Question: In a certain problem, the randomization test produces an exact two-sided p-value of 0.053, while the t-distribution approximation produces 0.047. One might say that since the p-values are on opposite sides of 0.05, they lead to quite different conclusions and, therefore, the approximation is not adequate. Comment on this statement.