All else being equal, if you cut the sample size in half, how does this affect the margin of error when using the sample to make a statistical inference about the mean of the normally distributed population from which it was drawn? M E = StartFraction z times s Over StartRoot n EndRoot EndFraction. The margin of error is multiplied by StartRoot 0.5 EndRoot. The margin of error is multiplied by StartRoot 2 EndRoot. The margin of error is multiplied by 0.5. The margin of error is multiplied by 2.