I have done a SEM model in laavan. I looked at direct effect of genes on cognition and indirect effect of gene on cognition via mediator (brain volume).
The code is below:
mydata <- apply(mydata, 2, scale)
model <- "# a path
Brain_volume ~ a * genes
# b path
Cognition ~ b * Brain_volume
# c prime path
Cognition ~ cp * genes
# indirect and total effects
ab := a * b
total := cp + ab
"
fit <- sem(model, data = mydata)
fit
summary(fit, fit.measures=TRUE, ci =TRUE)
This is the output:
lavaan 0.6.13 ended normally after 6 iterations
Estimator ML
Optimization method NLMINB
Number of model parameters 5
Used Total
Number of observations 7462 16383
Model Test User Model:
Test statistic 0.000
Degrees of freedom 0
lavaan 0.6.13 ended normally after 6 iterations
Estimator ML
Optimization method NLMINB
Number of model parameters 5
Used Total
Number of observations 7462 16383
Model Test User Model:
Test statistic 0.000
Degrees of freedom 0
Model Test Baseline Model:
Test statistic 473.475
Degrees of freedom 3
P-value 0.000
User Model versus Baseline Model:
Comparative Fit Index (CFI) 1.000
Tucker-Lewis Index (TLI) 1.000
Loglikelihood and Information Criteria:
Loglikelihood user model (H0) -20950.988
Loglikelihood unrestricted model (H1) -20950.988
Akaike (AIC) 41911.975
Bayesian (BIC) 41946.563
Sample-size adjusted Bayesian (SABIC) 41930.674
Root Mean Square Error of Approximation:
RMSEA 0.000
90 Percent confidence interval - lower 0.000
90 Percent confidence interval - upper 0.000
P-value H_0: RMSEA <= 0.050 NA
P-value H_0: RMSEA >= 0.080 NA
Standardized Root Mean Square Residual:
SRMR 0.000
Parameter Estimates:
Standard errors Standard
Information Expected
Information saturated (h1) model Structured
I have numerous missing values, and when I added missing="fiml" this is the output I got (the error msgs are highlighted in bold):
Warning: lavaan WARNING: 6456 cases were deleted due to missing values in exogenous variable(s), while fixed.x = TRUE.
lavaan 0.6.13 ended normally after 7 iterations
Estimator ML
Optimization method NLMINB
Number of model parameters 7
Used Total
Number of observations 9927 16383
Number of missing patterns 2
Model Test User Model:
Test statistic 0.000
Degrees of freedom 0
Error in solve.default(X[[i]], ...) : Lapack routine dgesv: system is exactly singular: U[3,3] = 0
This is the traceback output:
10.
solve.default(X[[i]], ...)
9.
FUN(X[[i]], ...)
8.
lapply(Wm, solve)
7.
lav_fit_fiml_corrected(lavobject, version = version)
6.
lav_fit_cfi_lavobject(lavobject = object, fit.measures = fit.measures,
baseline.model = baseline.model, standard.test = standard.test,
scaled.test = scaled.test)
5.
lav_fit_measures(object, fit.measures = "default", fm.args = fm.args)
4.
lav_object_summary(object = object, header = header, fit.measures = fit.measures,
estimates = estimates, ci = ci, fmi = fmi, std = std, standardized = standardized,
remove.step1 = remove.step1, cov.std = cov.std, rsquare = rsquare,
std.nox = std.nox, efa = efa.flag, fm.args = fm.args, modindices = modindices)
3.
.local(object, ...)
2.
summary(fit, fit.measures = TRUE)
1.
summary(fit, fit.measures = TRUE)
I understand the error but I don't understand why is it occuring with missing="fiml" when it's not occuring without it.
UPDATE: This analysis was run on a Linux machine. However, I don't get this error message when I ran the same script on my Windows machine.
Could this be related to some package? Can anyone please help?
Thanks!
