Resnicow and Vaughan [6
] present two key arguments why the prevailing behavior theories, such as the Theory of Planned Behavior, Social Cognitive Theory, and the Transtheoretical Model, are of limited use for healthful nutrition and physical activity promotion: (I) behavior change is influenced by such a complex set of interacting variables that it should be viewed as a chaotic system; (II) behavior change does not follow a linear pattern but rather occurs with 'quantum leaps'. I will briefly discuss both these issues, before addressing the consequences of the paradigm shift that Resnicow and Vaughan propose.
The complexity of nutrition and physical activity behavior change.
The fact that nutrition and physical activity behaviors are influenced by a complex set of interrelated variables is well recognized. First of all, as others and I argued in earlier contributions to the theory debate in the IJBNPA [4
], nutrition and physical activity are not behaviors but rather behavioral categories, i.e. complex collections of specific behaviors [7
]. Although the present day obesity epidemic suggests otherwise, people cannot eat an unlimited amount of food, and people have only a limited amount of time, so different specific nutrition behaviors and different sedentary and physical activity behaviors are not independent but may cluster, interact or compete, and each of these specific behaviors will have its own specific determinants.
I have argued, following a framework suggested by Rothshild [8
], that these determinants can be categorized in three groups: motivation, abilities and opportunities [4
]. Such a simplification in categories at least helps me to, for example, explain to students and health promotion practitioners that inducing behavior change is not merely a case of knowledge transfer, increasing risk perceptions, or otherwise convincing people that the pros of change outweigh the cons, but that teaching and practicing skills as well as environmental changes are often very much needed to get people to change their nutrition or physical activity behaviors. However, helpful as it may be, categorization of a complex set of determinants does not change the complexity and I very much agree with Resnicow that for many nutrition and physical activity behaviors, let alone public health-relevant behavioral categories, mapping a complete picture of mediation, moderation [10
], and clustering [11
] of all potential determinants is often impossible [5
]. Recently Glass and McAtee [13
] proposed a multilevel framework for the study of obesity-related health behaviors that also nicely illustrates the complexity of the issue. Resnicow posits that Chaos Theory truly recognizes such complexity and may therefore provide a better framework for understanding health behavior change.
Behavior change in quantum leaps instead of linear fashion.
Resnicow and Vaughan also argue that a key characteristic of the prevailing health behavior theories is invalid; the prevailing theories explicitly of implicitly treat behavior change as a linear event. According to Resnicow and Vaughan behavior change is much more likely to occur in 'quantum leaps'; 'mini-epiphanies' and 'tipping points' may occur that induce radical changes in behavior. Whether and when such dramatic changes occur are "virtually impossible to predict", and they may be heavily sensitive to initial conditions. Adopting this line of thinking makes nutrition and physical activity behavior change a random process.
I find Resnicow and Vaughan's contribution to the theory debate very useful. It forced me to rethink behavioral nutrition and physical activity theory and intervention approaches; based on the reactions to the theory debate key note at this year's ISBNPA conference, where Resnicow introduced his arguments, I was not the only one. Despite the strength of their arguments, I am not yet convinced. First of all, I want to see evidence. I agree that the scientific evidence for the prevailing health behavioral theories may be weak for many nutrition and physical activity behaviors. However, the direct empirical evidence for a chaotic nature of such behaviors or the quantum leap character of behavior change is absent. The authors do refer to evidence for quantum change in alcohol dependency behaviors, but such addictive behaviors may be different from nutrition and physical activity, i.e. more complex behavioral categories for which cessation is certainly not recommended. A next step to further build Resnicow and Vaughan's framework would therefore be to propose specific hypotheses to test the value of the framework. First of all, an empirical study to test whether nutrition and physical activity behavior change indeed occurs in quantum leaps, initiated by 'mini-epiphanies' and associated with 'tipping points' [6
] rather than in a linear fashion, is certainly possible.
Furthermore, there are empirically supported arguments why we should not disregard existing behavioral theories in order to embrace chaos.
First, certainly not all nutrition and physical activity behaviors are so complex that a Chaos Theory approach should be considered. For example, children's nutrition behaviors in schools are strongly dependent on what they can bring to school and on what is provided within schools; changing the school environment will directly influence what children eat at schools. Bere et al., for example, showed that cost free provision of fruits and vegetables in schools led to increased fruit and vegetable intakes among school children [14
]. Similarly, children's physical activity during school hours is strongly dependent on how many hours of physical education are provided, and increasing time for physical education does lead to increased physical activity levels [15
Second, the practical implications of the Chaos Theory paradigm seem limited, and I believe that Resnicow and Vaughan would agree that behavioral nutrition and physical activity interventions still ask for insight in people's motivations, abilities and opportunities, despite a possible chaotic nature of these behaviors. Resnicow and Vaughan propose some practical implications based on the chaotic and quantum character of health behavior change. They use the analogy of "ping-pong balls in a lottery machine" to illustrate what they mean: the role of health communications may be similar to spinning the lottery machine where ping pong balls that represent different motivations, abilities and beliefs in opportunities; health communicators should try to spin that machine periodically under various circumstances to improve the chances that the spinning will result in a mass-epiphany or a tipping point. But how should we spin the lottery machine? Should the spinning, i.e. the medium, strategy and content of health communications, be random? To me it still makes sense to have evidence-based insight in which motivations, abilities and self-efficacy beliefs are on those ping-pong balls, in order to target or tailor the machine spinning to this 'initial condition', and to improve the chances that the intervention is regarded as personally relevant which improves the chances that the health communications get sufficient attention [17