計(jì)量經(jīng)濟(jì)第四章ppt課件(完整版)
【正文】 0: bj = 0 H1: bj ?= 0 c 0 a/2 ?1 ? a? c a/2 TwoSided Alternatives 雙邊替代假設(shè) reject reject fail to reject Example: Student Performance and School Size continued(example ) 例子：學(xué)生表現(xiàn)與學(xué)校規(guī)模 ? We have estimated 我們已經(jīng)得到 ^math10=++ –.0002enroll () () () () ? If the question is: whether the number of teachers has impacts on student performance, we can form hypotheses of: H0: bstaff = 0 , H1: bstaff ?= 0. 如果問(wèn)題是：教師數(shù)目是否對(duì)學(xué)生表現(xiàn)有影響，我們可以檢驗(yàn)如下假設(shè)： H0: bstaff = 0 , H1: bstaff 不等于 0. ? The calculated t ratio is . The 5% critical value of standard normal is . Since , we fail to reject the null 計(jì)算得到的 t值為 。 Computing pvalues for t Tests 計(jì)算 t檢驗(yàn)的 p值 ? An alternative to the classical approach : If the calculated t statistic is used as critical value, what is the smallest significance level at which the null hypothesis would be rejected? 另一種想法：如果將算得的 t 統(tǒng)計(jì)量作為臨界值，那么使得零假設(shè)被拒絕的最小顯著水平是多少？ ? This level is known as the pvalue. For a twosided alternative, 這個(gè)水平稱(chēng)為 p 值。對(duì)于 95自由度的 t分布， 1%顯著水平下單邊檢驗(yàn)的臨界值為 ，拒絕零假設(shè)。在此替代假設(shè)下，我們并未規(guī)定 xj 對(duì) y影響的符號(hào)。當(dāng)我們用某一特定樣本計(jì)算此統(tǒng)計(jì)量時(shí)，我們得到這個(gè)檢驗(yàn)統(tǒng)計(jì)量的一個(gè)實(shí)現(xiàn)（ t）。 Background Review 背景知識(shí)回顧 ? Two kinds of mistakes are possible in hypothesis testing. 在假設(shè)檢驗(yàn)中存在兩種可能的錯(cuò)誤。Multiple Regression Analysis: Inference 多元回歸分析：推斷 (1) y = b0 + b1x1 + b2x2 + . . . bkxk + u Lecture Outline 本課提綱 ? CLM assumptions and Sampling Distributions of the OLS Estimators 經(jīng)典假設(shè)與 OLS估計(jì)量的樣本分布 ? Background review of hypothesis testing 假設(shè)檢驗(yàn)的背景知識(shí) ? Onesided and twosided t tests 單邊與雙邊 t檢驗(yàn) ? Calculating the p values 計(jì)算 p值 Assumption (Normality) 假設(shè) （正態(tài)） ? So far, we know that given the GaussMarkov assumptions, OLS is BLUE, 我們已經(jīng)知道當(dāng) Gauss－ Markov假設(shè)成立時(shí)， OLS是最優(yōu)線(xiàn)性無(wú)偏估計(jì)。 ? Type I error: reject the null hypothesis when it is in fact true. 第一類(lèi)錯(cuò)誤：當(dāng)零假設(shè)為真時(shí)拒絕零假設(shè)（棄真） ? Type II error: fail to reject the null when it is actually false. 第二類(lèi)錯(cuò)誤：當(dāng)零假設(shè)為假時(shí)未拒絕零假設(shè)（取偽） Background Review 背景知識(shí)回顧 ? Hypothesis testing rules are constructed to make the probability of mitting type I error fairly small. 我們建立一些假設(shè)檢驗(yàn)的規(guī)則使發(fā)生第一類(lèi)錯(cuò)誤的概率非常小。 Theorem t Distribution for the Standardized Estimators 定理 : 標(biāo)準(zhǔn)化估計(jì)量的 t分布 ? ?? ?? ?? ?。 ? For a twosided test, we set the critical value based on a/2 and reject H0 if the absolute value of the t statistic c. c is the percentile in the t distribution with nk1 degrees of freedom if a?. 對(duì)于雙邊檢驗(yàn)，我們根據(jù) a/2計(jì)算臨界值。 Computing pvalues for t Tests 計(jì)算 t檢驗(yàn)的 p值 ? The steps in classical hypothesis testing: 經(jīng)典假設(shè)檢驗(yàn)的步驟 ? State the null and the alternative hypothesis 表述零假設(shè)和替代假設(shè) ? Decide a significance level and find the related critical value
教學(xué)課件相關(guān)推薦
鄂ICP備17016276號(hào)-1