Test Fairness in the New Generation of Large-Scale Assessment