STAT 2507 Chapter Notes - Chapter 10: Alternative Hypothesis, Null Hypothesis, Test Statistic
Document Summary
Small-sample inferences for the di erence between two means: The automobiles are then operated for a speci ed number of miles, and the amount of wear is recorded for each tire. Instead, we chose to pair the measurements, comparing the wear for type a and type b tires on each of the ve automobiles. This experimental design, sometimes called a paired-di erence or matched pairs design, allows us to eliminate the car-to-car variability by looking at only the ve di erence measurements. di = x1i x2i. D = (cid:80) di n (cid:122) paired-di erence test of hypothesis for ( 1 2) = d: dependent samples: null hypothesis: h0 : d = 0, alternative hypothesis: Ha : d > 0 (or, ha : d < 0) Ha : d (cid:54)= 0 n = d sd/ n t = d 0 sd/ where n =number of paired di erences.