开发者:上海品职教育科技有限公司 隐私政策详情

应用版本:4.2.11(IOS)|3.2.5(安卓)APP下载

十种笠 · 2021年09月10日

critical value、 P value、significant level傻傻分不清

NO.PZ2018101001000048

问题如下:

Steven does the F-test of 8 variables and gets that the significance of F is 0.023, which is smaller than the significant level at 0.05, so he makes a conclusion that at least one of the dependent variables in the regression model makes a significant contribution to explain the independent variable. Is this conclusion most likely correct?

选项:

A.

Yes.

B.

No, it should be all of the independent variables in the regression model make a significant contribution to explain the dependent variable.

C.

No, it should be at least one of the independent variables in the regression model makes a significant contribution to explain the dependent variable.

解释:

C is correct.

考点: Regression Coefficient: F-test.

解析: The significance of F is 0.023就意味着P-value等于0.023,小于α = 0.05,所以拒绝原假设,得到的结论应该是at least one of the independent variables in the regression model makes a significant contribution to explain the dependent variable,题干中的结论把自变量和因变量说反了。所以应该选择C选项。

1、F value、critical value、 P value、significant level傻傻分不清

2、到底是哪个大于哪个拒绝原假设啊?拒绝原假设就=significant?

1 个答案

星星_品职助教 · 2021年09月11日

同学你好,

这是四个不同的概念,就以F检验为例,全部说明一下。

①F value(一般叫做F-statistic更多一些)是F检验中要计算的F统计量。

②计算出检验统计量后,要拿这个F value/statistic去和查表得到的F critical value去做对比,如果|F-statistic|>|F critical value|,则拒绝原假设

③p-value(在F 检验里也被叫做significance of F)是用来和significance level做对比的量。如果p-value<significance level,直接拒绝原假设

④significance level是人为设定的,例如一般的1%,5%等。

---------

以上②,③步就是拒绝原假设的原则,这两个步骤结论是一致的。

可以认为significant=拒绝原假设