Perceived Stress & Trait Self-Control Interact With Intention-Behavior Gap In Physical Activity PDF

Sport, Exercise, and Performance Psychology...

Sport, Exercise, and Performance Psychology © 2019 American Psychological Association 2020, Vol. 9, No. 2, 244 –260 ISSN: 2157-3905 http://dx.doi.org/10.1037/spy0000189 Perceived Stress and Trait Self-Control Interact With the Intention–Behavior Gap in Physical Activity Behavior Ines Pfeffer Chris Englert Medical School Hamburg—University of Applied Goethe University Frankfurt Sciences and Medical University This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. This document is copyrighted by the American Psychological Association or one of its allied publishers. Anett Mueller-Alcazar Medical School Hamburg—University of Applied Sciences and Medical University Regular physical activity is an empirically well-documented health behavior. Despite the fact that many individuals intend to exercise, they often fail to implement this intention into behavior. Planning is an effective technique for translating physical activity intentions into actions. The present study aimed to examine relevant mediators and moderators in order to explain how and for whom intentions translate into action. In a longitudinal study, 108 participants (Mage ⫽ 31.17 years, 61 women) from different jobs completed measures for intention and trait self-control at baseline (T1), action planning and coping planning 4 weeks later (T2), and physical activity behavior and perceived stress another 4 weeks later (T3). A moderated mediation analysis indicated that perceived stress and trait self-control interact with physical activity intention to predict behavior. In particular, in individuals with low levels of perceived stress and medium–to-high trait self-control, intention and physical activity correlate positively. Unexpectedly, action planning facilitated behavior enactment only in indi- viduals with low perceived stress and high trait self-control. In addition, when per- ceived stress and trait self-control were high, coping planning served as a significant mediator between intention and behavior. Our results highlight the moderating role of perceived stress and trait self-control in the context of intention and physical activity behavior. Under specific perceived stress levels, enactment of behavior based on intentions and planning is supported by high trait self-control. Examining potential mediators and moderators of the intention– behavior gap seems to be a fruitful approach to explain physical activity behavior. Keywords: action planning, coping planning, exercise, self-regulation, volition Physical activity behavior has been identified that the intention to be physically active is one as one of the most important health-related be- of the most important motivational determi- haviors that prevents widespread diseases and nants of actual physical activity behavior improves physical and mental health (Powell, (Ajzen, 1991; Bandura, 1986). However, inten- Paluch, & Blair, 2011; World Health Organiza- tions are not always translated into action (i.e., tion, 2010). It has been repeatedly demonstrated intention– behavior gap; Sheeran & Webb, This article was published Online First September 19, We thank Julia Zieske for her help with data collection. 2019. This research received no specific grant from any funding Ines Pfeffer, Department of Pedagogy, Medical School agency in the public, commercial, or not-for-profit sectors. Hamburg—University of Applied Sciences and Medical Correspondence concerning this article should be ad- University; Chris Englert, Department of Sport Psychol- dressed to Ines Pfeffer, Department of Pedagogy, Medical ogy, Goethe University Frankfurt; Anett Mueller-Alcazar, School Hamburg—University of Applied Sciences and Med- Department of Pedagogy, Medical School Hamburg— ical University, Am Kaiserkai 1, 20457 Hamburg, Germany. University of Applied Sciences and Medical University. E-mail: [email protected] 244  STRESS, SELF-CONTROL, AND PHYSICAL ACTIVITY 245 2016). The identification of the intention– Hamilton, 2017; Bélanger-Gravel, Godin, & behavior gap (i.e., the amount of unexplained Amireault, 2013; Carraro & Gaudreau, 2013; variance when predicting behavior from inten- Kwasnicka, Presseau, White, & Sniehotta, tions; Rhodes & de Bruijn, 2013; Sheeran & 2013; Pfeffer & Strobach, 2019; Scholz et al., Webb, 2016) has led to the development of 2008). Therefore, planning seems to be an im- theories proposing that health-related behavior portant technique by which intentions can be can best be explained by differentiating between effectively translated into actions. motivational processes, which support the for- Planning is thought to enhance perceptual mation of intentions, and volitional self- readiness for the mentally specified contextual regulatory processes, which help an individual cues, which increases automatic elicitation of This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. to translate an intention into action (Schwarzer the respective behavior (i.e., fast, efficient en- This document is copyrighted by the American Psychological Association or one of its allied publishers. et al., 2007; Sniehotta, Scholz, & Schwarzer, actment of the behavior without conscious in- 2005). The study focused on the intention– tent; Gollwitzer, 2014; Hagger & Luszczynska, behavior gap, as the factors that are relevant for 2014; Orbell, Hodgkins, & Sheeran, 1997; forming intentions are well understood (Ajzen, Sheeran, Webb, & Gollwitzer, 2005; Webb, 1991; Bandura, 1986; Sheeran & Webb, 2016), Sheeran, & Luszczynska, 2009). Action plan- whereas the factors that might help to bridge the ning is a task-facilitating technique and seems intention– behavior gap are less clear. There- to be more useful for the adoption of complex fore, we aim to examine relevant mediators and behaviors. Coping planning as a distraction- moderators of the intention– behavior gap in inhibition technique should be particularly fa- order to explain how and under which condi- vorable for behavioral maintenance (Scholz et tions intentions are translated into action. al., 2008). Based on a systematic review of randomized controlled trials, Kwasnicka et al. The Intention–Behavior Gap and Planning (2013) conclude that it is more efficacious to complement action plans with coping plans than Intentions are a necessary but not sufficient forming action plans alone. However, having prerequisite of physical activity behavior strong intentions to be physically active and (Sheeran & Webb, 2016; Sniehotta, Scholz, et subsequently generating physical activity action al., 2005). Planning has been shown to be a and coping plans does not necessarily lead to valuable technique explaining how health- the adoption and maintenance of a physically related intentions translate into action (Scholz, active lifestyle (Sniehotta, 2009). The effective- Schüz, Ziegelmann, Lippke, & Schwarzer, ness of planning seems to be variable, and fu- 2008; Sheeran, Milne, Webb, & Gollwitzer, ture research should examine under which con- 2005). The concept of planning is based on the ditions planning is an effective technique and idea of implementation intentions or so-called should consider potential moderators of the ef- if-then plans (Gollwitzer, 1999) and comprises ficacy of planning (Kwasnicka et al., 2013). (a) action planning and (b) coping planning (Sniehotta, Schwarzer, Scholz, & Schüz, 2005). Effects of Planning and Stress on the Action plans specify when, where, and how to Intention–Behavior Gap perform an intended behavior (e.g., “When I come home tonight, I will go for a run”). Cop- During stressful life periods, life might be ing plans anticipate potential barriers and risk perceived as uncontrollable, unpredictable, and situations that may impede the enactment of the overloaded (Cohen, Kamarck, & Mermelstein, action plans and mental representations of how 1983; Klein et al., 2016). Various barriers may to cope with these potential obstacles (e.g., arise (e.g., job demands, family commitments, “When I come home tonight and it is raining, I and time constraints) that might prevent the will go for a swim instead of going for a run.”). translation of health-related intentions into be- Action and coping planning were previously havior. Requirements of job, children, family, found to mediate the intention– behavior gap in and other factors could operate as barriers and observational and interventional studies of var- obstacles, which keep individuals from enacting ious health-related behaviors, and planning in- an intended behavior. Indeed, during times of terventions were shown to enhance behavior high stress, individuals are less likely to realize enactment (Arnautovska, Fleig, O’Callaghan, & health-related intentions, which leads to a big-  246 PFEFFER, ENGLERT, AND MUELLER-ALCAZAR ger intention– behavior gap, as was shown by havior seems to be inconclusive, and, thus, Payne, Jones, and Harris (2002). more research is needed to fully understand the Gollwitzer and Brandstätter (1997) stated potential interplay of these variables. Besides, that implementation intentions (i.e., planning) Fodor et al. (2014) did not examine the interac- might not necessarily generate an advantageous tion effect of job demands and resources di- effect on behavior beyond health intentions, rectly on the intention– behavior gap. Further- when the initiation of a behavior is easy to more, we argue that specific personal resources realize. In the occupational context, Budden and rather than job-related resources might play a Sagarin (2007) hypothesized that with increas- more important role for leisure time physical ing job stress, that is, under difficult conditions activity behavior enactment. During stressful This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. with regard to health-intention realization, the life periods, trait self-control might be a rele- This document is copyrighted by the American Psychological Association or one of its allied publishers. effect of action planning on physical activity vant personal resource that helps individuals to behavior should increase because forming ac- realize their health-related intentions and thus tion plans will be particularly beneficial in sit- reduce the intention– behavior gap. uations when goal progress is impeded (i.e., under high job stress conditions). Therefore, Effects of Planning, Stress, and Trait Self- forming action plans should be a facilitating Control on the Intention–Behavior Gap technique when perceived job stress is high. The results of their intervention study did not Whether people act on their health-related support this assumption. Forming action plans intentions or not can be explained by an indi- was not more beneficial under high perceived viduals’ ability to control him- or herself. Self- job-stress levels compared with low perceived control is the ability to bring one’s responses job-stress levels. The authors argue that plan- (i.e., thoughts, emotions, impulses, and behav- ning might not be effective for the enactment of iors) in line with (long-term) goals or plans complex behaviors in overloaded life situations (Baumeister & Vohs, 2016). Some individuals (i.e., high stress levels; Budden & Sagarin, are generally better at regulating themselves 2007), which might require less rigid and more than others (i.e., trait self-control; Baumeister, flexible behavioral techniques (Gollwitzer, Fu- Vohs, & Tice, 2007). Trait self-control repre- jita, & Oettingen, 2004). Consequently, Budden sents an individual’s general tendency in how and Sagarin (2007) claimed that action plans effectively self-control is exerted across a wide should be complemented by coping plans as a range of situations and contexts (Baumeister & fallback in case the initial action plan fails. A Vohs, 2016; Baumeister et al., 2007). Individu- shortcoming of this study is that self-reported als low in trait self-control tend to act more on planning was not assessed as a mediator of the their impulses and habits (Buckley, Cohen, intervention effect on behavior. Kramer, McAuley, & Mullen, 2014). High perceived stress levels may also prevent Research has found that higher trait self- the enactment of action and coping plans control is related to higher physical-activity lev- (Fodor, Antoni, Wiedemann, & Burkert, 2014). els (Englert, 2016; Hagger, Wood, Stiff, & In a longitudinal study in the context of fruit and Chatzisarantis, 2010). Furthermore, individuals vegetable consumption, Fodor et al. (2014) as- high in trait self-control might be more adept at sumed that the enactment of action and coping translating their physical activity intentions into plans will be challenging and impeded under action (i.e., reduced intention– behavior gap; conditions where high job demands meet low Hagger, 2013) by bridging potential barriers job resources. The authors expected that em- that may interfere with the execution of the ployees with high job demands are more likely intended plans. Individuals need, for example, to translate their action and coping plans into to overcome obstacles and barriers or to resist behavior when they experience high levels job the impulse to engage in less effortful and less resources compared with employees with low physically demanding sedentary behaviors levels of job resources (three-way interaction (Hagger et al., 2010). Consistent with these effect). Even though their results did not sup- assumptions, Allom, Panetta, Mullan, and Hag- port their assumption of a three-way interaction, ger (2016) found a positive direct relationship we suppose that evidence for the interplay be- between trait self-control and physical activity tween intention, planning, stress, and health be- levels. Bertrams and Englert (2013) showed that  STRESS, SELF-CONTROL, AND PHYSICAL ACTIVITY 247 higher levels of trait self-control were associ- icant three-way interaction effect, Intention ⫻ ated with a smaller intention– behavior gap (En- Stress ⫻ Trait Self-Control (Hypothesis 1). We glert & Rummel, 2016; Schöndube, Bertrams, assume that this relationship will be strongest Sudeck, & Fuchs, 2017), and in a prospective when perceived stress is low and trait self- study, Pfeffer and Strobach (2017) also found control high, and weakest when the perceived high trait self-control to be predictive of a stress level is high and trait self-control low. We smaller intention– behavior gap. also hypothesize that perceived stress and trait With regard to planning, we argue that form- self-control will moderate the action plan– ing action and coping plans can compensate for behavior association, which should be reflected poor self-control abilities. Specifying a cue– in a significant three-way interaction effect, Ac- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. behavior link will pass behavioral control to an tion Planning ⫻ Stress ⫻ Trait Self-Control This document is copyrighted by the American Psychological Association or one of its allied publishers. environmental cue, and the behavior is elicited (Hypothesis 2). This relationship will be stron- automatically when the respective cue occurs, gest when the perceived stress level is high and independently of one’s general self-control trait self-control low and weakest when the per- skills (Gollwitzer, 2014; Hagger & Luszczyn- ceived stress level is low and trait self-control ska, 2014; Orbell et al., 1997; Sheeran, Webb, high. Likewise, we expect that perceived stress et al., 2005; Webb et al., 2009). Planning of and trait self-control would moderate the coping physical activity behavior was shown to com- plan– behavior relationship, which should be re- pensate for poor executive functions abilities flected in a significant three-way interaction ef- (i.e., cognitive self-control abilities) in younger fect, Coping Planning ⫻ Stress ⫻ Trait Self- adults (Hall, Zehr, Ng, & Zanna, 2012). To the Control (Hypothesis 3). We propose that this best of our knowledge, the role of trait self- association will be strongest when stress is high control for the plan– behavior relationship has and trait self-control low and weakest when not been examined before and therefore is a stress level is low and trait self-control high. primary goal of the present study. Furthermore, we expect a moderated mediation effect and suppose that action planning and The Present Study coping planning will be mediators between in- tention and behavior, particularly when the per- The primary aim of the current study was to ceived stress level is high and trait self-control examine the interplay between planning, per- is low (Hypothesis 4). ceived stress, trait self-control, and the inten- tion– behavior gap. Perceived stress and trait Method self-control were independently shown to mod- erate the intention– behavior gap. Moreover, Participants and Procedure planning is known as an effective technique that helps individuals to reduce this gap (Bélanger- A longitudinal study with three points of Gravel et al., 2013; Kwasnicka et al., 2013; measurement (i.e., T1, T2, and T3) and a time Pfeffer & Strobach, 2019; Scholz et al., 2008). lag of 4 weeks between each point of measure- In line with Budden and Sagarin (2007), we ment was conducted. Data were obtained using assume that planning could be particularly ben- standardized scales and an online survey tool eficial when life is perceived to be overloaded for quantitative research (Software Unipark and uncontrollable and trait self-control abilities QuestBack EFS Survey 10.8 for academic re- are low, as mentally linking a specific situation search, Cologne, Germany). The survey link with a behavior will elicit the behavior automat- was distributed via e-mail and social networks ically in case the environmental cue appears by research assistants. The procedure was in (Gollwitzer, 1999). The two moderators (i.e., accordance with the ethical standards of the perceived stress and trait self-control) were ex- institutional research committee and with the pected to jointly impact the regression of the Declaration of Helsinki. As the most effective dependent variable on the independent variable. means to reduce common method variance and In line with Fodor et al. (2014), we hypothesize systematic measurement errors (which might that perceived stress and trait self-control will lead to an inflation or a deflation of the corre- jointly moderate the intention– behavior rela- lation of the examined variables) is to separate tionship, which should be reflected in a signif- the measure of the constructs, we placed a time  248 PFEFFER, ENGLERT, AND MUELLER-ALCAZAR lag between the assessment of the examined made on 4-point Likert scales ranging from 1 variables and also of the moderators (Chan et (not at all true) to 4 (absolutely true). Items of al., 2015). After providing informed consent, the Action Planning Scale were introduced by sociodemographic and control variables (i.e., the stem “I have made detailed plans for the age, gender, and past physical activity behavior) next weeks regarding...”, which was followed and the independent and moderator variables by six items (“... where to do my exercise”; (i.e., intention and trait self-control) were as- Pfeffer & Strobach, 2019). The introduction of sessed at T1, mediator variables (i.e., action and the Coping Planning Scale read, “I have made coping planning) were measured at T2, and the detailed plans regarding...”. This stem was dependent variable (i.e., physical activity be- followed by four items (e.g., “... what to do, if This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. havior) as well as perceived stress level were something intervenes”). Cronbach’s ␣ was cal- This document is copyrighted by the American Psychological Association or one of its allied publishers. measured at T3. culated at.93 for action planning and.90 for A total of N ⫽ 338 people participated in the coping planning in the current study, reflecting first point of measurement. Of these partici- very good internal consistencies for both scales. pants, n ⫽ 172 took part at T2 and n ⫽ 128 at The higher the score, the more detailed are an T3. Overall, 108 participants (32%; 61 females, individual’s action and coping plans. 47 male; Mage ⫽ 37.17 years; SDage ⫽ 12.57; Perceived stress. Subjective stress was as- age range: 19 – 65 years) completed question- sessed at T3 with the German version of the naires from all three points of measurement. Of Perceived Stress Scale (Cohen et al., 1983; these participants, 68% were employed full Klein et al., 2016), using 10 items to assess the time, 11% had a part-time job, and 2% received subjective stress levels over the past month. an hourly wage. Another 15% were students, Participants answered each item on 5-point Lik- and 4% participants marked the option “other.” ert scales ranging from 1 (never) to 5 (very often). It includes items such as “In the last Measures month, how often have you found that you could not cope with all the things that you had Physical activity intention. The intention to do?” and “In the last month, how often have to be physically active was assessed at T1 using you been able to control irritations in your a standardized scale in line with the theory of life?”. (Cronbach’s ␣ ⫽.84; ␣ ⫽.88 in the planned behavior (Ajzen, 1991). The items current study). were answered on 6-point Likert scales ranging Trait self-control. Trait self-control was from 1 (strongly disagree) to 6 (strongly agree). measured at T1 with the German short version The scale was introduced by the sentence of the Self-Control Scale (Bertrams & Dick- “Please indicate to which extent the following häuser, 2009; Tangney, Baumeister, & Boone, statements apply to yourself,” followed by three 2004). Items were rated on 5-point Likert scales items: “I intend to be physically active for at ranging from 1 (not at all) to 5 (very much). The least 30 min per day with moderate-to-vigorous items were preceded by the sentence “Please intensity,” “I plan to be...,” and “I am deter- indicate to which extent the following state- mined to be...” (Pfeffer, 2012; Pfeffer & ments generally apply to yourself,” which was Strobach, 2019). The criterion of at least 30 min followed by 13 items (e.g., “I am good at re- per day of moderate-to-vigorous physical activ- sisting temptation” and “I wish I had more ities is based on the World Health Organiza- self-discipline”). ␣ ⫽.80 in the current study. tion’s recommendation for physical activity be- Physical activity behavior. Physical activ- havior (WHO, 2010). For this scale and all the ity behavior was measured at T1 and T3 with other scales applied in this study, we calculated four items of the Godin Leisure-Time Exercise mean scores, with higher scores always reflect- Questionnaire (Godin & Shephard, 1985). Par- ing higher values on the respective variable. ticipants were asked to indicate how many times Internal consistency for the intention scale was they performed (a) vigorous (e.g., running, jog- calculated at ␣ ⫽.89 in the current study, ging, soccer, squash, basketball, roller skating, indicating good to very good reliability. vigorous swimming, and vigorous long distance Planning. Action and coping planning was bicycling) and (b) moderate (e.g., fast walking, assessed by applying the scales by Sniehotta, tennis, bicycling, volleyball, badminton, swim- Schwarzer, et al. (2005) at T2. Responses were ming, and dancing) physical activities for more  STRESS, SELF-CONTROL, AND PHYSICAL ACTIVITY 249 than 10 min during their spare time during the the model in order to prevent misleading inter- past month. Furthermore, participants indicated pretations of the findings. Predictor variables how long (minutes) these activities were per- were standardized (z transformed) before enter- formed on average per occasion. The moderate- ing them in the analysis. Age and sex were to-vigorous physical activity score was calcu- entered as covariates for the mediators and the lated by multiplying the frequency (during the dependent variable as these variables usually last 4 weeks) with the average duration per correlate with physical activity behavior occasion (minutes) for (a) vigorous and (b) (Caspersen, Pereira, & Curran, 2000). As past moderate physical activities and subsequently physical activity behavior is often the strongest summing up the minutes spent in vigorous-to- predictor of future behavior, we also controlled This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. moderate physical activities during the last 4 for physical activity behavior at T1. However, This document is copyrighted by the American Psychological Association or one of its allied publishers. weeks. The score represents the time spent this might lead to an underestimation of the (minutes) in moderate-to-vigorous physical ac- effects of the examined predictors in the model tivities during the last 4 weeks. This question- (Weinstein, 2007). We used the T1 measure of naire has shown adequate validity with other, trait self-control as moderator, as it captures a more objective measures of physical activity trait variable, which is relatively stable over (Godin & Shephard, 1985), and adequate reli- time. We chose to measure the perceived stress ability, r ⫽.86 (van Poppel, Chinapaw, Mok- measure at T3 as we were interested in per- kink, Van Mechelen, & Terwee, 2010). ceived stress as a state. In this way, we were Statistical Analyses able to examine if perceived stress during the past month might have been a correlate of not SPSS was used for data screening and data translating physical activity plans at T2 into analyses, supplemented by the macro PROCESS action at T3. (Hayes, 2013). Model 19 of PROCESS allows To test the three-way interaction hypotheses to test the moderated mediation model specified (Hypotheses 1–3), the product of the three re- in Figure 1 and also allows to test our hypoth- spective variables were calculated and inserted eses. In Hypotheses 1 to 4, we expected three- as predictor in the regression analyses. Signifi- way interaction effects as well as moderated cant three-way interaction effects were further mediation effects in the context of the inten- analyzed by allocating participants to groups of tion– behavior relationship. As our analyses high, average, and low scores of the respective were based on regression analyses, we tested for moderator variable(s) and predicting the depen- multicollinearity of the predictor variables in dent from the independent variable for these Figure 1. Conceptual moderated mediation model underlying the present study. T1 ⫽ first point of measurement; T2 ⫽ second point of measurement, 4 weeks after T1; T3 ⫽ third point of measurement, 8 weeks after T1; H1⫺H4 ⫽ Hypothesis 1– 4.  250 PFEFFER, ENGLERT, AND MUELLER-ALCAZAR groups separately (simple slopes at the mean as (Table 2) were calculated at ⱕ3.72, indicating well as at 1 SD below and above the mean of the low multicollinearity (tolerance ⱖ.27; Menard, stress and the trait self-control score respec- 1995). The model predicting action planning tively; Aiken & West, 1991; Hayes, 2013). To revealed 14% of explained variance in the de- determine the significance of the moderated me- pendent variable (T2), R2 ⫽.14, F(4, 103) ⫽ diation effects (Hypothesis 4), the bootstrapping 4.12, p ⫽.004. Past physical activity behavior approach with 10.000 bias-corrected bootstraps was the only significant predictor of action plan- was used (Aiken & West, 1991; Hayes, 2013; ning, b ⫽ 0.29, p ⫽.004, whereas intention did Hayes & Scharkow, 2013). not significantly predict the criterion, b ⫽ 0.16, p ⫽.087. For coping planning as a dependent This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Results variable, the model explained 23% of the vari- This document is copyrighted by the American Psychological Association or one of its allied publishers. Descriptive Statistics ance in this criterion, R2 ⫽.23, F(4, 103) ⫽ 7.50, p ⬍.001. Past physical activity behavior, Descriptive statistics and correlations among b ⫽ 0.30, p ⫽.002, as well as intention, b ⫽ study variables are illustrated in Table 1. Cor- 0.27, p ⫽.003, were significant predictors of relations among study variables were low to coping planning (Table 2). medium. With regard to moderate-to-vigorous The multiple linear regression analysis pre- physical activity behavior during the past 4 dicting physical activity behavior at T3 revealed weeks measured at T1, 5.6% of the participants a significant amount of explained variance, were not at all physically active, 15.8% were R2 ⫽.59, F(18, 89) ⫽ 7.15, p ⬍.001. Past active between 0.5 to 1 hr per week, 17.6% physical activity behavior was the strongest pre- between 1 and 2 hr per week, 2.8% between 2.0 dictor of behavior at T3, b ⫽ 327.71, p ⬍.001. and ⬍2.5 hr per week, indicating that 51.5% of As expected and in line with Hypotheses 1 to 3, the participants did not reach the recommended the three-way interactions Intention ⫻ Stress ⫻ physical activity level of 2.5 hr of moderate-to- Trait Self-Control, b ⫽ ⫺172.27, p ⫽.007, vigorous activities per week (WHO, 2010). Action Planning ⫻ Stress ⫻ Trait Self-Control, However, 48.5% of the participants had an ac- b ⫽ ⫺200.47, p ⫽.045, as well as Coping tivity level ⱖ2.5 hr per week and were suffi- Planning ⫻ Stress ⫻ Trait Self-Control, b ⫽ ciently physically active. 226.74, p ⫽.018, were statistically significant Moderated Mediation Model (Table 3). The three-way interaction Intention ⫻ Stress ⫻ The variance inflation factors for the predic- Trait Self-Control significantly increased the tors of physical activity behavior in this model amount of explained variance by 3%, R2 ⫽.03, Table 1 Descriptive Statistics and Bivariate Correlations Among Study Variables (N ⫽ 108) Variables 1 2 3 4 5 6 7 8 9 1. Age — 2. Sex.05 — 3. Physical activity (T1).12.19 — 4. Trait self-control (T1).40ⴱⴱⴱ ⫺.14.10 — 5. Intention (T1) ⫺.05.02.31ⴱⴱⴱ.20ⴱ — 6. Action planning (T2) ⫺.04 ⫺.01.34ⴱⴱⴱ.22ⴱ.26ⴱⴱ — 7. Coping planning (T2).09.02.40ⴱⴱⴱ.35ⴱⴱⴱ.36ⴱⴱⴱ.71ⴱⴱⴱ — 8. Perceived stress (T3).02.04 ⫺.14 ⫺.25ⴱⴱ ⫺.12 ⫺.06 ⫺.15 — 9. Physical activity (T3).24ⴱ.07.59ⴱⴱⴱ.27ⴱⴱ.35ⴱⴱⴱ.40ⴱⴱⴱ.43ⴱⴱⴱ ⫺.26ⴱⴱ — M 37.17 56.9%a 695.28 3.33 3.83 2.77 2.53 2.49 730.44 SD 12.57 — 677.55.58 1.62.90.92 0.75 703.69 Minimum 19 —.00 1.77 1.00 1.00 1.00 1.00 0.00 Maximum 65 — 3340.00 4.62 6.00 4.00 4.00 4.50 3960.00 Note. T1 ⫽ baseline; T2 ⫽ four weeks after baseline; T3 ⫽ eight weeks after baseline. a percentage of women. ⴱ p ⬍.05. ⴱⴱ p ⬍.01. ⴱⴱⴱ p ⬍.001.  STRESS, SELF-CONTROL, AND PHYSICAL ACTIVITY 251 Table 2 Regression Models Predicting Action and Coping Planning From Physical Activity Intention (Controlling for Age, Sex, and Physical Activity Behavior at T1) Action planning (T2) Coping planning (T2) Variables B SE t R2 B SE t R2.14ⴱⴱ.23ⴱⴱⴱ Age ⫺.04.09 ⫺.39.09.09 1.03 Sex ⫺.05.09 ⫺.51 ⫺.03.09 ⫺.34 Physical activity (T1).29ⴱⴱ.10 2.92.30ⴱⴱ.09 3.18 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Intention (T1).17†.10 1.73.27ⴱⴱ.09 2.99 This document is copyrighted by the American Psychological Association or one of its allied publishers. Note. T1 ⫽ baseline; T2 ⫽ four weeks after baseline. † p ⬍.10. ⴱ p ⬍.05. ⴱⴱ p ⬍.01. ⴱⴱⴱ p ⬍.001. F(1, 89) ⫽ 7.51, p ⫽.007. The conditional levels of trait self-control revealed that this two- effects of intention on physical activity behav- way interaction was a nonsignificant predictor ior at values of the moderators revealed that of behavior when self-control was low, b ⫽ intention was a significant predictor of behavior 68.04, t ⫽ 0.73, p ⫽.465, and a negative but when trait self-control was medium, b ⫽ 165. also nonsignificant predictor when trait self- 61, t ⫽ 2.24, p ⫽.028, and high, b ⫽ 307.80, control was at a medium level, b ⫽ ⫺104.23, t ⫽ 3.42, p ⫽.001, but only when perceived t ⫽ ⫺1.72, p ⫽.090. However, when trait stress was low at the same time. Other signifi- self-control was high, Intention ⫻ Perceived cant conditional effects of intention on behavior Stress was a significant negative predictor of were not found. The conditional effect of Inten- behavior, b ⫽ ⫺276.50, t ⫽ ⫺3.38, p ⫽.001 tion ⫻ Perceived Stress on behavior at different (Figure 2). Table 3 Regression Model Predicting Physical Activity Behavior (T3) From Intention (T1), Action and Coping Planning (T2), Trait Self-Control (T1) and Perceived Stress (T3) and Their Interactions (Controlling for Age, Sex, and Physical Activity at T1) Variables B SE t R2.59ⴱⴱⴱ ⴱ Age 121.71 57.50 2.12 Sex 6.44 51.74.12 Physical activity (T1) 327.71ⴱⴱⴱ 57.79 5.67 Intention (T1) 61.98 57.58 1.08 Trait self-Control (T1) 17.52 64.88.27 Action planning (T2) 110.49 79.51 1.39 Coping planning (T2) 38.73 78.59.49 Perceived stress (T3) ⫺92.34 55.65 ⫺1.66 Stress ⫻ Self-Control ⫺19.66 57.85 ⫺.34 Intention ⫻ Stress ⫺104.23† 60.79 ⫺1.72 Intention ⫻ Self-Control ⫺29.09 55.20 ⫺.53 Intention ⫻ Stress ⫻ Self-Control ⫺172.27ⴱⴱ 62.88 ⫺2.74 Action Planning ⫻ Stress ⫺83.75 88.73 ⫺.94 Action Planning ⫻ Self-Control ⫺8.62 81.53 ⫺.11 Action Planning ⫻ Stress ⫻ Self-Control ⫺200.48ⴱ 98.76 ⫺2.03 Coping Planning ⫻ Stress 55.82 82.06.68 Coping Planning ⫻ Self-Control 74.56 76.00.98 Coping Planning ⫻ Stress ⫻ Self-Control 226.74ⴱ 94.32 2.40 Note. T1 ⫽ baseline; T2 ⫽ four weeks after baseline; T3 ⫽ eight weeks after baseline. † p ⬍.10. ⴱ p ⬍.05. ⴱⴱ p ⬍.01. ⴱⴱⴱ p ⬍.001.  252 PFEFFER, ENGLERT, AND MUELLER-ALCAZAR This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. This document is copyrighted by the American Psychological Association or one of its allied publishers. Figure 2. Three-way interaction effect of Intention ⫻ Perceived Stress ⫻ Self-Control on physical activity behavior. T3 ⫽ eight weeks after baseline. The three-way interaction Action Planning ⫻ conditional effects of coping planning on phys- Stress ⫻ Trait Self-Control significantly in- ical activity behavior at values of the modera- creased the amount of explained variance by tors revealed that coping planning was no sig- 2%, R2 ⫽.02, F(1, 89) ⫽ 4.12, p ⫽.045. The nificant predictor of behavior irrespective of the conditional effects of action planning on phys- values in trait self-control and perceived stress ical activity behavior at values of the modera- level. However, the conditional effect of Cop- tors revealed that action planning was only a ing Planning ⫻ Perceived Stress on behavior at significant predictor of behavior when trait self- different levels of trait self-control revealed that control was high and perceived stress was low, the two-way interaction was a negative but not b ⫽ 384.45, t ⫽ 2.70, p ⫽.008. Other signifi- significant predictor of behavior when trait self- cant conditional effects of action planning on control was low, b ⫽ ⫺170.92, t ⫽ ⫺1.42, p ⫽ behavior were not found. The conditional effect.158, and a positive but not significant predictor of Action Planning ⫻ Stress on behavior at when trait self-control was medium, b ⫽ 55.82, different levels of trait self-control revealed that t ⫽ 0.68, p ⫽.498. When trait self-control was the two-way interaction was a positive but not high Coping Planning ⫻ Stress was a positive significant predictor of behavior when trait self- and significant predictor of behavior, b ⫽ 282. control was low, b ⫽ 116.72, t ⫽ 1.10, p ⫽ 57, t ⫽ 2.18, p ⫽.032 (Figure 4)..272, and a negative but not significant predictor The indices of the moderated mediation ef- when self-control was medium, b ⫽ ⫺83.75, fects proposed in Hypothesis 4 revealed a non- t ⫽ ⫺0.94, p ⫽.348, and high, b ⫽ ⫺284.21, significant effect for action planning, b ⫽ ⫺33. t ⫽ ⫺1.83, p ⫽.070 (Figure 3). 06, SE ⫽ 32.77, 95% confidence interval [CI] The three-way interaction Coping Plan- [⫺132.01, 6.94], but a significant effect for ning ⫻ Stress ⫻ Trait Self-Control significantly coping planning, b ⫽ 61.66, SE ⫽ 43.21, 95% increased the amount of explained variance by CI [0.16, 173.45]. Hypothesis 4 was therefore 3%, R2 ⫽.03, F(1, 89) ⫽ 5.78, p ⫽.018. The only partially supported. The indices of the con-  STRESS, SELF-CONTROL, AND PHYSICAL ACTIVITY 253 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. This document is copyrighted by the American Psychological Association or one of its allied publishers. Figure 3. Three-way interaction effect of Action Planning ⫻ Perceived Stress ⫻ Self- Control on physical activity behavior. T3 ⫽ eight weeks after baseline. ditional moderated mediation effects revealed havior in case the specified environmental cue that coping planning was a significant mediator appears. between intention and behavior when perceived Regarding Hypothesis 1, physical activity in- stress was high, b ⫽ 82.50, SE ⫽ 51.20, 95% CI tention was a significant positive predictor of [10.35, 224.73]. Other significant conditional behavior when the perceived stress level was moderated mediation effects were not observed. low and trait self-control was medium to high. This finding is in line with previous studies Discussion showing that trait self-control is a resource that helps reducing the intention– behavior gap (Al- The primary aim of the current study was to lom et al., 2016; Bertrams & Englert, 2013; examine the interplay between planning, per- Pfeffer & Strobach, 2017). This seems to be ceived stress, trait self-control, and the inten- particularly the case when the life situation is tion– behavior gap. It has been previously less stressful. Furthermore, irrespective of the shown that the level of perceived stress as well perceived stress level, intention was no signifi- as trait self-control act as moderators of the cant predictor of behavior for people with low intention– behavior gap and that planning seems trait self-control, medium-to-high stress levels to be an effective strategy to reduce this gap (Payne et al., 2002), or medium to high trait (Bélanger-Gravel et al., 2013; Pfeffer & self-control. That is, our results point out that Strobach, 2019; Scholz et al., 2008). In line with high physical activity intentions may mainly be Gollwitzer (1999) and Budde