A statistical test that indicates the likelihood of observing the difference if the true differ ence were zero. A large value of this statistic leads to nonacceptance of the null hypothesis that the true difference is zero.