Provides summary details for ctsemFit objects.
# S3 method for ctsemFit
summary(object, ridging = FALSE, timeInterval = 1, verbose = FALSE, ...)ctsemFit object as generated by ctFit.
if TRUE, adds a small amount of variance to diagonals when calculating standardised (correlation) matrices, should only be used if standardised matrices return NAN.
positive numeric value specifying time interval to use for discrete parameter matrices, defaults to 1.
Logical. If TRUE, displays the raw, internally transformed (when fitting with default arguments) OpenMx parameters and corresponding standard errors, as well as
additional summary matrices. Parameter transforms are described in the vignette, vignette('ctsem'). Additional summary matrices
include: 'discrete' matrices -- matrices representing
the effect for the given time interval (default of 1); 'asymptotic' matrices -- represents the effect as time interval
approaches infinity (therefore asymCINT describes mean level of processes at the asymptote, asymDIFFUSION describes total within-
subject variance at the asymptote, etc); 'standardised' matrices -- transforms covariance matrices to correlation matrices, and transforms
discreteDRIFT based on DIFFUSION, to give effect sizes.
additional parameters to pass.
Summary of ctsemFit object
Important: Although ctModel takes cholesky decomposed variance-covariance matrices as input,
the summary function displays the full variance-covariance matrices. These can be cholesky decomposed for comparison purposes using
t(chol(summary(ctfitobject)$covariancematrix)).
Standard errors are displayed in the $ctparameters section, however if ctFit was used with transformedParams=TRUE (the default, and recommended)
covariance matrix standard errors will have been approximated using the delta method. For
inferential purposes, maximum likelihood confidence intervals may be estimated using the ctCI function.
# NOT RUN {
## Examples set to 'donttest' because they take longer than 5s.
# }
# NOT RUN {
### example from Driver, Oud, Voelkle (2015),
### simulated happiness and leisure time with unobserved heterogeneity.
data(ctExample1)
traitmodel <- ctModel(n.manifest=2, n.latent=2, Tpoints=6, LAMBDA=diag(2),
manifestNames=c('LeisureTime', 'Happiness'),
latentNames=c('LeisureTime', 'Happiness'), TRAITVAR="auto")
traitfit <- ctFit(dat=ctExample1, ctmodelobj=traitmodel)
summary(traitfit,timeInterval=1)
# }
Run the code above in your browser using DataLab