The article discusses how tests are used to determine if a hypothesis is true or not. It explains that tests need to control the chance of wrongly rejecting a true hypothesis and also focus on detecting when a hypothesis is wrong. The researchers use sampling distributions to calculate these probabilities. They mention that exact distributions are hard to find in real-world cases, so they use approximate ones that are still valid. The main concern is how good these approximations are. The study mentions different types of tests like Bootstrap Test and Monte Carlo Test.