You can resist buying a candy bar while you're waiting in the checkout line at the grocery store--but you'll buy any pair of shoes that are on sale. Your best friend, in contrast, wouldn't dream of buying a pair of shoes he thinks he doesn't need, no matter how low the price--but he can't resist buying that same candy bar you so easily ignore. According to a new study in the Journal of Public Policy&Marketing, it is precisely those differences in self-control that researchers need to pay attention to when assessing the impact of public policies.

As the study reports, if you want to understand the effectiveness of a regulation or tax on a specific behavior, use a measure of self-control specific to that behavior.

"That may seem obvious. But our study is the first to compare general measures of self-control with so-called domain-specific measures of self-control," write the authors of the study, Kelly L. Haws (Vanderbilt University), Scott W. Davis (Texas A&M University), and Utpal M. Dholakia (Rice University). "And what we find is that domain-specific measures produce better predictions of consumer behavior when it comes to the impact of taxes or regulations intended to promote or discourage a particular behavior."

Consumer researchers routinely use either a general measure of self-control or a domain-specific measure--but without explaining or defending the measure chosen. What Haws, Davis, and Dholakia set out to find is which self-control measure is most appropriate for a given context. To that end, they conducted five studies that measured self-control in relation to two specific behaviors: eating and shopping.

The authors ultimately recommend a particular approach to measuring domain-specific self-control, an approach that adapts an existing general measure to the specific context at hand.

"The approach we use here can be adapted to assess individual self-control in other domains such as time management, alcohol consumption, and so on. Whether one is interested in spending, eating or another self-control behavior, greater consistency in assessing individual differences will facilitate both theoretical advancement and the testing and discovery of effective interventions in future research," the authors write.