一、中介 Mediator
* DV(Y): perf_tot
* IV(X): sUPB_cen
* 中介變數：Mediato: eUPB_cen
* 控制變數：control variable: smdis_cen
* data: Study_1.dta
. summ perf_tot eUPB_cen sUPB_cen smdis_cen
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
perf_tot | 176 5.460227 1.211976 1 7
eUPB_cen | 176 .0000167 1.246762 -1.1979 3.3021
sUPB_cen | 176 -1.52e-06 1.378339 -1.1411 4.8589
smdis_cen | 176 .000033 1.175343 -1.1342 4.3658
. pwcorr perf_tot eUPB_cen sUPB_cen smdis_cen , star(.05)
| perf_tot eUPB_cen sUPB_cen smdis_~n
-------------+------------------------------------
perf_tot | 1.0000
eUPB_cen | 0.0268 1.0000
sUPB_cen | 0.0128 0.3431* 1.0000
smdis_cen | -0.1620* 0.4357* 0.3108* 1.0000
. sem (eUPB_cen <- sUPB_cen smdis_cen) ( perf_tot <- eUPB_tot smdis_cen), nocapslatent
Endogenous variables
Observed: eUPB_cen perf_tot
Exogenous variables
Observed: sUPB_cen smdis_cen eUPB_tot
Fitting target model:
Iteration 0: log likelihood = -1382.4533
Iteration 1: log likelihood = -1382.4533
Structural equation model Number of obs = 176
Estimation method = ml
Log likelihood = -1382.4533
--------------------------------------------------------------------------------
| OIM
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
Structural |
eUPB_cen <- |
sUPB_cen | .207902 .0626352 3.32 0.001 .0851392 .3306648
smdis_cen | .3864473 .0734531 5.26 0.000 .2424819 .5304127
_cons | 4.25e-06 .0818244 0.00 1.000 -.1603687 .1603772
-------------+----------------------------------------------------------------
perf_tot <- |
smdis_cen | -.2209913 .0847031 -2.61 0.009 -.3870062 -.0549763
eUPB_tot | .1168051 .079851 1.46 0.144 -.0397001 .2733102
_cons | 5.203507 .1969408 26.42 0.000 4.81751 5.589504
---------------+----------------------------------------------------------------
var(e.eUPB_cen)| 1.178362 .1256138 .9561814 1.452169
var(e.perf_tot)| 1.405147 .1497892 1.140206 1.73165
--------------------------------------------------------------------------------
**. estat teffects**
Direct effects
-------------------------------------------------------------------------------
| OIM
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
Structural |
eUPB_cen <- |
sUPB_cen | .207902 .0626352 3.32 0.001 .0851392 .3306648
smdis_cen | .3864473 .0734531 5.26 0.000 .2424819 .5304127
------------+----------------------------------------------------------------
perf_tot <- |
smdis_cen | -.2209913 .0847031 -2.61 0.009 -.3870062 -.0549763
eUPB_tot | .1168051 .079851 1.46 0.144 -.0397001 .2733102
-------------------------------------------------------------------------------
Indirect effects
-------------------------------------------------------------------------------
| OIM
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
Structural |
eUPB_cen <- |
sUPB_cen | 0 (no path)
smdis_cen | 0 (no path)
------------+----------------------------------------------------------------
perf_tot <- |
smdis_cen | 0 (no path)
eUPB_tot | 0 (no path)
-------------------------------------------------------------------------------
Total effects
-------------------------------------------------------------------------------
| OIM
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
Structural |
eUPB_cen <- |
sUPB_cen | .207902 .0626352 3.32 0.001 .0851392 .3306648
smdis_cen | .3864473 .0734531 5.26 0.000 .2424819 .5304127
------------+----------------------------------------------------------------
perf_tot <- |
smdis_cen | -.2209913 .0847031 -2.61 0.009 -.3870062 -.0549763
eUPB_tot | .1168051 .079851 1.46 0.144 -.0397001 .2733102
-------------------------------------------------------------------------------
.
. sem (eUPB_cen <- sUPB_cen smdis_cen) ( perf_tot <- eUPB_tot smdis_cen), **vce(bootstrap,reps(200))
(running sem on estimation sample)**
Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.................................................. 50
.................................................. 100
.................................................. 150
.................................................. 200
Structural equation model Number of obs = 176
Log likelihood = -1382.4533 Replications = 200
--------------------------------------------------------------------------------
| Observed Bootstrap Normal-based
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
Structural |
eUPB_cen <- |
sUPB_cen | .207902 .0846609 2.46 0.014 .0419697 .3738344
smdis_cen | .3864473 .0888138 4.35 0.000 .2123754 .5605192
_cons | 4.25e-06 .0863928 0.00 1.000 -.1693226 .1693311
-------------+----------------------------------------------------------------
perf_tot <- |
smdis_cen | -.2209913 .0755332 -2.93 0.003 -.3690335 -.072949
eUPB_tot | .1168051 .0684562 1.71 0.088 -.0173666 .2509768
_cons | 5.203507 .1882887 27.64 0.000 4.834468 5.572546
---------------+----------------------------------------------------------------
var(e.eUPB_cen)| 1.178362 .1130481 .9763763 1.422133
var(e.perf_tot)| 1.405147 .1716706 1.105932 1.785317
--------------------------------------------------------------------------------
.
* 20221212 上課筆記
* DV(Y): perf_tot
* IV(X): sUPB_cen
* 中介變數：Mediato: eUPB_cen
* 控制變數：control variable: smdis_cen
* data: Study_1.dta
. summ perf_tot eUPB_cen sUPB_cen smdis_cen
pwcorr perf_tot eUPB_cen sUPB_cen smdis_cen , star(.05)
* sem (Y <- X Med) (Med <-X)
sem ( perf_tot <- eUPB_tot smdis_cen) (eUPB_cen <- sUPB_cen smdis_cen)
estat teffects
sem ( perf_tot <- eUPB_cen sUPB_cen) (eUPB_cen <- sUPB_cen )
estat teffects
sem ( perf_tot <- eUPB_cen sUPB_cen) (eUPB_cen <- sUPB_cen smdis_cen) (smdis_cen <- sUPB_cen)
estat teffects
sem (eUPB_cen <- sUPB_cen smdis_cen) ( perf_tot <- eUPB_tot smdis_cen), nocapslatent