Problem: A company manufactures ion thrusters for spacecraft, and is seeking to improve on their version 1 design. Cost limitations mean that test data is limited; they take 10 measurements of thrust from their version 1 design, and 10 measurements of thrust from their version 2 design. Explain how they might use a resampling method to determine whether there is a statistically significant increase in the thrust between the two versions. What assumptions are required? What benefits are there over a classical difference-of-sample-means T test?