Curbing “Addictive” Behavior In Soft Drink Consumption

Humans, unfortunately, seem to have a propensity for addictive behavior. Whether it be binge-eating, alcohol abuse or even more abstract behavioral patterns such as gambling, people seem to have great trouble avoiding these seductive actions and the negative health consequences thereof. Even though many sufferers are motivated to reduce their addictive behaviors, the strong impulses associated with these behaviors make this very difficult.

This is one of the main reasons that psychologists have been working on interventions that stand a better chance of addressing these impulses directly. One of the first forays in this direction was what is termed approach-avoidance training. The idea of approach-avoidance training is predicated on the existence of two basic motivational systems, one concerned with approach, such as appetitive behavior and interpersonal attraction, and the other concerned with avoidance, such as fear or disgust responses. After researchers noted that patients with addiction problems, such as alcohol abuse patients, had a bias towards approaching addiction-related stimuli even when they clearly had the explicit motivation not to do so, they attempted to address this bias directly. In a seminal study, alcohol abuse patients were trained to symbolically avoid alcohol over several sessions, using joysticks to push away alcohol-related pictures and pull soft drink pictures towards themselves. The researchers found that training patients’ avoidance bias towards alcohol helped them to avoid relapse even a year after the intervention.

Clearly, this work suggests that approach-avoidance training may be an extremely powerful tool in helping individuals combating even tenacious undesired behavior. For this reason, more recent studies have attempted to extend the effect of approach-avoidance training to other domains, from social stereotype reduction to obesity prevention. Unfortunately, these efforts have not always succeeded in producing an effect. At the same time, more foundationally oriented research has attempted to illuminate the precise processes underpinning approach-avoidance training effects in order to predict when it will be most effective.

Our recent contribution to the science on this subject tried to connect these two approaches. We conducted experiments on approach-avoidance training, using separate unfamiliar soft drinks as the targets of both approach and avoidance behaviors. Such a design would have allowed us to tease apart the determinants of the approach-avoidance training effect without worrying about extraneous factors, such as participants’ prior experience with the target drinks or large objective differences between the approached and avoided categories. At the same time, we wanted to contribute to extending approach-avoidance training to more everyday behaviors such as unhealthy consumption by using sugary soft drinks as the target stimuli in a nonclinical sample.

However, despite conducting four experiments on a total of three hundred participants, we could not find any evidence that approach-avoidance training affected participants’ subsequent consumption of the two drinks, nor did their explicit liking or spontaneous evaluation tendency change. In essence, participants responded to the drink they had pushed away with a joystick in the same way as they did to the drink they had pulled towards themselves. This sobering result persisted whether participants were reminded of the meaning of their push-pull movements via semantic labels (i.e. having to pull the word “approach” in the same way as they pulled one of the target drinks) or via perceptual consequences (i.e. having a pulled drink zoom towards them), whether participants were allowed to taste the drinks during their training session or not, and regardless of whether participants were induced to judge the drinks before tasting them or afterward.

At the same time, we did find evidence that people’s spontaneous response tendencies to approach and avoid were being affected by our training, albeit apparently not for very long. So what does this tell us? We might conclude from these studies that approach-avoidance training simply does not work for soft drinks, but this would be very much premature. There are many possible points at which the procedure might have failed to achieve the required impact in order to produce an effect on our measures.

For example, it is possible that participants did not construe the movements they were performing as approach-avoidance movements, but rather as meaningless arm movements. If this is the case, then future uncritical use of the semantic labeling and zooming visuals we implemented should be questioned. Leaving that point aside, it is also possible that participants did not transfer what they had learned during the approach-avoidance training to other contexts – a possibility in line with our finding that the changes our training caused in their behavioral approach-avoidance tendencies did not last outside of the training task itself. This would raise the question of what circumstances facilitate such a transfer and why this transfer did not occur in our study but did in other research that successfully produced an effect. Finally, it might be that the unfamiliarity of the drinks was problematic, as no existing response tendencies existed to be retrained. If so, forming new action tendencies in a training might not lead to the same consequences in the absence of concrete experience with the targets.

This last point, in particular, is underlined by other research from our lab, in which we also had participants perform approach-avoidance training toward soft drinks, but this time towards existing brands they were already familiar with. In these studies, we did find an approach-avoidance training effect on spontaneous evaluation responses, but only when participants already had positive spontaneous evaluations towards the drinks before the training. If participants showed initial positive responses towards a drink, approaching that drink tended to increase these responses while avoiding it tended to reduce them.

Taken together, all this research suggests that approach-avoidance training might not be the magic bullet we initially hoped for with regard to reducing unwanted behavior. We could not find the effect under laboratory conditions while controlling for various possible extraneous influences, and we are far from the only ones who have failed to produce this effect. However, there is sufficient extant research that does show the effects of approach-avoidance to believe that there is something there. Taken together, the checkered picture painted by these findings should lend us hope that the intervention can be optimized to be more reliable. And to that end, finding out when it doesn’t work is as important as finding out when it does.