Following over a century of research, pretend play has come to be known as an important feature of childhood. Children across cultures have been observed spontaneously engaging in pretend play, promoting the notion that pretend play serves a special developmental function. Early research described how pretend play provides a unique platform for children’s self-directed learning. Today, particularly in the U.S., pretend play is understood to be so important for young children that parents are encouraged to promote it and participate in it at home. How then does pretend play function in development when it is no longer spontaneously motivated by children? In this dissertation, I examine the prevailing assumptions regarding the role of pretend play in children’s lives and the role of adults in supporting pretend play. Using a mixed-methods approach, I describe how 60 families drawn from an urban American population vary in their practice of pretending at home in the preschool years. I then evaluate the relation between pretend play practices and learning outcomes at kindergarten. I find that children vary significantly in their investment in pretending at home, with distinct patterns of pretend play investment over time. Several children pretended rarely at all observations, and most children remained constant or decreased their pretend play time between three and four years of age. For the 25% of children who invested the most time in pretending, parents were also most likely to be participating. Parents and children pursued distinct goals during pretend play, and parents’ efforts to achieve learning objectives through pretend did not always produce the expected results. This study contributes new insights through the in-depth examination of pretending at home in a large, representative sample, and highlights the importance of examining the variation across families in order to better understand and better serve children’s learning and development.