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

返回

Jane Z · 2021年02月05日

不好意思，觉得之前的解答并不是很清楚

  1. CFA II
  2. Quantitative

NO.PZ2015120204000018

问题如下：

If an omitted variable is correlated with variables already included in the model, coefficient estimates will be biased and inconsistent and standard errors will also be inconsistent. Is this Statement correct?

选项：

A.

Yes.

B.

No, because the model’s coefficient estimates will be unbiased.

C.

No, because the model’s coefficient estimates will be consistent.

解释：

A is correct.

Chang is correct because a correlated omitted variable will result in biased and inconsistent parameter estimates and inconsistent standard errors.  

还是不是很懂。If an omitted variable is correlated with variables already included in the model, coefficient estimates will be biased and inconsistent and standard errors will also be inconsistent. Is this Statement correct?

  1. 虽然在missspecification里的第一点提到了omitted的变量，但是那里的条件并没有omitted variable is correlated with variables already included in the model这句话，为什么coefficient的estimates还是不准？
  2. 另外consistency不是说样本量越多，结果越准确吗？在之前提到的三种assumptionviolation里面，多重共线性里系数的估计也不准确了，但是仍然不会影响consistency。这里为什么就会影响到consistency？难道遗漏了变量就不是样本量越多越准确了吗？
添加评论
1 个答案
已采纳答案

星星_品职助教 · 2021年02月06日

同学你好，

①omitted variable bias成立需要有两个条件：

1）这个被遗漏的变量和方程里已有的X变量必须相关；

2）这个被遗漏的变量必须能解释Y；

只有满足了这两个条件，才能构成一个“omitted variable bias”

②omitted variable bias属于model misspecification的问题，会导致1）系数估计不准确和2）不满足consistency。这两个都是数学证明的结论，超出了CFA的范围，原版书里并没有涉及。

可以这样简单的去记忆，model misspecification是模型整体都错了，所以就都啥也不满足了。

而多重共线性只是OLS的假设没有满足，问题相对来说没有那么严重。虽然系数估计不准了，但统计学上的性质consistency还是满足的。

  • 1

    回答

  • 3

    关注

  • 672

    浏览
相关问题

NO.PZ2015120204000018 问题如下 If omittevariable is correlatewith variables alrea incluin the mol, coefficient estimates will biaseaninconsistent anstanrerrors will also inconsistent. Is this Statement correct? A.Yes. B.No, because the mol’s coefficient estimates will unbiase C.No, because the mol’s coefficient estimates will consistent. A is correct.Chang is correbecause a correlateomittevariable will result in biaseaninconsistent parameter estimates aninconsistent stanrerrors. 遗漏变量带来的结果不是异方差和自相关么？这两个问题都是不影响系数的啊？而且老师课上也说遗漏变量主要是影响残差项，和系数有什么关系？

2024-08-15 18:21 1 · 回答

NO.PZ2015120204000018 问题如下 If omittevariable is correlatewith variables alrea incluin the mol, coefficient estimates will biaseaninconsistent anstanrerrors will also inconsistent. Is this Statement correct? A.Yes. B.No, because the mol’s coefficient estimates will unbiase C.No, because the mol’s coefficient estimates will consistent. A is correct.Chang is correbecause a correlateomittevariable will result in biaseaninconsistent parameter estimates aninconsistent stanrerrors. 如上

2024-04-13 22:01 1 · 回答

NO.PZ2015120204000018 问题如下 If omittevariable is correlatewith variables alrea incluin the mol, coefficient estimates will biaseaninconsistent anstanrerrors will also inconsistent. Is this Statement correct? A.Yes. B.No, because the mol’s coefficient estimates will unbiase C.No, because the mol’s coefficient estimates will consistent. A is correct.Chang is correbecause a correlateomittevariable will result in biaseaninconsistent parameter estimates aninconsistent stanrerrors. 就这道题目衍生有两个问题什么是一致性？什么情况下会影响一致性（违反那些假设？或模型设定中有哪些错误？）及其原因

2023-10-17 11:53 3 · 回答

NO.PZ2015120204000018问题如下If omittevariable is correlatewith variables alrea incluin the mol, coefficient estimates will biaseaninconsistent anstanrerrors will also inconsistent. Is this Statement correct?A.Yes.B.No, because the mol’s coefficient estimates will unbiaseC.No, because the mol’s coefficient estimates will consistent.A is correct.Chang is correbecause a correlateomittevariable will result in biaseaninconsistent parameter estimates aninconsistent stanrerrors. 为什么会导致系数的inconsistent？老师不是说consistent的意思是变量个数n的增加，不会影响它的准确性吗。就算遗漏了某个变量x，但它能体现在残差项中，这个理解有问题吗？

2022-12-04 15:33 2 · 回答

NO.PZ2015120204000018 有问必答里之前有位助教的是影响了b1的估计就会导致inconsistent。但是多重共线性中会影响b1的估计也不会影响inconsistent，所以为什么ommitex会导致呢

2022-01-10 20:32 1 · 回答