You don’t have to have an equal sign in the null. It depends on what you are testing. I was just messing around. cpk123 Wrote: ------------------------------------------------------- > mwt9 > > not sure what is wrong with it… >
From JDV: If CFAI says that H0 contains the point of equality, then pretend that’s true. It really does make no difference (with continuous parameters). In the theory of all this stuff is ‘supremum’ of the test statistic over, say, the space of the null hypothesis. A removable discontinuity because you wrote Mu < 75 doesn’t change that from if you wrote Mu <= 75. It really is pretty arbitrary. At the end of the day, your conclusion is that we can reject H0 or not. Extending that to any belief or statements about whether or not Mu could be exactly 75 is taking things too far. The whole thread is here: http://www.analystforum.com/phorums/read.php?11,591722,593100#msg-593100 I reallly wasn’t trying to muck up the waters this close to test time. For all intensive purposes don’t put the equal sign in the Ha.