问一道题：NO.PZ2016062402000026 [ FRM I ]

NO.PZ2016062402000026 问题如下 Unr whcircumstances coulthe explanatory power of regression analysis overstate The explanatory variables are not correlatewith one another. The varianof the error term creases the value of the pennt variable increases. The error term is normally stribute An important explanatory variable is omittethinfluences the explanatory variables incluanthe pennt variable. If the true regression inclus a third variable z thinfluences both y anx, the error term will not be contionally inpennt of x, whiviolates one of the assumptions of the OLS mol. This will artificially increase the explanatory power of the regression. Intuitively, the variable x will appeto explain more of the variation in y simply because it is correlatewith z 看了之前的解答还是不明白“假设X为自变量，Y为因变量。如果存在第三个变量Z能够同时影响Y与X，并且这个变量Z没有纳入我们的模型方程中，那么X对于Y的力度就会被高估。这是因为，变量X其实本身是无法具有现在这么高的力度的，它之所以现在力度高，是因为它包含了一个被我们所遗漏掉的、和X相关的、对Y有作用的变量Z。所以，如果这种情况下是高估的”1.力度的评价是不是看R平方或者austeR平方？R平方越高说明力度越高？2.遗漏变量的时候，说明模型的回归效果不好，R平方应该会更低才对，为什么这个时候是overstate？当这个被遗漏的变量加入模型后，拟合效果更好，R平方应该上升

The varianof the error term creases the value of the pennt variable increases. The error term is normally stribute important explanatory variable is omittethinfluences the explanatory variables incluanthe pennt variable. If the true regression inclus a thirvariable z thinfluences both y anx, the error term will not contionally inpennt of x, whiviolates one of the assumptions of the OLS mol. This will artificially increase the explanatory power of the regression. Intuitively, the variable x will appeto explain more of the variation in y simply because it is correlatewith z 的情形，遗漏变量为什么会使回归的力度被夸大？遗漏产量不应该会导致回归效果不好吗？以至于回归力度下降？

老师，b是在说条件异方差的现象吗？这个现象会减少对y的力度吗？

为什么忽略了一个重要变量会让x疑似过度拟合呢？是不是因为这个变量被包含在残差项里呢？请，！