The Psychology of Eco-Consumption

Information programs to promote cellulosic biofuels may not achieve their objectives unless consumers can be induced to care about the information presented to them. The social psychology literature highlights two commonly used models to link psychological variables to environmentally related behaviors: the Theory of Planned Behavior (TPB) and the Norm Activation Theory (NAT). Other studies have compared the strength of these models or have adapted these models by adding additional variables, but few have compared across the alternative variable combinations noted in the literature. That is, most studies have added one or two psychological variables to the NAT or TPB models and have found that the additional variable is a significant factor influencing behavior. However, we are unfamiliar with any study that has included the full suite of examined variables within one model. This could be a problem in that the psychological variables are likely to be correlated. In turn, the output of these models may suffer from omitted variable bias; which could lead to erroneous conclusions about the importance of any specific variable. Previous findings that individual variables are significant in influencing behaviors may be incorrect. One objective here then is to start examining whether these ‘significant findings’ are robust, and if not, whether we can be more parsimonious in future modeling efforts. Economists often assume preferences are adequately proxied by the person’s socioeconomic characteristics or by the person’s participation in some environmental behavior. Recently, economists have begun to recognize that these characteristics are poor proxies especially since the proxies commonly used are binary (0/1) variables that provide relatively little detail, are usually not policy or program relevant and lack a theoretical justification. Thus, another objective is to compare the performance of models that incorporate these proxy variables with models incorporating psychological variables. Our results suggest: that a combined TPB and NAT model may be more effective than either model alone; that many variables seen as important in the literature may be less important than previously thought (allowing for more parsimonious models resulting in less costly data collection); and that common proxy variables like membership in an environmental group may not work that well. ∗Presented at the 2 International Workshop on Ecolabeling, Laboratory of Forest Economics, INRA–AgroParisTech. Nancy, France June 29-30.