Virtual worlds are emerging online environments that are beginning to be leveraged as a platform for consumer and marketing research. We compare the accuracy of responses obtained in virtual worlds with responses collected from traditional web-based research. We hypothesize that under certain conditions, data collected within a virtual world environment will generate more accurate responses than surveys administered through traditional Web browsers. We focus on low elaboration recognition tasks involving visual and text-based stimuli, and on how consumer motivational and characteristics of virtual worlds moderate accuracy on these tasks.