100 Statistical Tests
Gopal K Kanji | ISBN-13 978 1 4129 2376 7 | Pdf | 257 pgs | 2 mb
Having collected together a number of tests, it is necessary to consider what can be tested, and we include here some very general remarks about the general problem of hypothesis testing. Students regard this topic as one full of pitfalls for the unwary, and even teachers and experienced statisticians have been known to misinterpret the conclusions of their analysis.
Broadly speaking there are two basic concepts to grasp before commencing. First, the tests are designed neither to prove nor to disprove hypotheses. We never set out to prove anything; our aim is to show that an idea is untenable as it leads to an unsatisfactorily small probability. Second, the hypothesis we are trying to disprove is always chosen to be the one in which there is no change; for example, there is no difference between the two population means, between the two samples, etc. This is why it is usually referred to as the null hypothesis, H0. If these concepts were firmly held in mind, we believe that the subject of hypothesis testing would lose a lot of its mystique. (However, it is only fair to point out that some hypotheses are not concerned with such matters.)
...
No comments:
Post a Comment