Media effects scholars have largely focused on the effects of viewing media while ignoring how media have infiltrated viewers' environments in other ways. In particular, researchers should consider how merchandise from media texts may affect media effects or may produce unique media effects on their own. In this study, 41 girls ages 4-6 years were randomly assigned to play with either a doll of their favorite Disney Princess or a similar, non-media-related princess doll for ten minutes. They were then interviewed about their conceptions of their favorite Disney Princess. Using a new technique called "narrative string coding," the content of the participants' play was compared to the movie or movies featuring their favorite Disney Princesses. Quantitative analysis revealed that participants who played with Disney Princess dolls included more content in their play that exactly matched Disney Princess movies than did participants who played with other princess dolls. In addition, participants with more complex understandings of their favorite Disney Princesses as unique entities incorporated more content into their play that was completely different from Disney Princess movies than did children with less complex understandings of their favorite Disney Princesses. Qualitative analysis revealed that the relationship among the presence of media merchandise, the content media texts, and the aspects of their favorite Disney Princess that the participants' most liked was associated with differences in the participants' use of gender stereotypes in their fantasy play. These results indicate that scholars must consider the full media environment, including merchandise, to fully understand media effects. This may have particularly important implications for cultivation and media's effects on young children's cognitive development. Further, parents who are concerned about children learning stereotypes from media should be meticulous in choosing exactly what merchandise with which their children should be allowed to interact.