behavioural resolve).

Design

Two-week prospective.

Method

Participants were a random sample of university students who were subsequently randomized to groups who completed either measures of behavioural intention (n = 179) and behavioural resolve (n = 227) in relation to exercise and a 2-week follow-up of exercise behaviour.

Results

Comparing the two measures showed that the behavioural resolve construct explained significantly more variance in exercise behaviour than the standard intention construct (q = .35). Further comparison of the two measures showed that absolute discordance with subsequent exercise behaviour was considerably lower with behavioural resolve (mean = −.09) compared to behavioural intention (mean = −1.28).

Conclusions

These findings indicate that at least some of the intention–behaviour gap may be from inadequate measurement of the motivational domain, and this can be partially rectified with shifting to a behavioural resolve measure.

Highlights

► Resolve explained significantly more variance in exercise behaviour than intention. ► Discordance with exercise behaviour was lower with resolve compared to intention. ► Resolve effectively shrinks the gap between motivation and subsequent behaviour.

Introduction

Regular physical activity at a moderate to vigorous intensity has been linked to the reduction of over 25 chronic conditions including cardiovascular disease, type 2 diabetes, and several cancers (Warburton, Charlesworth, Ivey, Nettlefold, & Bredin, 2010). Unfortunately, it has been estimated that over 80% of adults fail to meet the minimum amount of physical activity to reap these benefits (Colley et al., 2011; Troiano et al., 2008). The need to promote physical activity is paramount. Understanding the basis for physical activity behaviours through sound behavioural theory should be helpful in order to inform interventions (Rhodes & Pfaeffli, 2010).

Many theories applied to physical activity include intention as the proximal antecedent to behavioural performance (Bandura, 2004; Fishbein et al., 2001; Noar & Zimmerman, 2005). Specifically, intention is considered the hinge between other physical activity antecedents and behavioural performance and interventions that can increase intention are expected to subsequently produce behavioural changes (Ajzen & Fishbein, 2005). Correlational tests of this proposition via theory of planned behaviour (Symons Downs & Hausenblas, 2005), protection motivation theory (Plotnikoff & Trinh, 2010), and variants of social cognitive theory (Roberts, Maddison, Magnusson, & Prapavessis, 2010) and self-determination theory (Hagger & Chatzisarantis, 2009) have been supportive. Further, meta-analyses have placed the point estimate of the intention–behaviour relationship as r = .50 (Armitage & Conner, 2001; Hagger, Chatzisarantis, & Biddle, 2002; McEachan, Conner, Taylor, & Lawton, 2011; Symons Downs & Hausenblas, 2005), which situates it within the large effect size range (Cohen, 1992) and larger than any other known correlate of physical activity.

Despite the relative predictive value of intention in the physical activity domain, tests of its absolute predictive utility show discordance with behaviour. For example, a recent meta-analytic examination of experimental changes in intention–behaviour relationships showed that medium-sized changes in intention resulted in trivial-sized changes in behaviour (r = .06) (Rhodes & Dickau, 2013). Passive prospective designs that separate the intention–behaviour relationship into quadrants also demonstrate considerable discordance of intention and behaviour (Godin, Shephard, & Colantonio, 1986; Sheeran, 2002). These studies typically show that half of all people who intend to be active fail to reach these intentions (Rhodes, Plotnikoff, & Courneya, 2008). This line of enquiry has generated the term “intention–behaviour gap” and initiated some researchers to develop post-motivational (e.g., planning, self-regulation, implementation intention) models meant to bridge the intention–behaviour relationship (Gollwitzer, 1999; Schwarzer, 2008; Sniehotta, 2009). These models suggest that behavioural performance includes a motivational phase leading to intention followed by a volitional phase beyond intention.

While the extension of intention models with self-regulation and planning constructs has had evidence for its predictive utility (Gollwitzer & Sheeran, 2006; Schwarzer, 2008), there are other possibilities that may explain the intention–behaviour discrepancy. One possibility may be measurement. Pure measurement effects such as scale correspondence or duration between intention–behaviour assessments can attenuate the relationship (Courneya, 1994; Symons Downs & Hausenblas, 2005). Domain assessment of motivation, however, may be the most prominent reason for the intention–behaviour gap in physical activity. Intention has been defined as decisions to perform a behavioural action (i.e., behavioural direction) in its most modest conceptualization, to the effort and time one is willing to spend to perform a behaviour (i.e., behavioural intensity) in its most encompassing measurement domain consideration (Sheeran, 2002). We postulate that intention measurement has typically aligned with the decisional aspects of the domain more than the intensity aspects and this may be part of the reason for the intention–behaviour gap.

For example, a recent systematic review on the moderators of the intention–behaviour relationship in the physical activity domain showed that intention strength, measured in this case as intention temporal stability, was the dominant explanation for intention–behaviour discordance (Rhodes & Dickau, 2012). This suggests that inadequate measurement of initial physical activity motivation may account for much of the intention–behaviour gap. There is also evidence for this line of reasoning in studies that have used social cognitive constructs to predict unsuccessful intenders (i.e., those who intended to act but did not) from successful intenders (those who intended to act and did). In these studies (Rhodes, Courneya, & Jones, 2003; Rhodes & Plotnikoff, 2006; Rhodes, Plotnikoff et al., 2008), affective attitude (i.e., enjoyment) and perceived behavioural control/self-efficacy predict intention–behaviour discordance, suggesting that some antecedent motivation may not be accounted for within intention.

The problem may reside in typical intention items that use “I intend to exercise” or “I plan to exercise” phraseology. While these statements provide direction (i.e., intention or no intention) and the scaling creates some assessment of strength (e.g., strongly agree to strongly disagree or frequency of days chosen to exercise), it seems plausible that the items neglect the context of how much effort will be expended to act on the intention across time. This type of assessment may be sufficient in simple or single act behaviours, but a repeated behaviour like physical activity (e.g., regular vigorous exercise) that requires one to organize large amounts of time, bring the body out of rest and undergo some pain/fatigue, use physical skills, arrange a suitable location, etc. requires considerable effort and attention that may not be assessed with these simple items. Indeed, the current problem with intention–behaviour discordance is from a large number of false positives (i.e., intenders who never act), which supports the notion that simple “I intend” items may fail to measure the effort people are willing to expend on physical activity.

This critique has received some prior attention. For example, Ajzen and Fishbein (1977) originally considered a concept called conditional intention (e.g., intention in the face of potential situations). It is possible that conditional intentions could improve upon the predictive capability of ordinary intention items because the items force participants to consider situations that may affect behavioural engagement, but the employment of this type of measure has not been tried to our knowledge. Of course, the potential weakness of these conditional statements is that their accuracy is contingent on whether these simulated situations occur with enough frequency to reflect a person's experiences (see also extended theorizing by Chatzisarantis, Hagger, Smith, & Phoenix, 2004).

Some researchers have examined poorly formed versus well-formed intentions in other behavioural domains and found that better intention formation can predict intention stability (Sheeran, 2002). This appears sound reasoning and parallels how planning (potentially an example of well-formed intentions) can out predict intention due to the detail of the intention formation. The problem with this approach is that it can obscure where motivation ends and volitional processes begin because one is not sure if planning is a post-motivational volitional process or a better indicator of motivation in the form of a well-formed intention. Probably the best use of contextualizing social cognition has been through the measurement of self-efficacy (Bandura, 1997), where confidence appraisals are made against the context of common barriers (e.g., time, fatigue). It seems conceivable that the use of these contextual prompts could also aid in understanding just how motivated an individual may be to perform physical activity.

Another area that may be neglected with simple “I intend” statements is intention within the context of other behavioural intentions or goals. Prior physical activity research has shown that intentions or goals to perform other behaviours is often negatively correlated with physical activity independent of physical activity intentions (Gebhardt & Maes, 1998; Gebhardt, Van Der Doef, & Maes, 1999; Presseau, Sniehotta, Francis, & Gebhardt, 2010; Rhodes & Blanchard, 2008, 2011; Rhodes, Blanchard, & Bellows, 2008; Riediger & Freund, 2004). This line of enquiry follows research on the selection and optimization of a goal (Freund & Baltes, 2002). Specifically, priorities of intentions and the subsequent attention control placed on these intentions may be neglected in simple statements about intention to perform physical activity. It may be this level of detail that helps better differentiate intention strength.

Thus, the purpose of this study was to compare a typical measure of exercise intention with a measure that attempts to deepen the motivational domain assessment via considerations of contextual barriers, energy and other competing goals. To distinguish this measure from a quarter of a century of intention research we have named the construct behavioural resolve. It was hypothesized that behavioural resolve would be a better predictor of exercise behaviour than behavioural intention. However, the absolute value considerations were also of consideration in our comparison of the two measures. It was hypothesized that resolve would have a lower mean score and a smaller level of discordance with behaviour. The theorizing behind these hypotheses followed the logic that resolve would elicit far fewer false positives and have superior predictive efficacy than intention by contextualizing motivation within the frame of self-regulation, goal prioritization and energy demands. The resolve construct is similar to the concept of conditional intention in its attempt to ground motivation to real life circumstances, yet our focus on goal prioritization and energy level may sidestep much of the dependency on any particular situation.

Section snippets

Participants and procedure

Four hundred and sixty-five participants were recruited through classes at a university on the west coast of British Columbia, Canada during the spring 2011 semester (January to February). There are approximately 19,500 students at this University in ten different faculties ranging from Fine Arts to Sciences. The institution's human ethic's review board approved the protocol for the study and all participants provided informed consent.

A list of all classes at the university was used to randomly

Results

A total sample of n = 179 for standard intention measure group and n = 227 for the resolve measure group was used in the analyses. The mean age of participants was 22.93 years (SD = 7.32); 64.9% were female, and the mean year of studies was 2.70 (SD = 1.19). In terms of vigorous exercise, the total sample reported a mean frequency of 2.69 bouts per week (SD = 2.19). Importantly, no differences in these descriptors were identified by measurement (intention/resolve) group (p > .60).

The mean

Discussion

The prominent intention–behaviour gap in physical activity and health behaviour research has prompted considerable interest in additional constructs that may augment this relationship. Still, the original measurement domain of behavioural motivation has seen negligible attention. It is possible that our measures of the intention construct do not assess the depth of motivation for a repeated behaviour like physical activity and this has resulted in at least some of this intention–behaviour

References (55)

  • Canadian Fitness and Lifestyle Research Institute

    2008 physical activity monitor facts and figures

    (2009)
  • Canadian Society for Exercise Physiology

    Canadian physical activity guidelines for adults 18–64 years

    (2011)
  • N.L.D. Chatzisarantis et al.

    The influences of continuation intentions on execution of social behaviour within the theory of planned behaviour

    British Journal of Social Psychology

    (2004)
  • G.W. Cheung et al.

    Evaluating goodness-of-fit indexes for testing measurement invariance

    Structural Equation Modeling

    (2002)
  • J. Cohen

    A power primer

    Psychological Bulletin

    (1992)
  • R.C. Colley et al.

    Physical activity of Canadian adults: accelerometer results from the 2007 to 2009 Canadian health measures survey

    Health Reports Statistics Canada

    (2011)
  • K.S. Courneya

    Predicting repeated behavior from intention: the issue of scale correspondence

    Journal of Applied Social Psychology

    (1994)
  • K.S. Courneya et al.

    Factors affecting the intention-physical activity relationship: intention versus expectation and scale correspondence

    Research Quarterly for Exercise and Sport

    (1994)
  • M. Fishbein et al.

    Factors influencing behavior and behavior change

  • A.M. Freund et al.

    Life-management strategies of selection, optimization, and compensation: measurement by self-report and construct validity

    Journal of Personality and Social Psychology

    (2002)
  • B. Gardner et al.

    A systematic review and meta-analysis of applications of the Self-Report Habit Index to nutrition and physical activity behaviors

    Annals of Behavioral Medicine

    (2011)
  • W.A. Gebhardt et al.

    Competing personal goals and exercise behaviour

    Perceptual and Motor Skills

    (1998)
  • W.A. Gebhardt et al.

    Conflicting activities for exercise

    Perceptual and Motor Skills

    (1999)
  • G. Godin et al.

    Assessment of leisure time exercise behavior by self-report: a concurrent validity study

    Canadian Journal of Public Health

    (1986)
  • G. Godin et al.

    A simple method to assess exercise behavior in the community

    Canadian Journal of Applied Sport Science

    (1985)
  • G. Godin et al.

    The cognitive profile of those who intend to exercise but do not

    Public Health Reports

    (1986)
  • P.M. Gollwitzer

    Implementation intentions: strong effects of simple plans

    American Psychologist

    (1999)
  • No funding was received for completion of this paper.

    View full text