Undertaking regular physical activity (PA) is one of the recommendations of public health information campaigns. Among the available forms of PA (e.g., those related to work, household or leisure), sport activity in their free time is one means that many individuals use to respond to these health messages. But each person responds in his or her own way, engaging in their preferred sport with greater or lesser amounts of practice. One of the most prominent factors determining individuals’ participation in their sport is their motivation. Many motivational theories and models have brought real contributions to explain the antecedents and processes that give rise to sports practice (Ryan, Reference Ryan2012). In France, the motivations most often mentioned by adults who engage in sport are pleasure and health (Gleizes & Pénicaud, Reference Gleizes and Pénicaud2017, November). The present study focuses on these health-related motivations, and the questions addressed are the following: What is the influence of health motivation on the amount of sports practice? What process can explain this influence?

Regulatory focus theory (Higgins, Reference Higgins1997), a well-established motivational theory and a useful framework in the health research area (Fuglestad, Rothman, & Jeffery, Reference Fuglestad, Rothman and Jeffery2008; Leder, Florack, & Keller, Reference Leder, Florack and Keller2015; Uskul, Keller, & Oyserman, Reference Uskul, Keller and Oyserman2008), argues for the existence of two regulatory systems (promotion and prevention) which guide people in their self-care behaviors. Self-regulation with a promotion focus is associated with growth and accomplishment needs and is reflected in a concern for the presence or absence of positive outcomes (gains and non-gains). Self-regulation with a prevention focus is associated with security and safety needs and is reflected in a concern for the presence or absence of negative outcomes (losses and non-losses) (Higgins, Reference Higgins1997). In the literature, regulatory focus has been studied as a situationally induced orientation (e.g., experimentally manipulated by framing health communications, Cesario, Grant, & Higgins, Reference Cesario, Grant and Higgins2004) and as a chronic individual-difference variable assessed with dispositional measures (e.g., Higgins et al., Reference Higgins, Friedman, Harlow, Idson, Ayduk and Taylor2001). To assess chronic regulatory focus, several studies in the health research area have used general measures of regulatory focus (e.g., Fuglestad et al., Reference Fuglestad, Rothman and Jeffery2008; Joireman, Shaffer, Balliet, & Strathman, Reference Joireman, Shaffer, Balliet and Strathman2012). Because the health domain involves specific situations that are not necessarily captured by general measures of regulatory focus, Gomez, Borges, and Pechmann (Reference Gomez, Borges and Pechmann2013) and Ferrer et al. (Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017) have recently highlighted the relevance of capturing chronic regulatory foci in the specific health context to study health behaviors. Chronic health regulatory foci measures have thus been developed. A chronic health promotion focus measure is supposed to capture concerns for improving the health state or attaining health-related gains. On the other hand, a chronic health prevention focus measure is supposed to capture concerns for protecting health state or avoiding health-related losses (Ferrer et al., Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017; Gomez et al., Reference Gomez, Borges and Pechmann2013). Gomez et al. (Reference Gomez, Borges and Pechmann2013) examined the influence of these two chronic health regulatory foci on the adoption of several health behaviors such as health information seeking, consumption of organic and functional foods, number of physician and pharmacist visits or drug consumption. More recently, Ferrer et al. (Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017) examined the influences of these two chronic health regulatory foci on intentions to perform several health behaviors including regular exercise.

The first aim of the present study was to extend Ferrer et al.’s (Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017) study on regular exercise intention by examining for the first time in the literature the influences of chronic health promotion and prevention regulatory foci on the amount of sports practice. According to Rothman, Wlaschin, Bartels, Latimer, and Salovey (Reference Rothman, Wlaschin, Bartels, Latimer, Salovey and Elliot2008), sports practice, balanced diet or sunscreen are among the health behaviors that promote health. These authors suggest that when people think about engaging in sport, they are more likely to attend to the presence or absence of favorable outcomes than to the presence or absence of unfavorable outcomes. Sports practice, being an activity that sustains a positive-outcomes approach could consequently fit better with a health promotion focus. One of the positive effects of fit is that people engage more strongly in what they are doing (Higgins, Reference Higgins2005). Thus, sports practice could be easier to undertake for people who experience a fit (Aaker & Lee, Reference Aaker and Lee2006). Moreover, while sports practice can lead both to immediate (e.g., pleasure) or distal positive consequences (e.g., general health), Hall and Fong (Reference Hall and Fong2007) have nevertheless evidenced that this health behavior has the particularity of involving an intertemporal trade-off by accruing present costs to achieve delayed rewards. In particular, those authors examined the perceived temporal proximity of benefits associated with physical activity and show that benefits become salient the equivalent of several hundred hours later (2007, p. 11). Fuglestad et al. (Reference Fuglestad, Rothman and Jeffery2008) also suggested that engaging in a regular exercise program will probably entail a longer initiation phase to fully master the new, effortful behaviors and to achieve positive outcomes. Thus, the importance that individuals attach to the future consequences of their behavior influences their willingness to engage in this health behavior (e.g., Gellert, Ziegelmann, Lippke, & Schwarzer, Reference Gellert, Ziegelmann, Lippke and Schwarzer2012). Promotion-focused individuals tend to adopt a distal temporal perspective whereas prevention-focused individuals tend to adopt a proximal temporal perspective (Pennington & Roese, Reference Pennington and Roese2003). Joireman et al. (Reference Joireman, Shaffer, Balliet and Strathman2012) also showed that a chronic general promotion focus was positively related to consideration of future consequences whereas a chronic general prevention focus tends to be positively related to consideration of immediate consequences. Thus, sports practice seems more in line with the distal temporal perspective of promotion-focused individuals. In addition, Ferrer et al.’s (Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017) results evidenced that health promotion focus was positively related to the intention to exercise regularly whereas health prevention focus was negatively related to this variable. Based on these theoretical supports and Ferrer et al.’s (Reference Ferrer, Lipkus, Cerully, McBride, Shepperd and Klein2017) empirical results, we thus expected that health promotion focus should be positively related with the amount of sports practice whereas health prevention focus should be negatively related with this variable.

Furthermore, while Ferrer et al.’s results (2017) support the idea that health promotion focus seems to be more favorable to physical activity, they do not inform us about the processes explaining this association. Thus, the second aim of this study was to explore for the first time in the literature a potential process behind the positive relationship expected between health promotion focus and sports practice. This process issue is important, both from a theoretical standpoint of acquiring, a clear and detailed understanding of how health promotion focus acts on sports practice, and from an applied standpoint of knowing how to support and facilitate sports practice. We proposed that the general set of resource allocation strategies SOC (Selection, Optimization and Compensation) could play a role in this process. Indeed, sports practice is a part of lifestyle behaviors which implies in individuals a self-management to overcome the perceived obstacles to practice (e.g., lack of time, family constraints, work – Salmon, Owen, Crawford, Bauman, & Sallis, Reference Salmon, Owen, Crawford, Bauman and Sallis2003; Gómez-López, Granero Gallegos, & Baena Extremera, Reference Gómez-López, Granero Gallegos and Baena Extremera2010). Now, the SOC model (Baltes & Baltes, Reference Baltes, Baltes, Baltes and Baltes1990) evidences a set of four interrelated action regulation strategies that help individuals to self-manage in an optimal way: Elective selection (i.e., developing and committing to a hierarchy of personal goals), loss-based selection (i.e., changes in the goal or the goal system in response to loss), optimization (i.e., engaging in goal-directed actions and means), and compensation (i.e., acquiring alternative means or support in response to a loss). For Freund and Baltes (Reference Freund and Baltes2002), SOC is “viewed as a coordinated ensemble of contextualized processes. Therefore, S, O, and C are considered conjointly and in their interaction for fully understanding their role in positive development” (p. 643). In the health domain, several studies have evidenced that use of SOC strategy has a positive influence on sports practice in a rehabilitation context (Ziegelmann & Lippke, Reference Ziegelmann and Lippke2007), on maintenance of sports practice among older adults (Gellert, Ziegelmann, Krupka, Knoll, & Schwarzer, Reference Gellert, Ziegelmann, Krupka, Knoll and Schwarzer2014) and on long-term adherence to a sport program among older women (Evers, Klusmann, Ziegelmann, Schwarzer, & Heuser, Reference Evers, Klusmann, Ziegelmann, Schwarzer and Heuser2012). The adoption of sports practice involves an initial motivation process that results in intention formation and subsequent self-regulation processes that address the pursuit of these goals (Schwarzer, Reference Schwarzer and Schwarzer1992). SOC strategy has been found to act as a mediator variable between planning (i.e., plans on ‘when’, ‘where’, and ‘how’ to perform a behavior) and exercise behavior (Ziegelmann & Lippke, Reference Ziegelmann and Lippke2007). It is thus proposed that SOC strategy constitutes an important step in the process leading to sports practice.

Furthermore, SOC strategy calls for flexibility for development of new goals in response to difficulties, refinement of goal-relevant means, and acquisition of new skills/resources (Freund & Baltes, Reference Freund and Baltes2002). Individuals with a promotion focus tend to approach desired end-states by engaging in flexible, explorative cognitive processing (Crowe & Higgins, Reference Crowe and Higgins1997). In addition, SOC is a strategy which serves a goal of achieving success (Baltes, Wynne, Sirabian, Krenn, & de Lange, Reference Baltes, Wynne, Sirabian, Krenn and de Lange2014), which is a goal congruent with promotion focus. Indeed, individuals with a promotion focus are attuned to emotions relating to the successful pursuit of positive outcomes (Higgins, Roney, Crowe, & Hymes, Reference Higgins, Roney, Crowe and Hymes1994). Thus, promotion focus and SOC strategy are matching because they are goal-congruent. Finally, Baltes et al.’s (Reference Baltes, Wynne, Sirabian, Krenn and de Lange2014) results showed that a chronic general promotion focus was positively related to the adoption of SOC strategy at work. Based on these theoretical supports and Baltes et al.’s (Reference Baltes, Wynne, Sirabian, Krenn and de Lange2014) empirical results, we hypothesized thus that chronic health promotion focus should also be positively related to the adoption of SOC strategy in sport. Moreover, given the link expected on the one hand between health promotion focus and SOC strategy, and the link evidenced in the literature between SOC strategy and sports practice on the other hand, SOC strategy was hypothesized to mediate the positive relationship between health promotion focus and amount of sports practice.

In sum, the aims of the present study were twofold: (a) To explore the direct links between promotion and prevention health regulatory foci and amount of sports practice; and (b) to examine whether SOC strategy has a mediating role in the relationship between health promotion focus and amount of sports practice. These links will be examined among a large sample of sport practitioners. In addition, as age, gender, health condition and education level may affect physical activity practice (e.g., Shaw & Spokane, Reference Shaw and Spokane2008), the effects of these variables were controlled in the present study.

In accordance with the above-mentioned frameworks, we formulated the following hypotheses:

Method

Participants

To be eligible for the study, participants should practice a sport. A total of 588 French volunteer participants took part in the study. Of the initial 588 participants, 75 were excluded because of missing data. The responses of 513 participants (51.7% men, 56.5% healthy) aged from 18 to 82 years (mean age = 36.96, standard deviation = 16.33, median score = 33, skewness value = .74) were consequently considered in this study. Most of the participants (97%) had completed secondary education. Demographic characteristics and health condition of the participants are presented in Table 1. All participants were recruited from various organizations, clubs and leisure centers and represented different sports, namely gym-based exercise (23%), ball sports (21%), racquet sports (17%), running (14%), walking (10%), nature sports (6%), swimming (3%), cycling (2%), fighting sports (2%) and yoga (2%). Furthermore, a high proportion of our sample had a relatively high level of sports practice in comparison with the amount of sports practice of the French population (European Commission, 2018): 50.7% of the participants practiced with high intensity, 91.6% practiced regularly, and 79.3% performed sports sessions for 60 minutes and more. The practice characteristics of the sample are presented in Table 2.

Table 1. Demographic Characteristics and Health Condition of the Sample

Note: Unhealthy participants could suffer from one or several diseases

Table 2. Practice Characteristics of the Sample

Note: Total score of amount of sports practice is the sum of scores of intensity (from 3 = high intensity to 1= slight intensity), frequency (from 4 = every day to 1 = irregularly) and duration (from 4 = 60 minutes and more to 1= less than 15 minutes).

Procedure

Data were collected anonymously via a cross-sectional, online self-report survey. The protocol met France’s ethical requirements for human research. Permission to conduct the study was first obtained from the administration. All participants provided informed consent prior to data collection.

Measures

Health regulatory focus. Gomez et al.’s (Reference Gomez, Borges and Pechmann2013) French health regulatory focus questionnaire was used to assess participants’ health regulatory focus. The questionnaire is composed of five items assessing health promotion focus (e.g., “I do not hesitate to embrace new experiences if I think they can improve my health”), and three items assessing health prevention focus (e.g., “I frequently think about the health problems I may have in the future”), presented in a random order. Participants responded on a scale from 1 = “not at all true of me” to 6 = “very true for me”. A Confirmatory Factor Analysis (CFA) was performed on the covariance matrix and the solution was generated by using maximum likelihood estimation. The two-factor model delineated by Gomez et al. (Reference Gomez, Borges and Pechmann2013) did not provide an acceptable fit to the data (RMSEA = .101; GFI = .947; CFI = .935). The health prevention focus item (i.e., “When I implement a health behavior, it’s because I want to protect myself from getting sick.”) exhibited high modification indices as well as moderate loadings on both factors. These results were similar to those obtained by Schmalbach et al. (Reference Schmalbach, Spina, Steffens-Guerra, Franke, Kliem, Michaelides and Zenger2017). In line with these authors’ procedure, a model excluding this item was tested and showed an improvement over the model previously examined (RMSEA = .089; GFI = .965; CFI = .959). Internal consistency was satisfactory for both the health promotion focus (α = .78) and the two-item health prevention focus (α = .77) subscales. Average scores were thus computed for each health regulatory focus subscale.

SOC strategy. Ziegelmann and Lippke’s (Reference Ziegelmann and Lippke2007) questionnaire, a version of Freund and Baltes’ (Reference Freund and Baltes2002) original questionnaire adapted to the physical activity context, was used to assess participants’ SOC strategy. The items were translated into French and slightly adjusted. The scale was composed of the stem “In general in my sport activity...” which was followed by three items assessing optimization strategy (e.g., “... I do everything possible to make my objectives come true”), and three items assessing elective selection strategy (e.g., “… I define my goals exactly and stick to them”). The second stem, “When it is getting more difficult in my sport activity...,” was followed by three items assessing compensation strategy (e.g., “... then I increase my efforts even more to continue to practice my sport activity”), and three items assessing loss-based selection strategy (e.g., “... then I temporarily reduce the practice of my sport activity”). Participants responded on a scale from 1 = “not at all true of me” to 6 = “very true for me”. A CFA was performed on the covariance matrix and the solution was generated by using maximum likelihood estimation. In line with Freund and Baltes (Reference Freund and Baltes2002), who proposed the SOC model as an integrated system, a hierarchical model of SOC strategy was constructed comprising a higher-order factor (SOC strategy) and four first-order factors (loss-based Selection, elective Selection, Optimization and Compensation). The model provided an acceptable fit to the data (RMSEA = .088; GFI = .927; CFI = .957) and internal consistency was satisfactory (α = .71). An average score was then computed for SOC strategy.

Amount of sports practice. Three items adapted from the sport and leisure activities subscale of the French Dijon Physical Activity Score (Robert et al., Reference Robert, Casillas, Iskandar, D’Athis, Antoine, Taha and van Hoecke2004) were used to assess amount of sports practice. This questionnaire is appropriate to assess usual sports practice in samples including elderly subjects aged up to 93 (Emile, Chalabaev, Stephan, Corrion, & d’Arripe-Longueville, Reference Emile, Chalabaev, Stephan, Corrion and d’Arripe-Longueville2014). One item assessed sport activity intensity (i.e., “The sport activity that you practice is of: (3) High intensity: significant muscular fatigue, to (1) slight intensity: Without muscular fatigue”). One item assessed sport activity frequency (i.e., “You are used to practicing your sport activity: (4) Every day, to (1) irregularly”). One item assessed sport activity duration (i.e., “The average duration of your sessions for sport activity is: (4) 60 minutes and more, to (1) less than 15 minutes”). In line with Robert et al.’s (Reference Robert, Casillas, Iskandar, D’Athis, Antoine, Taha and van Hoecke2004) procedure, the scores for the three items were added. The possible total score ranges vary between 3 (minimum) and 11 (maximum). The total score of sport and leisure activities subscale of this questionnaire has been shown to be reproductible in test-retest and to be correlated with peak VO2 and maximal power attained during effort test on a cycle ergometer and in walking tests (Gremeaux et al., Reference Gremeaux, Lemoine, Fargeot, D’Athis, Beer, Laurent and Casillas2008; Robert et al., Reference Robert, Casillas, Iskandar, D’Athis, Antoine, Taha and van Hoecke2004). The distribution of the sample on the scores of this measure is presented in Table 2.

Health condition. Consistent with prior research (e.g., Kempen, Jelicic, & Ormel, Reference Kempen, Jelicic and Ormel1997), a checklist of 18 diseases was used. Participants were asked to report whether they currently suffered from diseases diagnosed in the last 12 months. The total number of diseases diagnosed was used as a health condition index. The index ranged from 0 (none of the diseases) to 18 (all of the diseases).

Educational level. Participants indicated their highest educational qualification attained from 1 (no qualification) to 10 (PhD).

Statistical analysis

A model for evaluating the combined contribution (direct and indirect effects) of each variable—the promotion and prevention regulatory foci, SOC strategy, age, gender (coded as 0 for women and 1 for men), educational level and health condition—on amount of sports practice was tested. This model tested was built from our hypotheses. A covariance matrix, using Statistica-12, was generated from raw data (observed scores in each variable). A path analysis using Lisrel 9.1 (Scientific Software International) was used to examine the fit of this first model. The .05 level of significance was used for all statistical hypothesis testing. Beta represents the standardized regression coefficient. We used the model fit tests using conservative cutoffs for root-mean-square error of approximation (RMSEA) and comparative fit index (CFI). Finally, a bootstrapping method resample (Preacher & Hayes, Reference Preacher and Hayes2008) was employed to test the significance of the mediation effect of SOC strategy on the relationship between health promotion focus and amount of sports practice. Bootstrapping was set at 5000 samples with bias-corrected 95% confidence intervals.

Results

Descriptive statistics and correlations

Means, standard deviations and Pearson’s correlation coefficients for the variables are presented in Table 3. Furthermore, Khi–2 tests were conducted to examine gender differences. Only the amount of sports practice was impacted by gender. Men reported significantly higher amount of sports practice than women, χ^{Footnote 2}(8) = 45.67, p < .001.

Table 3. Descriptive Statistics and Matrix of Pearson r Correlation Coefficients among the Variables (N = 513).

Note: M= Mean; SD = Standard Deviation; ASP = Amount of Sports Practice.

* p < .05. **p < .01.

Hypothesized model path analysis

The results from the hypothesized model presented an acceptable fit (χ^{Footnote 2} = 36.15; df = 5; p < .001; RMSEA = 0.110; CFI = 0.94). Considering the exploratory nature of this hypothesized model, a second model (adjusted model) was then tested including modification indices^{Footnote 1} proposed by Lisrel. The second model tested, adjusted to the specific data, presented an excellent fit (χ^{Footnote 2} = 16.64; df = 5; p < .01; RMSEA = 0.067; CFI = 0.98). This model explained 24% of the variance of amount of sports practice. The results from the adjusted model, detailed below, are presented in Figure 1.

Figure 1. Path Diagram of Adjusted Model

ASP = Amount of Sports Practice. Standardized path coefficients are presented. Non-significant paths are not included in the figure.

*p = .05. **p < .01. ***p < .001.

Direct effect. Amount of sports practice was positively related with health promotion focus (β = .13, p < .01), SOC strategy (β = .28, p < .001), and negatively related with health prevention focus (β = –.20, p < .001).^{Footnote 2} Moreover, amount of sports practice was positively related with gender (β = .28, p < .001), weakly negatively related with health condition variable (β = –.08, p = .052), and not significantly related with educational level variable.

Indirect effect. SOC strategy was positively related with health promotion focus (β = .39, p < .001) and negatively related with age (β = –.24, p < .001).

Bootstrapping analyses indicated that the indirect contribution of health promotion focus to amount of sports practice through SOC strategy was significant, with a point estimate of .040 and a bias-corrected bootstrapped 95% CI from .13 to .29. Therefore, since zero was not contained in the bootstrapped CI (see Preacher & Hayes, Reference Preacher and Hayes2008), SOC strategy may be considered a significant partial mediator of the relationship between health promotion focus and amount of sports practice, while controlling for educational level, gender and health condition.^{Footnote 3}

Additional sub-components of SOC strategy path analyses

As Freund and Baltes (Reference Freund and Baltes2002) consider the SOC model an integrated system, the preceding analyses examined the composite score of SOC strategy as mediator of the link between health promotion focus and amount of sports practice. However, an existing empirical study suggests that specific strategies are more beneficial for exercise than others (Evers et al., Reference Evers, Klusmann, Ziegelmann, Schwarzer and Heuser2012). Thus, it is possible that sub-components of SOC strategy are more or less involved in this process. This possibility was not the central interest in the present work, but an ancillary analysis was conducted to examine it. Four path models examining the mediating role of each of the four sub-components of SOC strategy in the positive link between health promotion focus and amount of sports practice were thus successively run. Age,^{Footnote 4} gender, health condition and educational level effects were also controlled in these models. Results of these analyses confirmed the mediating roles of elective selection, optimization and compensation in the positive link between health promotion focus and amount of sports practice. Specifically, elective selection and compensation partially mediated the link between health promotion focus and amount of sports practice, and optimization totally mediated this link. On the other hand, loss-based selection was not associated with health promotion focus and was negatively associated with amount of sports practice. Fit indices, beta coefficients and bootstrapped CI for each model are shown in Table 5.

Table 4. Summary of the Fit and Pathways of Hypothetical Model of ASP for each Age Group

Note: ASP = Amount of Sports Practice. In each model gender, health condition and educational level effects were controlled.

Table 5. Summary of the Fit and Pathways of Models of ASP for each Sub Component of SOC

Note: ASP = Amount of Sports Practice. In each model gender, educational level and health condition effects were controlled.