This article addresses the need for systematic theory comparison and development in environmentally significant behavior (ESB) research. Using logistic regression (N = 398), models based on Schwartz's norm-activation theory (NAT) and Ajzen's theory of planned behavior (TPB) were compared as explanations of drivers' intentions to reduce or maintain their car use for commuting. NAT explained more variance (McFadden R² = .342). A model using NAT and TPB constructs was also tested. This explained more variance than either individual theory (McFadden R² = .379). A personal-normative variable (NAT) and perceived behavioral control (TPB) were the only statistically-significant predictors of intentions in the model derived from both theories. It is argued that combining NAT and TPB constructs accounts for a range of influences on car-use intentions that neither individual theory fully captures. A combined model may also apply to other ESBs, especially those perceived as reducing personal utility (i.e., entailing sacrifice).