Abstract

Introduction

Even rewarding activities have an addiction potential (Allegre, Souville, Therme & Griffiths, 2006; Terry, SzaboGriffiths, 2004). Exercise, for example, may become maladaptive if taken to extremes (Ackard, Brehm & Steffen, 2002; Hall, Kerr, Kozub & Finnie, 2007; Hamer & Karageorghis, 2007). Indicators of maladaptive exercise include working out several times a day, or for longer periods than recommended, obsessing over calories expended, anger if interrupted, cancelling or avoiding social or occupational responsibilities in order to exercise, and centring daily schedules around exercise (Adams & Kirkby, 2002; Terry et al., 2004). Terms applied to such behaviour include exercise dependence (Hamer & Karageorghis, 2007; Hausenblas & Symons Downs, 2002), over-exercising (Long, Smith, Midgley & Cassidy, 1993), heightened commitment to exercise (Yates, Edman, Grago & Crowell, 2001), obligatory exercise (Ackard et al., 2002; Hall et al., 2007), and exercise addiction (Annett, Cripps & Steinberg, 1995; Mathers & Walker, 1999). Importantly, exercise dependence is considered distinct from the behaviour of the ‘gym enthusiast’ by the symptoms of tolerance, withdrawal, lack of control, intention efforts, time, reduction in other activities, and continuance evident in the dependent individual (Berczik et al., 2012; de Coverley Veale, 1987; Freimuth, Moniz & Kim, 2011).

Exercise dependence has long been studied with reference to compulsive running among athletes (Chapman & DeCastro, 1990; Freimuth et al., 2011; Hall et al., 2007; Ogden, Veale & Summers, 1997) and in relation to eating disorders among women (Ackard et al., 2002; Mond, Hay, Rodgers & Owen, 2006). It can be a secondary addiction to an eating disorder, or occur without an eating disorder as a primary addiction. Generally, weight and shape concerns are lower in those with primary exercise dependence, as the purpose of exercise, when an eating disorder is present, is almost exclusively for weight and shape control (Freimuth et al., 2011).

Smith, Hale and Collins (1998) proposed that exercise dependence among bodybuilders is a specific exemplar of this behaviour. While weight training may be initiated for innocent reasons such as improving strength and fitness, over time it may take on an exaggerated importance. This is particularly likely in situations where physique and self-esteem are intertwined (Hamer & Karageorghis, 2007). Research concerning bodybuilding dependence has focused on socio-demographic, motivational and psychological correlates. For example, Smith et al. (1998) and Hurst, Hale, Smith and Collins (2000) determined that dependent persons were more likely to begin weight training to improve self-esteem. Stronger, leaner, muscular males, for example, may be per ceived as socially dominant, and more likely than their smaller counterparts to receive respect and admiration from their peers, which may foster an obsession with bodybuilding (Cafri, Van den Berg & Thompson, 2006; Smith & Stewart, 2012). Hurst et al. (2000) further noted that the social atmosphere and support gained through weight training were associated with dependence.

Psychologically, dependence has been linked to both muscle dysmorphia (Cafri et al., 2006; Giardino & Procidano, 2012; Smith & Hale, 2004) and lower life satisfaction (Smith & Hale, 2005). Further, a high drive for muscularity has been associated with the susceptibility to other addictions (Hale, Roth, DeLong & Briggs, 2010). Finally, bodybuilding dependence has been shown to be more likely among persons from lower socioeconomic backgrounds and those not currently involved in a romantic relationship (Smith & Hale, 2005). Given the growing literature concerning this phenomenon, the current study explored potential psychological, motivational, emotional and behavioural concomitants of bodybuilding dependence, with a focus on anger, hostility and aggression.

Methods

Results

Questionnaires were completed by 101 men (mean age = 30.6 years, SD = 9.3, range 18-67). Participants had an average weight training history of 10.1 years (SD = 8.3). Most indicated gymnasium attendance at least a few days a week (70.3%), with weight training most days they attended (88.1%), conducted at a moderate to moderately high intensity (72.1%) for between 30 and 90 minutes (70.3%) One third (34.3%) reported not being involved in a romantic relationship. The sample was relatively well educated, although 23.3% reported only secondary education or lower, with 21.2% reporting a trade qualification. Summary statistics for other study variables are shown in Table 1. A total BDS score was also computed, solely to derive a putative classification of dependence, defined as greater than one standard deviation above the mean (n = 18, 17.8%).

discussion

This study was predicated on the assumption that bodybuilding dependence is a specific, identifiable form of the general phenomenon of exercise dependence (Adams & Kirkby, 2002; Annett et al., 1995). While the current sample of undifferentiated males engaged in regular weight training does not allow a valid commentary on precise incidence figures for bodybuilding dependence, an impressive variability of scores was evident among the sample, with at least one participant recording the maximum score for each of social dependency, training dependency and mastery. The basic proposition that there would be positive and significant associations between bodybuilding dependence and anger, hostility, aggression, and stress, was largely supported at the bivariate level. While the pattern of relationships between the measures was not consistent, the evidence was suggestive of the potential relevance of these constructs to the dependent individual.

A unique advantage of the current study was the evaluation of motives for bodybuilding. From a targeted list of 21 possible reasons for weight training, three plausible motives were derived: personal challenge, physique anxiety and mood control. Of these, mood control shared the strongest positive associations with dependence. This is particularly salient given the study’s focus on anger, hostility and aggression. Interestingly, mood control was directly associated with hostility only, suggesting they may both be cognitive self-evaluations. Further, the placement of these motives in the multivariate models allowed a commentary on their unique contribution to dependence, after shared variance with other predictors had been accounted for.

The variable pattern of results provided an opportunity to comment on the discriminant validity of the dependence constructs. While social dependency assesses the desire to be in the bodybuilding environment, training dependency literally reflects dependence on the activity, and mastery distinguishes the desire to be in control of the training schedule (Hurst et al., 2000; Smith & Hale, 2004, 2005; Smith et al., 1998). In the current study social dependency was particularly relevant to variables such as personal challenge, hostility, anger and physical aggression. These findings concur with those of Smith et al. (1998) who underlined the link between social dependency and perceptions of self-worth. In a similar vein, training dependency was associated with both personal challenge and mood control. Perhaps these associations highlight the compulsive, driven behaviour of training dependent persons. In contrast, mastery was the only dependence category associated with physique anxiety. This suggests that mastery taps a distinctly unique component of exercise dependence compared with social and training dependency (Hurst et al., 2000). Nevertheless, based on the current data the overarching conclusion must be that there are more commonalities among the three dependence subscales than differences between them. It may be judicious therefore for future studies to consider measurement models that offer more discriminating quantifications of these constructs.

There was no evidence within the current sample that individuals reporting dependence were unique in terms of relationship status, education level, age, or years engaged in bodybuilding. Some of these results are contrary to those documented by Smith and Hale (2005). It is possible that these variant results are attributable to participant characteristics. Based on the description given for the previous sample (Smith & Hale, 2005), the current participants are likely to have been better educated, with a greater proportion in a steady relationship. Further, the comment has been made that much of the research regarding bodybuilding dependence has been invariant with regard to culture (Hurst et al., 2000). While the current study contributes to this deficit, the results may simply reflect the sociocultural idiosyncrasies of the Australian male (Huggins, 1998; Taylor, Stewart & Parker, 1998). Indeed, the very decision to study only males (Hurst et al., 2000; Smith & Hale, 2005) may also offer a source of variance concerning the potential correlates of bodybuilding dependence. More generally, there is the possibility of selection bias among our participants as they were a sample of convenience. Future research may therefore benefit from the use of a control group of non-dependent weightlifters or exercisers, to enable a more effective comparison of demographic characteristics.

While it was not possible to infer causal relationships among the variables measured due to the cross-sectional nature of the study, it was nevertheless appropriate to consider the potential for causality. This was driven both by the desire to provide a framework for hypothesis testing in future research, and to acknowledge the transactional hypothesis (Smith et al., 2004) which emphasises stress disposing towards hostility and anger, to the manifestation of outward aggression, potentially leading to dependence. On balance, with the exception of training dependency, the current models reflect this proposition. This is characteristic of the transactional model (Smith et al., 2004), but not necessarily exclusive to bodybuilding or bodybuilding dependence. A unique path in our models was, however, the negative coefficient between stress and physical aggression. This should be viewed with caution as it is quite possibly a statistical artefact arising, for example, from the multicollinearity among predictor variables (e.g., stress and hostility). An alternative proposition is perhaps that weight training provides an appropriate and socially acceptable avenue to express aggressive tendencies, thereby attenuating feelings of stress. It may also be the case that engagement in physical exercise becomes a coping skill, which may contribute to, or be a reason for, dependence (Berczik et al., 2012). Future research may benefit from investigating the contribution of more specific coping skills on dependence, such as problem-focused or emotion-focused coping.

As stated previously, the models discussed above are not a commentary on the way in which the measured constructs definitively co-relate, but rather represent a platform from which alternative models, perhaps with samples of varying composition, might be tested. Indeed, the order of the predictors we have examined is somewhat interchangeable. We reiterate that the data presented are designed to generate hypotheses rather than to evaluate definitive hypotheses. Future research might reasonably evaluate alternative causal orders and/or compare associations between anger, hostility, aggression, and stress among competitive bodybuilders, amateur weightlifters and persons who exercise for general fitness (Smith et al., 1998). Future studies may also usefully include both pre- and post-training measures of these variables, along with physiological markers such as serum endorphin levels (Hamer & Karageorghis, 2007). Such designs provide the opportunity to explore the proposition that weight training improves mood, and decreases stress levels. Indeed, unless such complex longitudinal studies are undertaken, the true causal nature of the relationships considered in the current study will not be definitively understood.

References