Young children are powerful causal learners: they readily track statistical contingencies between causes and effects, and they can use this evidence to infer general rules for a system (e.g., red blocks, but not blue blocks, will cause this machine to play music). However, little is known about the ways in which children 1.) transfer the causal rules they form in one context to produce new outcomes, and 2.) select relevant causal variables based on evidence. In this talk, I'll discuss new data that support preschoolers' ability to perform this kind of abstract transfer. I will also discuss ongoing investigations into the question of variable choice. I situate both of these topics within an "interventionist" philosophical framework for understanding causation, considering both causal intervention and causal explanation. Finally, I will explore the implications of this project for understanding open questions about the role of "placeholder structures" and variables in conceptual development.