Much research has examined the reciprocal relations between a child’s spontaneous focus on number (SFON) in the preschool years and later mathematical achievement. However, this literature relies on several different tasks to assess SFON with distinct task demands, making it unclear to what extent these tasks measure the same underlying construct. Moreover, prior studies have investigated SFON in the context of small sets exclusively, but no work has explored whether children demonstrate SFON for large sets and how this relates to children’s math ability. In the current study, preschoolers were presented four distinct SFON tasks assessing their spontaneous attention to number for small (Experiment 1) and large (Experiment 2) sets of numbers. Results revealed performance across the four distinct SFON tasks was unrelated. Moreover, preschooler’s SFON for small sets (1–4 items) was significantly stronger than that for large sets (10–40 items), and analyses revealed that number knowledge was only associated with SFON for small sets and not large. Together, findings suggest that SFON may not be a set-size-independent construct and instead may hinge upon a child’s number knowledge, at least in the preschool years. The role of number language and how it relates to children’s SFON are discussed.