Talk Abstract: Political scientists have long been interested in understanding the intensity of voter preferences, but this endeavor has rarely been incorporated into the study of representation and disagreements over measurement remain. I measure intensity of preferences using a simple extension of conjoint analysis. I provide some evidence of the face validity of this measurement approach. This strategy requires large samples, so I use the approximately 500,000 cases from Nationscape. I show that simple survey marginals do not show large biases when it comes to the balance of intensity in the general population. However there are large differences in intensity across issues. Conservatives care much more about the issues on which they hold conservative views, and liberals care much more about the issues on which they hold liberal views. This may amplify polarization and affect the outcomes of primary elections. People with moderate views care more about the issues on which their views match the reigning consensus. This contradicts work that argues that moderate views are less deeply held, and supports the argument that moderate stances are beneficial in general elections.