In performing a hypothesis test based on a simple random paired sample to compare the means of two populations, knowing that the paired difference variable has a symmetric distribution that is far from normal, the paired t-test is acceptable if the sample size is small or moderate. Is this true? Explain.