Hi there, I am trying got run an GSEM model and I have two variables that come up as 1 (constrained) in the output.
The code is:
gsem (leave <- threat private q12 q10 q46_n, logit) ///
(mit_wcs <- threat private q12 q10 q46_n, logit) ///
(protect_wcs <- threat private q12 q10 q46_n, logit) ///
(threat <- age2 educ q120 q46_n q12 q10 pov_bi) ///
(private <- age2 educ q120 q46_n q12 q10 pov_bi) ///
(threat -> q53_reverse q54_reverse q59_reverse q60_reverse q61_reverse q62_reverse) ///
(private -> q50_reverse q51_reverse q57_reverse q63_reverse) ///
[pweight = weight], latent(threat private) nocapslatent vce(robust)
The relevant output with the problem in bold is:
Log pseudolikelihood = -117269.2
( 1) [leave]threat = 1
( 2) [mit_wcs]private = 1
------------------------------------------------------------------------------------
| Robust
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
leave |
q12 | -.0970142 .4749732 -0.20 0.838 -1.027945 .8339162
q10 | -.133152 .497925 -0.27 0.789 -1.109067 .8427631
q46_n | -.3341718 .4581183 -0.73 0.466 -1.232067 .5637236
threat | 1 (constrained)
private | -1.276096 1.337587 -0.95 0.340 -3.897719 1.345526
_cons | -.624116 .5745823 -1.09 0.277 -1.750277 .5020447
-------------------+----------------------------------------------------------------
mit_wcs |
q12 | -.7155051 .4560428 -1.57 0.117 -1.609333 .1783224
q10 | .8055143 .5845385 1.38 0.168 -.34016 1.951189
q46_n | .2032035 .4261993 0.48 0.634 -.6321318 1.038539
threat | .0189815 .3317193 0.06 0.954 -.6311763 .6691394
private | 1 (constrained)
_cons | .7758378 .3943119 1.97 0.049 .0030007 1.548675
-------------------+----------------------------------------------------------------
protect_wcs |
q12 | -1.054207 .5047535 -2.09 0.037 -2.043506 -.0649082
q10 | 1.11825 .6723743 1.66 0.096 -.1995796 2.436079
q46_n | .4871357 .4865278 1.00 0.317 -.4664413 1.440713
threat | -.2663928 .3895127 -0.68 0.494 -1.029824 .497038
private | .7906234 .5874189 1.35 0.178 -.3606964 1.941943
_cons | 1.59111 .4427315 3.59 0.000 .7233727 2.458848
I very much want threat and private to return useful coefficients, standard errors, and p values in leave and mit_wcs, respectfully. Is there anything I can do to achieve this? Also, why is this happened?
The code is:
gsem (leave <- threat private q12 q10 q46_n, logit) ///
(mit_wcs <- threat private q12 q10 q46_n, logit) ///
(protect_wcs <- threat private q12 q10 q46_n, logit) ///
(threat <- age2 educ q120 q46_n q12 q10 pov_bi) ///
(private <- age2 educ q120 q46_n q12 q10 pov_bi) ///
(threat -> q53_reverse q54_reverse q59_reverse q60_reverse q61_reverse q62_reverse) ///
(private -> q50_reverse q51_reverse q57_reverse q63_reverse) ///
[pweight = weight], latent(threat private) nocapslatent vce(robust)
The relevant output with the problem in bold is:
Log pseudolikelihood = -117269.2
( 1) [leave]threat = 1
( 2) [mit_wcs]private = 1
------------------------------------------------------------------------------------
| Robust
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
leave |
q12 | -.0970142 .4749732 -0.20 0.838 -1.027945 .8339162
q10 | -.133152 .497925 -0.27 0.789 -1.109067 .8427631
q46_n | -.3341718 .4581183 -0.73 0.466 -1.232067 .5637236
threat | 1 (constrained)
private | -1.276096 1.337587 -0.95 0.340 -3.897719 1.345526
_cons | -.624116 .5745823 -1.09 0.277 -1.750277 .5020447
-------------------+----------------------------------------------------------------
mit_wcs |
q12 | -.7155051 .4560428 -1.57 0.117 -1.609333 .1783224
q10 | .8055143 .5845385 1.38 0.168 -.34016 1.951189
q46_n | .2032035 .4261993 0.48 0.634 -.6321318 1.038539
threat | .0189815 .3317193 0.06 0.954 -.6311763 .6691394
private | 1 (constrained)
_cons | .7758378 .3943119 1.97 0.049 .0030007 1.548675
-------------------+----------------------------------------------------------------
protect_wcs |
q12 | -1.054207 .5047535 -2.09 0.037 -2.043506 -.0649082
q10 | 1.11825 .6723743 1.66 0.096 -.1995796 2.436079
q46_n | .4871357 .4865278 1.00 0.317 -.4664413 1.440713
threat | -.2663928 .3895127 -0.68 0.494 -1.029824 .497038
private | .7906234 .5874189 1.35 0.178 -.3606964 1.941943
_cons | 1.59111 .4427315 3.59 0.000 .7233727 2.458848
I very much want threat and private to return useful coefficients, standard errors, and p values in leave and mit_wcs, respectfully. Is there anything I can do to achieve this? Also, why is this happened?
Comment