Because the picture based on previous research is fairly muddy, Sharon and Woolley decided to to children's beliefs about fantasy vs. real entities more directly. They presented children ages 3, 4, and 5 with colored line drawings of real and fantasy entities. One of the real entities (Michael Jordan) was specific, while the other was generic (a drawing of a child), two of the fantasy entities were generic (e.g., a monster) and two were specific (e.g., Santa Claus). They asked the children 12 yes/no questions about various properties. The children were to answer yes if they believed the entity had that property, and no if they believed it did not. In addition, they asked children to categorize the entities as real, pretend, or "not sure." What they found was interesting. Four and five-year old children were clearly able to distinguish real from pretend entities in the property attribution task. The results of their property attributions were similar to those of adults. Things were different in the categorization task, though. In all three age groups, the children categorized the entities correctly less than 40% of the time. However, this is due in large part to children categorizing many of the pretend entities as "not sure" (most tended to classify Santa and the Easter Bunny, the specific fantasy entities, as real). This indicates that instead of believing that the pretend entities are real, children are actually uncertain about their status. From this data, Sharon and Woolley drew the following conclusions:
Our results suggest the following tentative sketch of the possible course of development of the fantasy/reality distinc tion: The very young child may initially be somewhat unsure about attributing human-like properties to various entities. With experience , children acquire increasing knowledge about everything in their world—both about real entities and their properties, and about such socially supported myths as Santa Claus and the Easter Bunny. Thus, there is the simultaneous development of beliefs considered correct (e .g. dinosaurs are real) and of beliefs considered incorrect but age-appropriate (e .g. Santa is real). But at the same time, as children believe in the reality of fantasy figures, or are unable to say with certainty that they are pretend, they treat them very differently from real entities in terms of the properties and abilities they are willing to grant. In this way, children seem to place fantastical entities in a separatecategory—neither unqu stionably real nor pretend, but somewhe re in between. This category could then form a natural bridge to the adult cate gory of fantastical entities. Thus, rather than having misplaced the boundary between real and fantastical entities, young children are still in the process of actively constructing it. (p. 308)
* All of the information is from the Sharon and Woolley paper. To find citations for the individual points, you'll have to read the paper.