PSM方法求助 - Stata专版

0关注
1粉丝

大专生

90%

还不是VIP/贵宾

-

0%

威望: 0 级
论坛币: 406 个
通用积分: 0.6000
学术水平: 0 点
热心指数: 0 点
信用等级: 0 点
经验: 223 点
帖子: 17
精华: 0
在线时间: 105 小时
注册时间: 2014-2-8
最后登录: 2024-10-14

楼主

塞外晓木匠 发表于 2018-2-1 11:54:31 |AI写论文

是否 +2 论坛币

k人参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

stata用diff命令做kernel PSM的DID，做出来的结果感觉和陈强老师那本书上列示的结果不一样，感觉有点怪。我用的是有十一年的面板数据，我的命令是diff houseprice,treat(treated) period(after) id(code4) kernel cov(pergdp secondary tertiary loan real_estate fixed_asset deposit) rep logit sup test。命令都是diff命令中的实例，然后我的期数为11期，因为每年都会有几个城市开通机场，想看看其对房价的影响，但是出来的结果是这样的：

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

回帖推荐

黃河泉发表于6楼查看完整内容

根据你讲的"因为每年都会有几个城市开通机场"，我怀疑你不可以用 diff 来估计你的模型，如果我没误解，diff 指令需要大家"同一时间"开通 (我指的是 treatment group)。

沙发

塞外晓木匠 发表于 2018-2-1 11:55:02

. diff houseprice,treat(treated) period(after) id(code4) kernel cov(pergdp ///
>  secondary tertiary loan real_estate fixed_asset deposit) rep logit sup test
Report - Propensity score estimation with logit command
Atention: _pscore is estimated at baseline

Iteration 0: log likelihood = -1432.3021
Iteration 1: log likelihood = -1355.2746
Iteration 2: log likelihood = -1354.4134
Iteration 3: log likelihood = -1354.4113

Logistic regression                            Number of obs =    2074
                                                      LR chi2(7)    =    155.78
                                                      Prob > chi2    =    0.0000
Log likelihood = -1354.4113                Pseudo R2    =    0.0544

------------------------------------------------------------------------------
   treated |    Coef.    Std. Err.       z    P>|z|    [95% Conf. Interval]
-------------+----------------------------------------------------------------
   pergdp |  -.0626399 .0272417 -2.30 0.021 -.1160327 -.009247
secondary | 5.081141 .6769763    7.51 0.000    3.754292    6.40799
tertiary |    5.85948 1.012762    5.79 0.000    3.874504 7.844457
      loan |    -.2268431 .1413841 -1.60 0.109 -.5039507 .0502646
real_estate | 5.680924 1.160947    4.89 0.000    3.40551 7.956339
fixed_asset |  -1.652177 .2364751 -6.99 0.000 -2.115659 -1.188694
   deposit | -.1746447 .2030676 -0.86 0.390 -.5726499 .2233606
   _cons | -3.403146 .6025458 -5.65 0.000 -4.584115 -2.222178
------------------------------------------------------------------------------
TWO-SAMPLE T TEST
Test on common support

Number of observations (baseline): 2074
         Before       After
Control: 963          -          963
Treated: 1111       -          1111
            2074          -

t-test at period = 0:
--------------------------------------------------------------------------------
> --------------
Weighted Variable(s) | Mean Control | Mean Treated | Diff. | |t| |
>  Pr(|T|>|t|)
---------------------+------------------+--------------+------------+---------+-
> --------------
houseprice          | 2.764          | 3.085       | 0.321    |  3.85 |
> 0.0001***
pergdp             | 2.612          | 2.770       | 0.158    |  1.68 |
> 0.0937*
secondary          | 0.505          | 0.506       | 0.001    |  0.27 |
> 0.7898
tertiary          | 0.352          | 0.354       | 0.002    |  0.50 |
> 0.6193
loan                | 0.711          | 0.719       | 0.008    |  0.43 |
> 0.6703
real_estate       | 0.081          | 0.082       | 0.002    |  0.64 |
> 0.5231
fixed_asset       | 0.581          | 0.580       | -0.001    |  0.16 |
> 0.8724
deposit             | 0.673          | 0.666       | -0.008    |  0.65 |
> 0.5134
--------------------------------------------------------------------------------
> --------------
*** p<0.01; ** p<0.05; * p<0.1
Attention: option kernel weighs variables in cov(varlist)
Means and t-test are estimated by linear regression