This study analyzes the probability of buying products from a particular online store given a set of alternative vendors. We use a multinomial logit choice model to analyze experimental data of consumer choice in two product categories - books and computers, each of which represents goods that differ along the risk dimension. Books represent search-quality, cheaper products, while computers represent experience-quality, more expensive products. Our modeling approach enabled us to capture the competitive environment of Internet shopping and at the same time obtain diagnostic information about the salient factors in the choice process. The results indicate the existence of differences in terms of the dimensions considered by consumers when buying high- vs. low-risk goods. Most notably, we found that when purchasing computers online, aspects of uncertainty and risk were more salient than when purchasing books online. We discuss the implications of these findings and areas for future research.