Yousra DHAINI
May 27, 2020, 7:59:43 AM5/27/20
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to lavaan
Hello world,
Please i find a problem when running a sem model, when i run the summary() it shows the estimation but NAs in all Std.Err values and no z-values ( univariate Normality is not valid but multivariate Normality is accepted and i scaled my dataframe) still don't know where the problem come from.
I wonder if anyone can help me.
Thanks.
Terrence Jorgensen
May 27, 2020, 10:26:55 AM5/27/20
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to lavaan
when i run the summary() it shows the estimation but NAs in all Std.Err values and no z-values
Does it say that the optimizer did not find a solution? Check lavInspect(fit, "converged")
It is possible your model is not identified, but you would need to share more information about your model and data (e.g., a reproducible example with R script).
Terrence D. Jorgensen
Assistant Professor, Methods and Statistics
Research Institute for Child Development and Education, the University of Amsterdam
Message has been deleted
Yousra DHAINI
May 27, 2020, 6:06:52 PM5/27/20
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to lavaan
| 17:09 (il y a 5 heures) |  | ||
| ||||
Yes sir, it doen't converge to a solution ( the result of lavInspect(fit, "converged") is False)
My model specification is as follows:
Well, i'm running a SEM with 3 latent variables( Access,developement and Utilisation) Aiming to creating another latent variable which is financial inclusion using these 3 variables. I tested normality of the distribution of my variables(univariateNormality is accepted for 2 variables but multivariateNormality is accepted by shapiro test p-value=0.05185 ), so i used both the ML and MLR method and i scaled data before running this code.
My model specification is as follows:
model.full<-' # latent variables
Accès =~ Couv.Comm.Rurale+Couv.Comm.Urbain+Couv.Comm.Total+Densite.Res.Bq.Rural+Densite.Res.Bq.Urbain+Densite.Res.Bq.Total+Densite.pointsIOB.MTO.Rural+Densite.pointsIOB.MTO.Urbain+Densite.pointsIOB.MTO.Total
Dev =~ Indice.Rur+Pov.Mon+Pov.Mult+Pov.Glob+Vuln.Urb+Vuln.Rur+Vuln + PIB.hab+DCFM.hab+Poids.Emp.Informel
Utilisation =~ Equ.Comptes.Rural+Equ.Comptes.Urbain +Equ.Comptes.Total+Dep.Moy.Cpte+Dep.PIB+Cred.PIB +DepAV.DCFM+Prop.Femmes.Nouv.Cred +Prop.Jeunes.Nouv.Cred+Prop.Jeunes.Cred +Prop.Femmes.Cred
# residual variances observed variables
Couv.Comm.Rurale ~~ Couv.Comm.Rurale
Couv.Comm.Urbain ~~ Couv.Comm.Urbain
Couv.Comm.Total ~~ Couv.Comm.Total
Densite.Res.Bq.Rural~~ Densite.Res.Bq.Rural
Densite.Res.Bq.Urbain~~ Densite.Res.Bq.Urbain
Densite.Res.Bq.Total~~ Densite.Res.Bq.Total
Densite.pointsIOB.MTO.Rural~~ Densite.pointsIOB.MTO.Rural
Densite.pointsIOB.MTO.Urbain~~ Densite.pointsIOB.MTO.Urbain
Densite.pointsIOB.MTO.Total ~~ Densite.pointsIOB.MTO.Total
Indice.Rur ~~ Indice.Rur
Pov.Mon ~~ Pov.Mon
Pov.Mult ~~ Pov.Mult
Pov.Ext ~~ Pov.Ext
Pov.Glob ~~ Pov.Glob
Vuln.Urb ~~ Vuln.Urb
Vuln.Rur ~~ Vuln.Rur
Vuln ~~ Vuln
PIB.hab ~~ PIB.hab
DCFM.hab ~~ DCFM.hab
Poids.Emp.Informel ~~ Poids.Emp.Informel
Equ.Comptes.Rural ~~ Equ.Comptes.Rural
Equ.Comptes.Urbain ~~ Equ.Comptes.Urbain
Equ.Comptes.Total ~~ Equ.Comptes.Total
Dep.Moy.Cpte ~~ Dep.Moy.Cpte
Dep.PIB ~~ Dep.PIB
Cred.PIB ~~ Cred.PIB
DepAV.DCFM ~~ DepAV.DCFM
Prop.Femmes.Nouv.Cred ~~ Prop.Femmes.Nouv.Cred
Prop.Jeunes.Nouv.Cred ~~ Prop.Jeunes.Nouv.Cred
Prop.Jeunes.Cred ~~ Prop.Jeunes.Cred
Prop.Femmes.Cred ~~ Prop.Femmes.Cred
# factor variances
Accès~ Accès
Dev ~ Dev
Utilisation ~ Utilisation
# factor covariances
Accès~ Dev
Dev ~ Utilisation + Accès
Utilisation ~ Accès + Dev
'Here's the lines of code i runned:
1-ML
fit <- sem(model.full, data=data)
summary(fit, fit.measures = TRUE, standardized = TRUE)
The resut i get:
> fit <- sem(model.full, data=data)
Warning message:
In lav_model_estimate(lavmodel = lavmodel, lavpartable = lavpartable, :
lavaan WARNING: the optimizer warns that a solution has NOT been found!
> summary(fit)
lavaan 0.6-6 did NOT end normally after 10000 iterations
** WARNING ** Estimates below are most likely unreliable
Estimator ML
Optimization method NLMINB
Number of free parameters 69
Number of observations 75
Model Test User Model:
Test statistic NA
Degrees of freedom NA
Parameter Estimates:
Standard errors Standard
Information Expected
Information saturated (h1) model Structured
Latent Variables:
Estimate Std.Err z-value P(>|z|)
Accès =~
Couv.Comm.Rurl 1.000
Couv.Comm.Urbn -0.571 NA
Couv.Comm.Totl 3.468 NA
Dnst.Rs.Bq.Rrl 0.493 NA
Dnst.Rs.Bq.Urb 2.462 NA
Dnst.Rs.Bq.Ttl 7.014 NA
Dnst.IOB.MTO.R 1.088 NA
Dnst.IOB.MTO.U -0.929 NA
Dnst.IOB.MTO.T 4.682 NA
Dev =~
Indice.Rur 1.000
Pov.Mon 1.524 NA
Pov.Mult 1.245 NA
Pov.Glob 1.610 NA
Vuln.Urb 1.374 NA
Vuln.Rur 1.519 NA
Vuln 1.803 NA
PIB.hab -0.788 NA
DCFM.hab -0.822 NA
Pds.Emp.Infrml 0.631 NA
Utilisation =~
Equ.Compts.Rrl 1.000
Equ.Cmpts.Urbn 2.323 NA
Equ.Compts.Ttl 3.486 NA
Dep.Moy.Cpte 3.332 NA
Dep.PIB 3.782 NA
Cred.PIB 3.218 NA
DepAV.DCFM 3.987 NA
Prp.Fmms.Nv.Cr 0.720 NA
Prp.Jns.Nv.Crd 1.374 NA
Prop.Jeuns.Crd -0.392 NA
Prop.Femms.Crd 1.394 NA
Regressions:
Estimate Std.Err z-value P(>|z|)
Accès ~
Accès -3.702 NA
Dev ~
Dev -2.410 NA
Utilisation ~
Utilisation -5.552 NA
Accès ~
Dev -0.242 NA
Dev ~
Utilisation 1.177 NA
Accès 0.618 NA
Utilisation ~
Accès 1.493 NA
Dev -0.020 NA
Variances:
Estimate Std.Err z-value P(>|z|)
.Couv.Comm.Rurl 1.011 NA
.Couv.Comm.Urbn 0.978 NA
.Couv.Comm.Totl 0.720 NA
.Dnst.Rs.Bq.Rrl 0.980 NA
.Dnst.Rs.Bq.Urb 0.848 NA
.Dnst.Rs.Bq.Ttl -0.102 NA
.Dnst.IOB.MTO.R 0.960 NA
.Dnst.IOB.MTO.U 0.966 NA
.Dnst.IOB.MTO.T 0.498 NA
.Indice.Rur 0.634 NA
.Pov.Mon 0.298 NA
.Pov.Mult 0.526 NA
Pov.Ext 0.985 NA
.Pov.Glob 0.213 NA
.Vuln.Urb 0.428 NA
.Vuln.Rur 0.301 NA
.Vuln 0.015 NA
.PIB.hab 0.803 NA
.DCFM.hab 0.786 NA
.Pds.Emp.Infrml 0.868 NA
.Equ.Compts.Rrl 0.919 NA
.Equ.Cmpts.Urbn 0.650 NA
.Equ.Compts.Ttl 0.230 NA
.Dep.Moy.Cpte 0.294 NA
.Dep.PIB 0.095 NA
.Cred.PIB 0.342 NA
.DepAV.DCFM -0.004 NA
.Prp.Fmms.Nv.Cr 0.953 NA
.Prp.Jns.Nv.Crd 0.869 NA
.Prop.Jeuns.Crd 0.976 NA
.Prop.Femms.Crd 0.865 NA
.Accès 0.017 NA
.Dev 0.631 NA
.Utilisation 0.023 NA
> lavInspect(fit, "converged")
[1] FALSE
2-MLR
fit<-sem(model=model.full,data=data,std.lv=TRUE,estimator = "MLR",se = "robust", test = "Satorra-Bentler")
summary(fit, fit.measures = TRUE, standardized = TRUE)
avInspect(fut, "converged")
The results i got:
> fit<-sem(model=model.full,data=data,std.lv=TRUE,estimator = "MLR",se = "robust", test = "Satorra-Bentler")
Error in solve.default(A1) :
system is computationally singular: reciprocal condition number = 2.34478e-37
In addition: Warning messages:
1: In lav_model_hessian(lavmodel = lavmodel, lavsamplestats = lavsamplestats, :
lavaan WARNING: Hessian is not fully symmetric.
Max diff = 112.140213034585
(Max diag Hessian = 291.802586691273)
2: In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats, :
lavaan WARNING:
The variance-covariance matrix of the estimated parameters (vcov)
does not appear to be positive definite! The smallest eigenvalue
(= -2.206684e-10) is smaller than zero. This may be a symptom that
the model is not identified.
>summary(fit, fit.measures = TRUE, standardized = TRUE)
lavaan 0.6-6 did NOT end normally after 10000 iterations
** WARNING ** Estimates below are most likely unreliable
Estimator ML
Optimization method NLMINB
Number of free parameters 63
Number of observations 75
Model Test User Model:
Test statistic NA
Degrees of freedom NA
Parameter Estimates:
Standard errors Sandwich
Information bread Observed
Observed information based on Hessian
Latent Variables:
Estimate Std.Err z-value P(>|z|) Std.lv Std.all
Accès =~
Couv.Comm.Rurl 1.000 1.001 1.007
Couv.Comm.Urbn -0.238 NA -0.238 -0.240
Couv.Comm.Totl 0.632 NA 0.633 0.637
Dnst.Rs.Bq.Rrl 0.306 NA 0.307 0.309
Dnst.Rs.Bq.Urb -0.114 NA -0.115 -0.115
Dnst.Rs.Bq.Ttl 0.961 NA 0.962 0.966
Dnst.IOB.MTO.R 0.204 NA 0.204 0.206
Dnst.IOB.MTO.U -0.948 NA -0.949 -0.952
Dnst.IOB.MTO.T 0.119 NA 0.119 0.120
Dev =~
Couv.Comm.Rurl 1.000 1.058 1.065
Couv.Comm.Urbn 0.043 NA 0.045 0.046
Couv.Comm.Totl 0.129 NA 0.137 0.138
Dnst.Rs.Bq.Rrl 0.263 NA 0.279 0.281
Dnst.Rs.Bq.Urb -0.237 NA -0.250 -0.252
Dnst.Rs.Bq.Ttl -0.889 NA -0.940 -0.944
Dnst.IOB.MTO.R 0.261 NA 0.276 0.278
Dnst.IOB.MTO.U -0.153 NA -0.162 -0.162
Dnst.IOB.MTO.T -0.316 NA -0.335 -0.337
Utilisation =~
Couv.Comm.Rurl 1.000 1.155 1.163
Couv.Comm.Urbn 0.002 NA 0.003 0.003
Couv.Comm.Totl 0.360 NA 0.415 0.418
Dnst.Rs.Bq.Rrl 0.365 NA 0.421 0.424
Dnst.Rs.Bq.Urb 0.235 NA 0.271 0.273
Dnst.Rs.Bq.Ttl 0.744 NA 0.860 0.863
Dnst.IOB.MTO.R 0.183 NA 0.212 0.213
Dnst.IOB.MTO.U 0.547 NA 0.632 0.634
Dnst.IOB.MTO.T 0.308 NA 0.356 0.358
Regressions:
Estimate Std.Err z-value P(>|z|) Std.lv Std.all
Accès ~
Accès -0.702 NA -0.702 -0.702
Dev ~
Dev 1.283 NA 1.283 1.283
Utilisation ~
Utilisation -0.125 NA -0.125 -0.125
Accès ~
Dev -0.007 NA -0.008 -0.008
Dev ~
Utilisation -0.309 NA -0.337 -0.337
Accès -0.289 NA -0.273 -0.273
Utilisation ~
Accès -0.574 NA -0.497 -0.497
Dev -0.014 NA -0.013 -0.013
Variances:
Estimate Std.Err z-value P(>|z|) Std.lv Std.all
.Couv.Comm.Rurl -0.571 NA -0.571 -0.578
.Couv.Comm.Urbn 0.925 NA 0.925 0.937
.Couv.Comm.Totl 0.698 NA 0.698 0.708
.Dnst.Rs.Bq.Rrl 0.833 NA 0.833 0.845
.Dnst.Rs.Bq.Urb 0.783 NA 0.783 0.795
.Dnst.Rs.Bq.Ttl -1.287 NA -1.287 -1.296
.Dnst.IOB.MTO.R 0.904 NA 0.904 0.916
.Dnst.IOB.MTO.U -0.933 NA -0.933 -0.940
.Dnst.IOB.MTO.T 0.715 NA 0.715 0.726
Indice.Rur 0.987 NA 0.987 1.000
Pov.Mon 0.987 NA 0.987 1.000
Pov.Mult 0.987 NA 0.987 1.000
Pov.Ext 0.987 NA 0.987 1.000
Pov.Glob 0.987 NA 0.987 1.000
Vuln.Urb 0.987 NA 0.987 1.000
Vuln.Rur 0.987 NA 0.987 1.000
Vuln 0.987 NA 0.987 1.000
PIB.hab 0.987 NA 0.987 1.000
DCFM.hab 0.987 NA 0.987 1.000
Pds.Emp.Infrml 0.987 NA 0.987 1.000
Equ.Compts.Rrl 0.987 NA 0.987 1.000
Equ.Cmpts.Urbn 0.987 NA 0.987 1.000
Equ.Compts.Ttl 0.987 NA 0.987 1.000
Dep.Moy.Cpte 0.987 NA 0.987 1.000
Dep.PIB 0.987 NA 0.987 1.000
Cred.PIB 0.987 NA 0.987 1.000
DepAV.DCFM 0.987 NA 0.987 1.000
Prp.Fmms.Nv.Cr 0.987 NA 0.987 1.000
Prp.Jns.Nv.Crd 0.987 NA 0.987 1.000
Prop.Jeuns.Crd 0.987 NA 0.987 1.000
Prop.Femms.Crd 0.987 NA 0.987 1.000
.Accès 1.000 0.998 0.998
.Dev 1.000 0.893 0.893
.Utilisation 1.000 0.749 0.749
Warning message:
In .local(object, ...) :
lavaan WARNING: fit measures not available if model did not converge
> lavInspect(fit, "converged")
[1] FALSE
Reply all
Reply to author
Forward
0 new messages