. predict res, r
. gen res_abs=abs(res)
. reg  res_abs lavgprs lincpc

. predict sig_hat
. reg  lpackpc lavgprs lincpc [aw = sig_hat]

. reg  lpackpc lavgprs lincpc
      Source |       SS       df       MS              Number of obs =     528
-------------+------------------------------           F(  2,   525) =  162.03
       Model |  9.65857683     2  4.82928842           Prob > F      =  0.0000
    Residual |  15.6472676   525  .029804319           R-squared     =  0.3817
-------------+------------------------------           Adj R-squared =  0.3793
       Total |  25.3058444   527   .04801868           Root MSE      =  .17264
------------------------------------------------------------------------------
     lpackpc |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     lavgprs |  -.7852949   .0503104   -15.61   0.000    -.8841293   -.6864604
      lincpc |   .3040118   .0531955     5.71   0.000     .1995097    .4085139
       _cons |   7.686531   .1703546    45.12   0.000     7.351871    8.021191
------------------------------------------------------------------------------

. hettest year
Breusch-Pagan / Cook-Weisberg test for heteroskedasticity
         Ho: Constant variance
         Variables: year
         chi2(1)      =     7.76
         Prob > chi2  =   0.0053

3# Elisa129

下面是我自己做的第一问，和你给的答案一样，但第二问不知道怎么做，FGLS应该要用prais命令吧，请问你的第二问答案怎么还是做的OLS回归啊？

. use http://www.stata-press.com/data/imeus/cigconsump

. keep if year == 1985 | year == 1995

(432 observations deleted)

. regress lpackpc lavgprs lincpc

Source |
SS
df
MS
Number of obs =
96

-------------+------------------------------
F(
2,
93) =
45.42

Model |
2.78667621
2
1.3933381
Prob > F
=
0.0000

Residual |
2.85290424
93
.03067639
R-squared
=
0.4941

-------------+------------------------------
Adj R-squared =
0.4832

Total |
5.63958045
95
.059364005
Root MSE
=
.17515

------------------------------------------------------------------------------

lpackpc |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]

-------------+----------------------------------------------------------------

lavgprs |
-.8867271
.128371
-6.91
0.000
-1.141647
-.6318077

lincpc |
.435617
.140242
3.11
0.003
.1571241
.7141099

_cons |
7.780854
.3321711
23.42
0.000
7.121228
8.44048

------------------------------------------------------------------------------

. estat hettest year

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity

Ho: Constant variance

Variables: year

chi2(1)
=
4.31

Prob > chi2
=
0.0379

Hettest is used to test whether the estimated variance of the residuals from a regression are dependent on the values of the year variable. The null hypothesis that the variance of the residuals is homogenous. Therefore, if the p-value is very small, we would have to reject the hypothesis and accept the alternative hypothesis that the variance is not homogenous. So in this case, the evidence is against the null hypothesis that the variance is homogeneous, so we need to refit the model.

看了下书，好像是对的，500论坛币已经给了，请查收。

另外还有2道题目，悬赏1000论坛币，你如果有时间，也看看吧。