This dissertation investigates consumer valuation of car characteristics with the special focus on two non-physical attributes of an automobile such as safety and comfort. Consumer valuation of automobile attributes is of interest to car manufacturers who supply the characteristics, consumers who purchase them, and policy makers who regulate the automobile. This study uses two approaches to accomplish this goal. The first one is the traditional hedonic method which calculates consumer willingness to pay for measurable components of safety and comfort, whereas the second one combines these components into comfort and safety indexes. It is argued in this study that these individual components, which can make a car safer or more comfortable, are evaluated by consumers in the broader context of safety and comfort before the final choice is made. It is also argued that this aggregation can be justified by a high degree of multicollinearity between various car attributes which has been observed in the previous hedonic studies of the automobile market. Included here is also a comprehensive discussion of econometric problems associated with the characteristics approach. The computational part is based on the new and the most extensive data set used in the hedonic literature of the automobile market. The study concludes by presenting the set of price and income elasticities of demand for the safety and comfort related variables. The results of both methods indicate that many car attributes are Giffen goods, which implies a positive relationship between the marginal willingness to pay and quantity purchased. The main reasons for these findings could be attributed to the impact of the government quality standards affecting automobiles and the shortcomings of the hedonic procedure (treatment of nonlinearities).