CFAI Reading 12 Example # 1 pg 354 There are two sets of Null and alternative hypotheses on this page. One for b1, one for b2. Can someone help me interpret the reasoning behind the two one tailed tests and the hypothesis tests chosen here. I would have thought these are two tailed tests.
“”""""""""""""" on most cases “”""""""""""" On Multiple regressions you want to test that each slope coefficient (in your example b1 and b2) is statistically different than 0, which is essentially a two tailed test. basic rule, if the test has an = sign only it will be a two tailed test. You can to test correlation different than 0 because you want to see that the coefficients have statistical significance for testing that they have some explanatory power of the dependant variable (if they are not good at explaining the dependant variable then they should not be stat. sig different than 0) “”"""" specific cases “”"""" In specific cases however and this is the trick for the exam, read the item set figure out what the appropriate tests should be. In this case, you the test is that “competition leads to lower costs” so you expect that the relation is greater the number of market makers. hence your H0 is >= 0 because of what you are tryting to prove. So just watch out when you read the item set to determine whether you are expecting a negative relationship where a one tailed test is more appropriate