STA 215 Chapter Notes - Chapter 22: Astrological Sign, Null Hypothesis, Sample Size Determination
Document Summary
Addresses how closely observed numbers it null model. 256 heads of the largest fortune 400 companies. 12 hypothesized proporions for each sign: (you would use a z-test for one astrological sign: are execuives more likely to be pisces? and ask if the true proporion of execuives with that sign is equal to 1/12. ) Assumpions and condiions: counted data data must be counts for the categories of a categorical variable, independence randomness, to be generalized for a whole populaion, sample size at least 5 counts in each cell. Find the expected values from null hypothesis model. Expected value = (total # of observaions)(proporion of observaions) Example: 12 zodiac signs, 1/12 of 256 execuives, expected value is 21. 333: 2. Square the residuals and ind the sum: 4. Find df: test the hypothesis look up p-value. Can only reject the null or fail to reject the null; can never conirm that a theory (the null hypothesis) is in fact true.