What is a likely lurking variable in this case that could explain this observed relationship? (Hint: this lurking variable affects both the type of light with which a child sleeps and the child's later nearsightedness.)