Last football season the starting quarterback of a professional football team completed just 55% of his attempted passes. The owner of this team has given the quarterback the ultimatum that he must improve his completion percentage during the off season or he will be fired. The owner plans to use a random sample of 50 passing attempts at the end of spring training to test the hypotheses where p represents the true proportion of passes this quarterback can complete after training during the off season. If the owner wants to minimize the chance of firing the quarterback when he really has improved, should he use a significance level of 0.01 or 0.10