我使用stata帮助文档里面自带的数据集和代码可以成功估计潜变量的中介效应 (结果见下)

use http://www.stata-press.com/data/r14/sem_sm2.dta, clear (Structural model with measurement component)
qui sem (F1->educ66 occstat66)(F2->anomia66 pwless66)(F3->anomia67 pwless67)(F2<-F1)(F3<-F1 F2)
medsem, indep(F1) med(F2) dep(F3) stand rit

Significance testing of indirect effect (standardised)
±-------------------------------------------------------------------------+
Estimates | Delta | Sobel | Monte Carlo
|--------------------------------------------------------------------------|
Indirect effect | -0.426 | -0.426 | -0.444

Std. Err. | 0.033 | 0.032 | 0.032

z-value | -12.867 | -13.104 | -13.964

p-value | 0.000 | 0.000 | 0.000

Conf. Interval | -0.491 , -0.361 | -0.489 , -0.362 | -0.499 , -0.397
|--------------------------------------------------------------------------|

Baron and Kenny approach to testing mediation
STEP 1 - F2:F1 (X -> M) with B=-0.524 and p=0.000
STEP 2 - F3:F2 (M -> Y) with B=0.813 and p=0.000
STEP 3 - F3:F1 (X -> Y) with B=-0.134 and p=0.000
As STEP 1, STEP 2 and STEP 3 as well as the Sobel’s test above
are significant the mediation is partial!

RIT = (Indirect effect / Total effect)
(0.426 / 0.560) = 0.760
Meaning that about 76 % of the effect of F1
on F3 is mediated by F2!

±-------------------------------------------------------------------------+
Note: to read more about this package help medsem

但是我用同样的方式估计我自己的模型就报错 (见下)

qui sem (Quality -> b*)(Personality ->c* )(Face -> d*)(Sensitivity -> a* )(Attitude -> e* ) (Action -> f* )(Quality Personality Face->Sensitivity)(Sensi tivity->Attitude Action),standard
SEM结果正常估计, 略
medsem, indep(Quality Personality Face) med(Sensitivity) dep(Attitude Action) stand mcreps(5000) zlc rit rid
Attitude Action:Sensitivity not found
r(111);

请问这是为什么呢?