Participation in physical activity (PA) is linked to a variety of health benefits [1]. However, approximately 1.4 billion people are still insufficiently active according to a new global survey [2]. More specifically, in Europe, namely, in Spain, physical inactivity affects 27% of the adult population [2]. Physical inactivity is a leading risk factor worldwide for morbidity and mortality, and is a significant economic burden [3]. In light of this evidence, PA promotion has become a critical public health goal for policy-makers and practitioners.
Recent research aims to understand and motivate people to engage in regular physical activity (PA) by using health behavior and behavioral change theories as a framework for intervention [4]. Evidence suggests that theory-based interventions are more effective in promoting adherence to PA than those that are not [5]. However, some studies have reported only small to moderate effect sizes for theory-based interventions (e.g., E.S. =.12-.156, E.S. =.357), leading to debates about their appropriateness [8–10]. To enhance intervention effectiveness, it is essential to identify relevant mediators that can help determine why a theory-based intervention is successful or not [11].
Mediators are intervening variables between an independent variable (e.g., intervention) and a dependent variable (e.g., change in PA) [12]. Relevant mediators in interventions should lead to higher PA levels while redundant ones should be discarded [11]. Understanding the relevant mediators of PA opens new possibilities for theoretical integration, reducing redundancy and exploiting each theory’s strengths [13,14]. Structural equation modelling (SEM) provides a flexible framework for performing mediation analysis and estimating latent variables based on theory [15]. However, reviews have reported that SEM is more the exception than the rule [11,12]. More innovative studies are needed to focus on key mediators of PA to improve future interventions [15,16–18]. In this regard, more emphasis is needed on psychological constructs due to their significant impact on PA compared to other factors (e.g., biological, environmental, and behavioural) [16]. In this context, body image (BI) has received significant attention, as it is likely to play a role in PA behaviour [19].
Body image and physical activity
The term body image (BI) refers to a multidimensional construct that consists of three components: (1) a behavioural component involving body-related behaviours (i.e., checking behaviours such as fixating on body parts in the mirror); (2) a perceptual component involving the perception of body characteristics (i.e., estimates regarding the size and shape of their body and body parts); and (3) a cognitive-affective component (i.e., how people think and feel about their bodies) [20]. Body image dissatisfaction (BID) is a related concept defined as a person’s persistent negative attitudes and feelings regarding his or her physical appearance [20]. BID is widespread, with prevalence ranging from 11–72% in women and 8–61% in men [21]. More recently, a survey found that 20% of adults experienced shame, 34% felt down or low, and 19% felt dissatisfied with their BI [22]. Furthermore, BID has a significant impact on weight status so that people who are overweight or obese are at higher risk of BID [21]. On the other hand, BID is strongly associated with an impaired health-related quality of life [23]. Therefore, BID has become a global public health concern due to its increasing prevalence and potentially severe effects on human health [24].
Given the above, to implement interventions to reduce BID prevalence, a large body of research in the PA domain exists. In this regard, different systematic reviews and meta-analyses over the years have concluded that PA significantly changes the physical self while also increasing self-esteem, all of which leads to a more positive BI [25,26]. However, most research has focused on how PA interventions can promote positive BI rather than on the reverse association (BI effects on PA onset/maintenance) despite being a bidirectional relationship [27–30].For example, a recent review from 1990 to 2019 (N = 155) found only one intervention targeting BI and testing the effect on adherence to PA [31]. Similarly, another review of more than 210 studies from 2008 to 2018 detected the same pattern in the literature and concluded that “this focus precludes an understanding of BI as a predictor of PA and/or sports participation” [19] (p.6). Thus, although it is plausible to suggest that negative BI is associated with lower PA levels and that positive BI is associated with higher PA levels, the literature is limited, and there is no full understanding of whether BI is a barrier or a facilitator [19,32–34].
Otherwise, despite theoretical frameworks being crucial for PA research, as previously stated, research on the relationship between PA and BI has been largely atheoretical to date. For example, Martin-Grinis, Basset, & Conlin35 pointed out that only half of the revised studies (N = 12) employed a theoretical framework. Similarly, Sabiston et al. [19] reported that only 33.5% of the studies were guided by an identifiable theoretical framework, limiting any “understanding of mechanisms that could be targeted in interventions” [19] (p.6). In this context, it has been argued that a person's motivation could be particularly informative for explaining the relationship between BI and PA and hence serve as a mediating variable framework [34]. To this end, self-determination theory (SDT) is a prominent theoretical framework of human motivation that can be advantageous to apply within the field of BI because it specifies the various motivational regulations that underpin PA behaviour, thereby providing potential targets for tailored PA interventions [19,26,31,34].
Self-determination theory and physical activity
SDT has been extensively investigated in the PA domain. A prominent systematic review of more than 60 studies from 1960 to 2011 found a positive association between more autonomous forms of motivation (i.e., identified and integrated regulations) and PA behaviours, in terms of frequency, duration, and intensity, in a variety of populations and settings [36]. In contrast, they found no association or a negative association between PA behaviour and the controlled forms of regulation (i.e., external and introjected regulations). Additionally, variables such as sex, BMI, and age have been shown to influence motivational regulations. Regarding sex, some authors suggest that more controlled regulations predict PA behaviour for men, whereas more autonomous regulations predict PA behaviour for women [37]. These findings highlight the importance of considering the role of individual differences when studying SDT within the PA field [36].
Self-determination theory and body image
As previously stated, BID is extended in our modern societies. Women are thought to adopt a self-objectified view of themselves and to be evaluated on the basis of their appearance [20]. According to SDT, these beauty standards, which are an external demand or a socially constructed contingency, can be equivalent to a form of social pressure that fosters more controlled regulations (i.e., external or introjected) and drives women to engage in compulsory or inconsistent PA [38]. Consequently, this focus diminishes intrinsic motivation, where activity is done for the inherent feeling of enjoyment and satisfaction [36]. This has important consequences on people's physical and mental wellbeing, leading to drastic strategies such as extreme dietary restraint and ultimately EDs [38]. However, despite their importance, very few studies have tested motivational regulations as a mediating variable framework for the effects of BI on the onset/maintenance of PA [39].
Relationship between body image, self-determination theory, and physical activity
In a series of studies, researchers have examined the mediating role of motivational regulation in the relationship between body size discrepancies, BID and PA participation. Markland [40] found that self-determined motivation regulations (i.e., identified and intrinsic) mediated the relationship between body size discrepancies and PA behavior among women. Brunet et al. [39] similarly found that motivational regulations partially mediated the relationship between body size discrepancies and PA participation in young women. More and Phillips [41] extended this research to include the role of sex and found that BID was related to less frequent cardiovascular and strength-based activity in both men and women due to lower levels of intrinsic regulation. Women's relationship between BID and PA was partially mediated by controlled regulations.
Overall, the body of research that tests explicit hypotheses derived from SDT has recently been described as a ‘cutting edge’ [42] and ‘emergent’ topic in the BI field [26]. Despite its importance and the potential implications for evidence-based practice, it is important to highlight some limitations. First, most of the aforementioned studies relied on body size discrepancies (between actual and ideal body size) as a proxy for BID. However, some authors have argued that body size discrepancies do not always reflect BID, nor do they indicate the magnitude of any dissatisfaction [43]. Therefore, it is recommended to evaluate BID with a special focus on characteristics such as body appreciation and the value of appearance [44]. In fact, it has recently been argued that the mediating role of motivational regulations in the relationship between BI and PA has not been explicitly tested and is therefore “a potential avenue for further research” [34] (p.161). Relatedly, it has been recommended that function-related aspects of BI are also important to measure since evidence suggests that some people place greater emphasis on function (what the body can do) than appearance (how the body looks) [35]. Second, most studies in the field had nonrepresentative samples that mostly focused on young women, discarding other subpopulations. This trend is consistent with prior epidemiological research that reports BID to be an ‘epidemic’ among young women [21]. Nonetheless, more study focusing on diverse samples, such as men, is required because BID affects these subpopulations as well [39]. Finally, research on this topic has relied on standard regression methods for mediation analysis [39,41,44]. Future research would benefit from adopting an SEM framework as a multivariate statistical technique [45].
The present study
In light of the evidence, within an SEM framework, this study aimed to elucidate the relationship between BI and PA, considering the possible mediating effect of motivational regulations using a large representative sample of the Spanish population. As a secondary aim, given the importance of individual differences, and in particular sex, as previously stated, we conducted a moderation analysis to elucidate whether the proposed model is similar for men and women (see Fig. 1). Given the above, the following hypotheses were proposed:
Hypothesis 1
(H1): A positive BI will show significant positive direct effects on PA participation.
Hypothesis 2
(H2): More autonomous forms of motivation (i.e., intrinsic, introjected and identified) and lower levels of controlled regulations (i.e., external and introjected) will show significant positive and negative indirect effects between BI and PA participation.
Hypothesis 3
(H3): Men will be influenced by a more autonomous form of motivation (i.e., intrinsic, introjected and identified), and women will be influenced by more controlled regulations (i.e., external and introjected) with regard to the type of motivational regulations as a mediator of the relationship between BI and PA participation.