The teaching of the concepts of statistical tests of hypotheses to non-statisticians
作者:
Shrikant I. Bangdiwala,
期刊:
Journal of Applied Statistics
(Taylor Available online 1989)
卷期:
Volume 16,
issue 3
页码: 355-361
ISSN:0266-4763
年代: 1989
DOI:10.1080/02664768900000043
出版商: Carfax Publishing Company
数据来源: Taylor
摘要:
The logic underlying the formulation of statistical tests of hypothesis can be counterintuitive for the non-mathematician, e.g. to test whether two treatments are different, why suppose they are equal? When introducing the topic of hypothesis testing, it is easy to present the formal fiamework for the testing procedure without explaining the logic behind it. In courses for statisticians, one may often (unjustifiably) rely on the understanding of probability concepts as a foundation for understanding statistical inference, but in courses taught to non-statisticians where there is minimal discussion of probability, it is essential that explanations must be based on concepts the students can readily understand. The method proposed here for teaching the concept of hypothesis testing makes an analogy to the judicial system, whereby a person is assumed innocent until sufficient evidence warrants a verdict of guilty. Analogies for the different elements of statistical tests are presented and discussed, together with a classroom fiamework for discussion of statistical tests.
点击下载:
PDF (461KB)
返 回