The most commonly used measure of spatial cognition to assess both adults and children is mental rotation. However, little is known about the cognitive strategies that children use to solve this task. Understanding how and when children may employ different mental rotation strategies can illuminate the development of mental rotation ability and help clarify previous mixed findings on the developmental trajectory of children's mental rotation skills. Thus, in this study, we investigated what strategies children use in a new mental rotation task and whether their strategy use would be influenced by the test instructions. In Experiment 1, we found that the types of strategies children used in our mental rotation test differed from strategies reported in previous research, suggesting that strategy use is dependent on test design. In Experiment 2, we found that children's strategy use can be malleable; changing the test instructions reduced one type of erroneous strategy, flipping. Our findings suggest that different tests labeled “mental rotation tests” may actually be measuring different abilities.