[回归分析求助] 如何用stata做多重中介效应检验 [推广有奖]

11楼

jiemingshi0 发表于 2019-4-19 16:49:44 |只看作者 |坛友微信交流群

https://stats.idre.ucla.edu/stata/faq/how-can-i-analyze-multiple-mediators-in-stata/
可以看看这个，多重中介的

使用道具举报

12楼

山惟木子 发表于 2020-6-25 14:45:50 |只看作者 |坛友微信交流群

xianyun198109 发表于 2019-2-15 21:21
至今尚未解决，
同求：IV->MV1->MV2->DV 双重中介效应检验

你好，请问你现在解决了吗？

使用道具举报

13楼

kokowaah797

发表于 2020-6-28 09:06:16 |只看作者 |坛友微信交流群

直接建立结构方程模型SEM会比较方便，在菜单"Statistics>SEM(structural equation modeling)>model building and estimation"，直接画出你的多重中介路径，然后点击上方“estimate”图标，就可以了

使用道具举报

14楼

kokowaah797

发表于 2020-6-28 09:12:28 |只看作者 |坛友微信交流群

例如这样哈，随便找了个数据跑的[img]Version:0.9 StartHTML:0000000105 EndHTML:0000002250 StartFragment:0000000136 EndFragment:0000002217 <HTML><BODY><TABLE BORDER><tr><td>. sem (carrier -> storage, ) (carrier -> consume, ) (storage -> percentage, ) (storage</td><td>-> </td></tr><tr><td>> consume, ) (consume -> percentage, ), nocapslatent</td></tr><tr><td> Endogenous variables</td></tr><tr><td> Observed: storage consume percentage</td></tr><tr><td> Exogenous variables</td></tr><tr><td> Observed: carrier</td></tr><tr><td> Fitting target model:</td></tr><tr><td> Iteration 0: log likelihood = -3150.6169 </td></tr><tr><td>Iteration 1: log likelihood = -3150.6169 </td></tr><tr><td> Structural equation model Number of obs = 474</td></tr><tr><td>Estimation method = ml</td></tr><tr><td>Log likelihood = -3150.6169</td></tr><tr><td> </td></tr><tr><td>OIM</td></tr><tr><td>Coef. Std. Err. z P>z [95% Conf. Interval]</td></tr><tr><td> Structural </td></tr><tr><td>storage </td></tr><tr><td>carrier .2598609 .0457312 5.68 0.000 .1702294 .3494924</td></tr><tr><td>_cons 3.266003 .2216903 14.73 0.000 2.831498 3.700508</td></tr><tr><td> consume </td></tr><tr><td>storage -.1465384 .0352931 -4.15 0.000 -.2157117 -.0773651</td></tr><tr><td>carrier .1693156 .0363164 4.66 0.000 .0981368 .2404943</td></tr><tr><td>_cons 1.566516 .2056783 7.62 0.000 1.163394 1.969638</td></tr><tr><td> percentage </td></tr><tr><td>storage .156289 .0384182 4.07 0.000 .0809908 .2315873</td></tr><tr><td>consume .1216222 .050053 2.43 0.015 .0235201 .2197243</td></tr><tr><td>_cons 1.434682 .209252 6.86 0.000 1.024556 1.844809</td></tr><tr><td> var(e.storage) 1.999029 .1298509 1.76006 2.270442</td></tr><tr><td>var(e.consume) 1.180261 .0766662 1.03917 1.340508</td></tr><tr><td>var(e.percentage) 1.465849 .0952171 1.290618 1.664872</td></tr><tr><td> LR test of model vs. saturated: chi2(1) = 12.76, Prob > chi2 = 0.0004</td></tr></TABLE></BODY></HTML>[/img]

截屏2020-06-28 09.10.22.png (542.3 KB)

截屏2020-06-28 09.10.22.png

使用道具举报

15楼

kokowaah797

发表于 2020-6-28 09:55:14 |只看作者 |坛友微信交流群

我晕，结果没贴上来，我重新发一下

. sem (carrier -> storage, ) (carrier -> consume, ) (storage -> percentage, ) (storage ->
> consume, ) (consume -> percentage, ), nocapslatent
Endogenous variables
Observed: storage consume percentage
Exogenous variables
Observed: carrier
Fitting target model:
Iteration 0: log likelihood = -3150.6169
Iteration 1: log likelihood = -3150.6169
Structural equation model Number of obs = 474
Estimation method = ml
Log likelihood = -3150.6169
----------------------------------------------------------------------------------
| OIM
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
Structural |
storage |
carrier | .2598609 .0457312 5.68 0.000 .1702294 .3494924
_cons | 3.266003 .2216903 14.73 0.000 2.831498 3.700508
---------------+----------------------------------------------------------------
consume |
storage | -.1465384 .0352931 -4.15 0.000 -.2157117 -.0773651
carrier | .1693156 .0363164 4.66 0.000 .0981368 .2404943
_cons | 1.566516 .2056783 7.62 0.000 1.163394 1.969638
---------------+----------------------------------------------------------------
percentage |
storage | .156289 .0384182 4.07 0.000 .0809908 .2315873
consume | .1216222 .050053 2.43 0.015 .0235201 .2197243
_cons | 1.434682 .209252 6.86 0.000 1.024556 1.844809
-----------------+----------------------------------------------------------------
var(e.storage)| 1.999029 .1298509 1.76006 2.270442
var(e.consume)| 1.180261 .0766662 1.03917 1.340508
var(e.percentage)| 1.465849 .0952171 1.290618 1.664872
----------------------------------------------------------------------------------
LR test of model vs. saturated: chi2(1) = 12.76, Prob > chi2 = 0.0004