【正文】
ossible in hypothesis testing. 在假設(shè)檢驗中存在兩種可能的錯誤。 ? The significance level (level) of a test is the probability of a Type I error. 一個檢驗的顯著性水平是發(fā)生第一類錯誤的概率。如果為 5%的檢驗中錯誤地拒絕零假設(shè)。 ? The set of values of the test statistic for which the test rejects the null is the rejection region, and the values of the test statistic for which it does not reject the null is the acceptance region. ? 假設(shè)檢驗中,使得零假設(shè)被拒絕的檢驗統(tǒng)計量的取值范圍稱為 拒絕域 ,使得零假設(shè)不能被拒絕的檢驗統(tǒng)計量的取值范圍成為 接受域 。當(dāng)我們用某一特定樣本計算此統(tǒng)計量時,我們得到這個檢驗統(tǒng)計量的一個實(shí)現(xiàn)( t)。來估計分布,因為我們要用注意這是一個假設(shè)下,有在221j221j ? ~ ??C L M1:f r e ed o m of d eg r ee s t h eN o t e?by es t i m at e t oh av e w eb ec a u s e n o r m al ) ( v son d i s t r i b u t i a is t h i sN o t e)( ~ ?? sas s u m p t i o n C L M U n d er t h essbbbssbbbttseknttseknjjknjj????????The t Test (cont) t檢驗 ? Knowing the sampling distribution for the standardized estimator allows us to carry out hypothesis tests 知道標(biāo)準(zhǔn)化估計量的樣本分布后,便可以進(jìn)行假設(shè)檢驗 ? Start with a null hypothesis 由零假設(shè)出發(fā) ? For example, H0: bj=0 () 例如, H0: bj=0 ? If accept null, then accept that xj has no partial effect on y, controlling for other x’s 如果接受零假設(shè),則認(rèn)為控制 x其它分量后, xj對 y沒有邊際影響。 The t Test (cont) ? ?? ?0?j0?jHt??t?H ,h y p o t h e s i s n u l l a c c e p t t h e o w h e t h e r td e t e r m i n e t or u l er e j e c t i o n a w i t h a l o n g s t a t i s t i c o u r u s e t h e n w i l lWe)(?? ?f o r s t a t i s t i c t h e f o r m t on e e df i r s t eo u r t e s t w p e r f o r m To定是否接受零假設(shè),統(tǒng)計量和拒絕條件來決然后利用統(tǒng)計量:的先要構(gòu)造為了進(jìn)行檢驗,我們首:jjjjsettsettjjbbbbbbbb??t Test: OneSided Alternatives t檢驗:單邊替代假設(shè) ? Besides our null, H0, we need an alternative hypothesis, H1, and a significance level 除了零假設(shè)外,我們需要替代假設(shè) H1,并設(shè)定顯著性水平 ? H1 may be onesided, or twosided H1可以是單邊或雙邊的 ? H1: bj 0 and H1: bj 0 are onesided H1: bj 0 和 H1: bj 0 是單邊的 ? H1: bj ? 0 is a twosided alternative H1: bj ? 0是雙邊替代假設(shè) ? If we want to have only a 5% probability of rejecting H0 if it is really true, then we say our significance level is 5% 如果我們愿意在 5%的概率上錯誤地拒絕實(shí)際上為真的零假設(shè),則說我們的顯著水平為 5% OneSided Alternatives (cont) 單邊替代假設(shè) ? Having picked a significance level, a, we look up the critical value c in a t distribution with n – k – 1 degree of freedom . 取定顯著性水平 a后,找到自由度為 n – k – 1的 t分布的臨界值 OneSided Alternatives (cont) ? If H0: bj = 0 versus H1: bj 0, we reject H0 if tbj c, fail to reject H0 if tbj =c. 如果 H0: bj = 0對 H1: bj 0,當(dāng) tbj c時我們拒絕H0,當(dāng) tbj =c,則不能拒絕 H0 ? Because t distribution is symmetric, if H0: bj = 0 versus H1: bj 0, we reject H0 if tb