MGMT 1050 Chapter Notes - Chapter 15: Form 10-Q, Test Statistic, Null Hypothesis
Document Summary
We will learn three chi-squared tests: goodness of fit test, a test for normality, test of a contingency table. Goodness of fit test: for every category, the expected value must be 5 or more, this condition can be met by increasing the sample size. In the market share for soft drinks, pepsi has a 48% share, coke has a 42% share and other brands have a 10% share. We are interested to see if the promotional activities have been successful: h0: the market share are the same from university to national market share, h1: the market shares are different. If there were only two categories, this would be a binomial problem (z test for p1 p2) k. =(cid:229) e i i observed cell frequency f i expected cell frequency e i number of categories k. Note how every expected frequency is greater than 5. The test stat is calculated through the differences between the survey results and expected frequency.