Previous studies in the area of leisure constraints reported weak negative relationships or, in some cases, non-relationships between perceived constraints and participation in sport and leisure activities [34, 42]. Jackson et al.  proposed that these weak relationships might be due to the negotiation of leisure constraints. Although everybody faces and reports constraints, there are individuals who overcome them with the development of successful negotiation strategies. A variety of behavioral variables have been suggested in the literature to influence the success of these negotiation strategies, such as motivation and personality . In the current study, we aimed to test if attitudes and perceived behavioral control interact with constraints and intervene within their relationship with intention to continuing physical activity participation. Based on the negotiation proposition , we aimed to test if the constraints that older individuals report have a direct impact on intention, or an indirect one through their impact on perceived behavioral control and attitudes. We would expect that individuals who feel that they can control their behavior and/or will have positive attitudes toward physical activity would be more likely to overcome the influence of constraints.
The results of the study revealed some clear patterns. First of all, the element of subjective norms was not significantly associated with older individuals’ intention to participate in physical activity programs. This finding supports previous studies [12, 24, 43]. It seems that older individuals are somewhat independent on their decisions regarding physical activity participation. This might be due to the organized nature of the physical activity programs. All the individuals who participated in the research were members of KAPI and participated in the programs provided by KAPI. It is not clear if the results would have been the same if the question had been about physical activity participation in an unorganized setting. Previous studies have suggested that middle-aged and old individuals usually follow some stereotypes, do not always feel comfortable to participate in exercise and recreational programs, and are influenced by social contacts .
In terms of the relationship between constraint dimensions and the elements of the theory of planned behavior, the results revealed significant statistical associations among the variables. Both attitudes and perceived behavioral control were shown to be significantly correlated with all the constraint dimensions. As expected, all the correlations were negative. These results support the study conducted by Alexandris and Stodolska .
In terms of the role of perceived behavioral control and attitudes as mediators of the relationship between constraints and intention, some clear patterns were also revealed. First of all, both the variables were shown to be mediators, which suggests that constraints influence intention both directly and indirectly through their negative effects on attitudes and perceived behavioral control. Subsequently, individuals who face or perceive constraints are more likely to overcome them if they have developed positive attitudes toward physical activity and/or feel in control of their behavior. In terms of the strength of the mediation, perceived behavioral control was shown to be the strongest mediator. These findings support the negotiation proposition  and further extend it by suggesting two more behavioral variables (attitudes and perceived behavioral control) that could be included in future constraint models. As previously discussed, so far, only motivation had been shown to interact with constraints [10, 28].
In terms of the impact of the attitudes and perceived behavioral control on the specific constraint dimensions, the results indicated that attitudes fully mediated the influence of accessibility and facilities dimensions on intentions, and partially mediated the relationship between psychological and partners constraints and intention. This suggests that the psychological and lack-of-partners constraints have also a direct impact on individuals’ intention to continuing physical activity participation. It should be noted that psychological constraints are perceived intrapersonally [18, 33], whereas lack-of-partners constraints are perceived interpersonally. Previous studies have reported intrapersonal constraints as the most powerful predictor of behavior [8, 33].
On the other hand, perceived behavioral control had a strong mediation effect on intrapersonal, interpersonal, and structural constraints. Whereas the mediation effect of perceived behavioral control on structural constraints was in a degree expected, considering previous studies in the area of leisure constraints , the mediation of perceived behavioral control on intrapersonal constraints was not expected to be so strong. Intrapersonal constraints have been found to be the most important blocking factors of behavior ; it seems that the alteration of perceptions of control might minimize the inhibiting role of intrapersonal constraints and enhance intention and, subsequently, participation behavior. These results support previous studies [12, 24, 43] suggesting that perceived behavioral control is the most powerful predictor of intention and behavior within the theory of planned behavior. This proposes that practitioners should target perceived behavior control to reduce its effects. Consultation, education, psychological support, design, and delivery of appropriate programs are among the strategies that should be applied to reduce the influence of perceived behavioral control.
In conclusion, the study provided evidence that attitudes and perceived behavioral control interact with intention, and this interaction determines, in a large degree, older individuals’ decision to continue taking part in physical activities. On the other hand, the subjective norm component was not shown to be significantly related with constraints. Furthermore, the results indicated that both attittudes and perceived behavioral control strongly mediated the relationship between constraints and intention, with the behavioral control component being the strongest mediator.
Study limitations and future research
It should be pointed out that the study was a cross-sectional one, which means that we were not able to trace individuals’ future participation levels. The model tested was only based on individuals’ intention to continue participating in physical activities. Future studies could adopt a longitudinal approach to test a model based on actual physical activity levels. This would give researchers the chance to further explore the role of constraints within individuals’ decision making, considering also the “negotiation” construct proposed by Jackson . Furthermore, the sample of the study was not large enough to ensure generalizations of the results. The model examined in the present study should be verified with larger samples in which detailed demographic information should be reported to test possible influences of socio-economic groups on the constraints–intention relationships.
In the present study, we tested the mediation effects of attitudes on the relationship between constraints and intention using regression models, following Baron and Kenny’s  guidelines. Because we did not have an established theoretical framework for the relationships between constraints and the components of the theory of planned behavior, we did not apply structural modeling, which can be a suggested methodological approach for future studies that will aim to test the model proposed in the present study. It would also be interesting to test the model on different age groups and investigate possible changes on the perception of constraints and the influence of perceived behavioral control on intention with advancing age.