In this paper I am going to perform a one-tailed, one-sample tì-test in order to test the null hypothesis that the average number of ounces of cereal per box equals sixteen. A manufacturer of a particular brand of cereal maintains there is an average of sixteen ounces of cereal per box. On the other hand, a consumer group asserts the average is fewer than sixteen ounces per box. The consumer group is going to file a class action lawsuit for false advertising if the average is fewer than sixteen ounces per box. I was hired by the consumer group to determine whether the consumer group is correct or the cereal manufacturer is correct.
According to Sprinthall (2003, p. 249), ?Sometimes, however, a researcher not only assumes that a mean difference between samples will occur but also predicts the direction of the differenceì. When this happens, the statistical decision is not based on both tails of the distribution, but only one.
In this instance, the sign of the t ratio is crucial. When conducted in this way, the t test is called a one-tail t test. The calculation of a one- tail t test is identical to that of the two-tail t. The only differences are in the way the alternative hypothesis is written and in the method used for looking up the table value of t [italics in original].
I randomly selected 20 supermarkets, and from each supermarket I randomly selected one box of the manufacturer?s cereal. I then weighed the cereal in each box for all twenty boxes. I came up with numbers ranging from 12.4 being the lowest to 18.4 being the highest weight of the cereal. I then used the one-tailed, one-sample t test to test the null hypothesis that the average number of ounces of cereal...