Bug In summary() function

[Brandon McCormick]

I posted previously about an error in the summary function, but I found out more info that may help trace the bug. 

I am running a mediation analysis with a binary predictor, three latent mediators, and two outcome variables. I also have 3 latent covariates and 3 other covariates with n = 3324. The model converged normally. My model used multiple reputations. So the fit object is a lavaan.mi

Previously, when I called the summary() function, I received this error 

Error in dimnames(W) <- dimnames(B) : 'dimnames' applied to non-array

I updated all packages to make sure that wasn't the problem and restarted R. 

Now I can call the Summary() function and I get a list of parameters with an estimated value next to each. If I try to call any additional arguments such as standardized = TRUE or fit.measures = TRUE, I get an error stating they are unused arguments. I get this even when I call lavaan::summary(fit, standardized = TRUE). If I call parameterestimates(), I can get the estimates and I can also get the standardized values if I call standardized = TRUE. However, if I add additional arguments such as se = TRUE, I do not get any additional information. I am able to get fit measures with the fitMeasures() function. 

I tried to replicate the error in the HS data set (with and without mi), but could not get it to replicate. Therefore, I am thinking it is a problem with the size of the lavaan.mi object (about 250 MB).

Here is my code:

library(lavaan)

library(semTools)

library(Amelia)

set.seed(6819)

mTBI_SEM_Data_Amelia <- amelia(mTBI_SEM_Data, m = 40, noms = c("Ethnicity", "HI_Bi_B21", "Sex"), sqrts  = c("NV1", "NV2", "NV3", "NV4", "NV5", "NV6", "NV7", "NV8", "NV9", "NV10", "NV11", "NV12", "NV13", "NV15", "NV16", "NV17", "NV18", "NV19", "NV20", "NV21", "VO1", "VO2", "VO3", "VO4", "VO5", "VO6", "VO7", "VO8", "VO9", "VO10", "VO11"), empri = .01*nrow(mTBI_SEM_Data))

mTBI_SEM_Data_Amelia <- transform(mTBI_SEM_Data_Amelia, NVtot = NV1 + NV2 + NV3 + NV4 + NV5 + NV6 + NV7 + NV8 + NV9 + NV10 + NV11 + NV12 + NV13  + NV15 + NV16 + NV17 + NV18 + NV19 + NV20 + NV21)

mTBI_SEM_Data_Amelia <- transform(mTBI_SEM_Data_Amelia, VOtot = VO1 + VO2 + VO3 + VO4 + VO5 + VO6 + VO7 + VO8 + VO9 + VO10 + VO11)

mTBI_SEM_Data_Amelia <- transform(mTBI_SEM_Data_Amelia, VOtot_log = log(VOtot +1))

mTBI_SEM_Data_Amelia <- transform(mTBI_SEM_Data_Amelia, NVtot_log = log(NVtot +1))

mTBI_SEM_Data_Imps <- mTBI_SEM_Data_Amelia$imputations

mTBI_Off_Med2 <- '

  

#Wave 1 LV’s 

  

# IM_W1 

  

IM_W1 =~ NA* IM1_W1 + IM2_W1 + IM3_W1 + IM4_W1 + IM5_W1

IM_W1 ~~ 1* IM_W1

IM1_W1 ~~ IM2_W1

IM2_W1 ~~ IM5_W1

IM3_W1 ~~ IM5_W1

#AD_W1  

AD_W1 =~ NA* AD1_W1 + AD2_W1 + AD3_W1 + AD4_W1 + AD5_W1 + AD6_W1 + AD7 _W1 + AD8_W1 + AD9 _W1 + AD10_W1 + AD11_W1 + AD12_W1 + AD13_W1 + AD14_W1

AD_W1 ~~ 1* AD_W1

AD1_W1 ~~  AD2_W1 + AD6_W1 + AD7_W1 + AD11_W1

AD3_W1 ~~  AD4_W1 + AD7_W1 + AD8_W1 + AD10_W1 + AD14_W1 

AD4_W1 ~~ AD9_W1

AD5_W1 ~~  AD2_W1 + AD4_W1 

AD6_W1 ~~  AD7_W1

AD8_W1 ~~ AD4_W1 + AD5_W1 + AD9_W1 + AD10_W1

AD11_W1 ~~ AD12_W1

AD10_W1 ~~ AD9_W1 + AD12_W1 + AD13_W1

AD13_W1 ~~ AD4_W1 + AD12_W1

ASB_CUT_W1 =~ NA* CU1_W1 + CU2_W1 + CU3_W1 + CU4_W1 + ASB1_W1 + ASB2_W1 + ASB3_W1 + ASB4_W1 + ASB5_W1 + ASB6_W1 + ASB7_W1 + ASB8_W1 +  ASB9_W1 + ASB10_W1

ASB_CUT_W1  ~~ 1* ASB_CUT_W1

ASB2_W1 ~~ ASB1_W1 + ASB4_W1 + ASB8_W1 + ASB10_W1 + CU1_W1 

ASB3_W1 ~~  ASB2_W1 + ASB8_W1  + ASB10_W1 + CU1_W1

ASB4_W1 ~~  ASB5_W1

ASB8_W1 ~~ ASB6_W1 + ASB9_W1 + ASB10_W1 + CU1_W1

CU2_W1 ~~ ASB1_W1 + ASB6_W1 + ASB9_W1 + CU1_W1 

CU3_W1 ~~  ASB2_W1 + ASB3_W1 + ASB8_W1 + ASB10_W1 + CU1_W1 + CU4_W1

CU4_W1 ~~ ASB1_W1 + ASB3_W1 + ASB10_W1 + ASB2_W1 

#Based on MI

IM1_W1 ~~  IM3_W1 + ASB1_W1 + ASB2_W1 + AD7_W1 + AD14_W1 

IM2_W1 ~~  IM4_W1  + AD3_W1 + AD5_W1 + AD9_W1 + AD10_W1  + AD12_W1 + AD14_W1 + ASB4_W1 + ASB7_W1 + ASB2_W1 + ASB1_W1 

IM3_W1 ~~  ASB2_W1 + AD13_W1 + CU2_W1

IM4_W1 ~~ IM5_W1  + AD3_W1 + AD4_W1 + AD5_W1 + ASB1_W1 + ASB5_W1 + ASB6_W1 + AD1_W1 + AD8_W1 + AD10_W1 + AD9_W1 + CU2_W1 + AD13_W1

IM5_W1 ~~ ASB2_W1 + AD12_W1 + AD14_W1 

AD2_W1 ~~ ASB7_W1 + AD8_W1

AD1_W1 ~~  ASB4_W1

AD3_W1 ~~ CU1_W1 + CU2_W1 + AD5_W1 + ASB10_W1 + CU3_W1

AD4_W1 ~~  ASB7_W1 + CU2_W1

AD5_W1 ~~ CU2_W1 + ASB1_W1 + ASB3_W1 + ASB4_W1 + ASB6_W1 + ASB5_W1 + ASB7_W1 + ASB8_W1 + ASB9_W1 

AD11_W1 ~~ AD6_W1 + AD7_W1 + ASB10_W1

AD8_W1 ~~ ASB1_W1 + ASB4_W1 + ASB7_W1  + CU2_W1

AD7_W1 ~~ ASB1_W1 + ASB7_W1

AD9_W1 ~~ ASB2_W1 + ASB3_W1+ ASB10_W1 

AD10_W1 ~~ AD11_W1 + ASB9_W1 +  ASB10_W1

AD12_W1 ~~ ASB3_W1 + ASB8_W1

AD13_W1 ~~ CU2_W1 + ASB4_W1 + ASB6_W1

AD14_W1 ~~ AD12_W1+ ASB1_W1 + ASB8_W1

ASB5_W1 ~~ ASB9_W1 

ASB3_W1 ~~ ASB9_W1

#Wave 2 LV’s 

#IM_W2

IM_W2 =~ NA* IM1_W2 + IM2_W2 + IM3_W2 + IM4_W2 + IM5_W2

IM_W2 ~~ 1* IM_W2

IM2_W2 ~~ IM4_W2

IM1_W2 ~~ IM4_W2

#AD_W2

AD_W2 =~ NA* AD1_W2 + AD2_W2 + AD3_W2 + AD4_W2 + AD5_W2 + AD6_W2 + AD7_W2 + AD8_W2 + AD9_W2 + AD10_W2 + AD11_W2 + AD12_W2 + AD13_W2 + AD14_W2

AD_W2 ~~ 1* AD_W2

AD2_W2 ~~  AD1_W2 + AD11_W2

AD3_W2 ~~ AD4_W2 + AD10_W2 + AD14_W2

AD4_W2 ~~  AD5_W2

AD7_W2 ~~ AD6_W2 + AD12_W2 + AD14_W2

AD8_W2 ~~ AD4_W2 + AD9_W2 + AD10_W2

AD9_W2 ~~ AD10_W2 + AD13_W2

AD12_W2 ~~ AD14_W2

#ASB_CUT_W2

ASB_CUT_W2 =~ NA* CU1_W2 + CU2_W2 + CU3_W2 + CU4_W2 + ASB1_W2 + ASB2_W2 + ASB3_W2 + ASB4_W2 + ASB5_W2 + ASB6_W2 + ASB7_W2 + ASB8_W2 +  ASB9_W2 + ASB10_W2

ASB_CUT_W2  ~~ 1* ASB_CUT_W2

CU1_W2 ~~   CU4_W2

CU2_W2 ~~  CU1_W2 + CU3_W2 + CU4_W2 + ASB9_W2

CU3_W2 ~~  CU1_W2 + CU4_W2

ASB2_W2 ~~  ASB1_W2 + ASB4_W2

ASB5_W2 ~~  ASB9_W2

ASB6_W2 ~~ ASB3_W2 +  ASB7_W2 + ASB8_W2 + ASB10_W2

ASB7_W2 ~~ ASB3_W2 + ASB4_W2 + ASB3_W2 + ASB10_W2 

#based on MI

IM1_W2 ~~   IM3_W2 + IM2_W2 + ASB2_W2 

IM2_W2 ~~   AD3_W2 + AD10_W2 + AD14_W2

IM3_W2 ~~ AD4_W2

IM4_W2 ~~   AD3_W2 + AD4_W2 + AD6_W2 + ASB6_W2 

IM5_W2 ~~   IM3_W2 + AD5_W2

ASB2_W2 ~~  ASB5_W2

ASB3_W2 ~~  AD3_W2  + AD9_W2 

AD3_W2 ~~  CU2_W2

AD7_W2 ~~  AD8_W2 + CU1_W2

AD12_W2 ~~  ASB9_W2

AD14_W2 ~~ AD11_W2 + ASB2_W2 + ASB8_W2

ASB7_W2 ~~ AD4_W2 + AD5_W2 + AD7_W2 + AD8_W2

#correlate LV’s W2

IM_W2 ~~ c* AD_W2 + ASB_CUT_W2

AD_W2 ~~ ASB_CUT_W2

#correlate LV’s W1

IM_W1 ~~ c* AD_W1 + ASB_CUT_W1

AD_W1 ~~ ASB_CUT_W1

#Regression paths 

#direct effects

NVtot_log ~ C1 *  HI_Bi_B21 

VOtot_log ~ C2 * HI_Bi_B21 

#Mediation: A paths

IM_W2 ~ A1 * HI_Bi_B21 

AD_W2 ~ A2 * HI_Bi_B21 

ASB_CUT_W2 ~ A3*  HI_Bi_B21 

#Mediation: B paths

NVtot_log ~ B1_1 * IM_W2

NVtot_log ~ B2_1 * AD_W2

NVtot_log ~ B3_1 * ASB_CUT_W2 

VOtot_log ~ B1_2 * IM_W2

VOtot_log ~ B2_2 * AD_W2

VOtot_log ~ B3_2 * ASB_CUT_W2 

#Effect estimates

indirect1_1 := A1 * B1_1

indirect2_1 := A2 * B2_1

indirect3_1 := A3 * B3_1

indirect1_2 := A1 * B1_2

indirect2_2 := A2 * B2_2

indirect3_2 := A3 * B3_2

total1_1 := C1 + (A1 * B1_1)

total2_1 := C1 + (A2 * B2_1)

total3_1 := C1 + (A3 * B3_1)

total1_2 := C2 + (A1 * B1_2)

total2_2 := C2 + (A2 * B2_2)

total3_2 := C2 + (A3 * B3_2)

#control paths: W1 LV’s

IM_W2 ~ IM_W1

AD_W2 ~ AD_W1

ASB_CUT_W2 ~ ASB_CUT_W1

NVtot_log ~ IM_W1 + AD_W1 + ASB_CUT_W1

VOtot_log ~ IM_W1 + AD_W1 + ASB_CUT_W1

#control paths: other vars

IM_W2 ~ Age_W3 + Sex + Ethnicity + Neighborhood_SES + PC_EDU

AD_W2 ~ Age_W3 + Sex + Ethnicity + Neighborhood_SES + PC_EDU

ASB_CUT_W2 ~ Age_W3 + Sex + Ethnicity + Neighborhood_SES + PC_EDU

NVtot_log ~ Age_W3 + Sex + Ethnicity + Neighborhood_SES + PC_EDU

VOtot_log ~ Age_W3 + Sex + Ethnicity + Neighborhood_SES + PC_EDU'

fitmTBI_Off_Med2 <- sem.mi(mTBI_Off_Med2 , data = mTBI_SEM_Data_Imps)

summary(fitmTBI_Off_Med2 , standardized=TRUE, fit.measures=TRUE, ci=TRUE)

[Terrence Jorgensen]

I posted previously about an error in the summary function

Please don't start a new thread about the same topic.  It makes it difficult to keep responses consistent.

 

I am thinking it is a problem with the size of the lavaan.mi object (about 250 MB).

That's pretty big, and you have quite a large model.  You could test this hypothesis by fitting a smaller part of your model (fewer variables) to a subset of your data (smaller N).

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen

 

[Brandon McCormick]

I apologize for the multiple posts (I will delete the original), this was the result of the sessioninfo():

R version 4.0.0 (2020-04-24)

Platform: x86_64-apple-darwin17.0 (64-bit)

Running under: macOS Mojave 10.14.6

Matrix products: default

BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib

LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib

Random number generation:

 RNG:     Mersenne-Twister 

 Normal:  Inversion 

 Sample:  Rounding 

 

locale:

[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:

[1] Amelia_1.7.6   Rcpp_1.0.4.6   semTools_0.5-3 lavaan_0.6-6  

loaded via a namespace (and not attached):

[1] compiler_4.0.0 tools_4.0.0    foreign_0.8-78 mnormt_1.5-7   pbivnorm_0.6.0 stats4_4.0.0  

I am relatively new to SEM, would it be better to test each mediation separately (three separate models)? (or do you happen to know any good papers who have used a similar method I could reference). 

[David Simmonds]

That's a really complicated model. I would start with the assumption that there's something problematic on the model, not R. Did you get small pieces working by themselves?

Best Regards,
David

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/8832757d-de2e-40d4-8d6a-21b79f77a77fo%40googlegroups.com.

[Brandon McCormick]

Yes, the model works up until the final model with covariates. It converges normally, but won’t display normally. fitMeasures() indicates good fit RMSEA = .029, CFI = .90

On Jun 4, 2020, at 5:53 PM, David Simmonds <davidms...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/CAMxdSn-JC7oioXf3ZJ%2BMRJ0hV2dT_vqVkvoULhnva0AxG83%2Bgg%40mail.gmail.com.

[David Simmonds]

Did you try to display the path model, using semPaths?

Best Regards,
David

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/E60FA62C-6D90-4B44-BE11-D7504E958386%40gmail.com.

[Terrence Jorgensen]

Error in dimnames(W) <- dimnames(B) : 'dimnames' applied to non-array

This is strange.  Can you try vcov(fitmTBI_Off_Med2, type = "within") and vcov(fitmTBI_Off_Med2, type = "between"), to see whether either of those return something unexpected?  (i.e., in both cases we expect a matrix with as many rows/columns as there are estimated parameters)

[Brandon McCormick]

This is what I got

vcov(fitmTBI_Off_Med2, type = "within")

Error in UseMethod("vcov") :
  no applicable method for 'vcov' applied to an object of class "c('lavaan.mi', 'lavaanList')"
vcov(fitmTBI_Off_Med2, type = "between")
Error in UseMethod("vcov") :
  no applicable method for 'vcov' applied to an object of class "c('lavaan.mi', 'lavaanList')"

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/f85c30f8-f3aa-49b6-8391-98f6d5a5406fo%40googlegroups.com.

--

Brandon McCormick

Doctoral Student

Clinical Psychology - Psychology and Law 
The University of Alabama
101 McMillan
Tuscaloosa, AL 35401 
Phone 205-460-8678
b.mccorm...@gmail.com

[Brandon McCormick]

David, 

It seems SemPaths doesn't accept lavaan.mi objects. But I was able to display it in lavaanPlot

[Terrence Jorgensen]

Error in UseMethod("vcov") : 

  no applicable method for 'vcov' applied to an object of class "c('lavaan.mi', 'lavaanList')"

I can't think of why this might occur.  I am using the same version of R on the same OS as you.  This makes it seem that R isn't finding the right information in semTools NAMESPACE.  When you run methods("vcov"), you should see vcov,lavaan.mi-method in the vector of results.  

Is it possible you are loading the lavaan.mi object into your workspace, which was created in an older version of R or semTools or lavaan, then saved to a file so you could save time by simply loading into your workspace?

[Brandon McCormick]

I think that could be possible, I can try creating a new workspace. 

On Jun 5, 2020, at 5:14 AM, Terrence Jorgensen <tjorge...@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/04f3d3fa-36ef-47e2-ba5e-95dc569952bdo%40googlegroups.com.

[Brandon McCormick]

I created a new workspace, and re-ran everything. 

I still received these errors:

summary(fitmTBI_Off_Med2)

Error in dimnames(W) <- dimnames(B) : 'dimnames' applied to non-array

vcov(fitmTBI_Off_Med2, type = "within")

Error in dimnames(W) <- dimnames(B) : 'dimnames' applied to non-array

This is the session info

R version 4.0.0 (2020-04-24)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Mojave 10.14.6

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base    

other attached packages:
[1] Amelia_1.7.6   Rcpp_1.0.4.6   semTools_0.5-3 lavaan_0.6-6  

loaded via a namespace (and not attached):

[1] MASS_7.3-51.5  compiler_4.0.0 parallel_4.0.0 tools_4.0.0    foreign_0.8-78 mnormt_1.5-7  
[7] pbivnorm_0.6.0 stats4_4.0.0  

[David Simmonds]

Brandon,

Just out of curiosity, I would love to see the path diagram. I guess I should have mentioned that previously. 

Best Regards,
David

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/CAF4pV9zvqk4aetQhD6i61ya5DjG%3DM5-J1ApS%2B%3DmEfcKXncYmBg%40mail.gmail.com.

[Terrence Jorgensen]

I created a new workspace, and re-ran everything. I still received these errors:

Then I would need a reprex (with enough data to reproduce the error that I cannot replicate) to track down the problem.  You can look into dump() to simply send me any objects as R syntax files.  

[Brandon McCormick]

Hi Terrence, 

I emailed you privately, 

Brandon

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/124316e4-3ae1-43ba-814b-095bb65cc6f1o%40googlegroups.com.

[Terrence Jorgensen]

Okay, your original problem somehow became a different problem.

When I load your workspace, sessionInfo() shows me this list of packages (note that my version numbers might differ from yours)

attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base    

loaded via a namespace (and not attached):

 [1] compiler_4.0.0     emmeans_1.4.7      estimability_1.3   tools_4.0.0      

 [5] coda_0.19-3        lavaan_0.6-6       mvtnorm_1.1-0      mnormt_1.5-7      

 [9] grid_4.0.0         pbivnorm_0.6.0     xtable_1.8-4       stats4_4.0.0      

[13] semTools_0.5-3.906 lattice_0.20-41   

Whereas a fresh session in RStudio shows me only these namespaces loaded:

loaded via a namespace (and not attached):

[1] compiler_4.0.0 tools_4.0.0   

I ran the help-page example and got the later problems you described (i.e., the summary() showed only "est.ave" because it used the class?lavaanList method rather than the class?lavaan.mi method, and R could not find vcov() or any other methods for lavaan.mi objects).  So I saved your lavaan.mi object by itself, rather than within a workspace: 

saveRDS(fitmTBI_Off_Med2, "lavMI.rds")

Then restarted R and tried just loading semTools and imported your object:

library(semTools)
fit <- readRDS("lavMI.rds")
summary(fit)

Now I got your original error about dimnames(W).  Unlike you, I was perfectly able to run

vcov(fit, type = "within") # Error in dimnames(W) <- dimnames(B) : 'dimnames' applied to non-array
vcov(fit, type = "between") # no problem, because this is calculated before W (within)

So the later problem (that your R could not find the right vcov() method) must be due to your attached namespaces, which I can only guess would be fixed if you no longer load anything but semTools before trying to inspect your results.  You could try removing one loaded package at a time to see which one is the culprit -- I would be interested to know.

But the first problem (finally) is that the within-imputations VCOV cannot be calculated because your model object did not save the VCOV from each individual imputation.  Here is what you can see in your object's first imputation:

fit@vcovList[[1]]

$se

[1] "standard"

$information

[1] "expected" "expected"

$vcov

NULL

The $vcov element should be a matrix, the output of vcov(lavaan(model = parTable(fit), data = fit@DataList[[1]])) , which you get by fitting the model to your first imputation only in lavaan.  That works fine, so I'm not sure why it wasn't saved in any of your imputations when you fit the model using runMI():

str(lapply(fit@vcovList, "[[", i = "vcov"))

List of 40

 $ : NULL

 $ : NULL

     ... (all 40)

 $ : NULL

 $ : NULL

I tried restarting R again to re-fit your model using the information inside the object, to see if it saved the $vcov element this time:

library(semTools)
fit <- readRDS("lavMI.rds")
fit2 <- lavaan.mi(model = parTable(fit),
                  data = lapply(fit@DataList[1:3], head, n = 500), # subset already takes several minutes
                  show.progress = TRUE)
fit2@vcovList

It did not.  So I tried using lavaan() on the individual data sets:

fitList <- lapply(lapply(fit@DataList[1:3], head, n = 500),
                  function(dat) lavaan(model = parTable(fit), data = dat))
lapply(fitList, function(x) vcov(x)[1:5, 1:5])

vcov() is available from each imputation, so I think there is something going wrong inside lavaanList().  I wonder if it is related to your model being so massive (380 parameters means VCOV is 380x380), or that your variables appear to be profoundly nonnormal:

hist(fit@DataList[[1]]$CU1_W2) # for example

[Brandon McCormick]

Thank you for unpacking all of this, I will do some further investigating!

On Jun 14, 2020, at 4:59 AM, Terrence Jorgensen <tjorge...@gmail.com> wrote:



<Auto Generated Inline Image 1>

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/88a632be-fe29-428f-a324-683560d18bf1o%40googlegroups.com.

<Auto Generated Inline Image 1>