Tools for Model Specification in the Latent Variable Framework

Tools for model specification in the latent variable framework (add-on to the 'lava' package). The package contains three main functionalities: Wald tests/F-tests with improved control of the type 1 error in small samples, adjustment for multiple comparisons when searching for local dependencies, and adjustment for multiple comparisons when doing inference for multiple latent variable models.


-- mode: org --

  • Version 1.4 <2018-09-21 Fri> ** User visible change
  • modelsearch2 is now only based on the score statistic.
  • Version 1.3.5 <2018-05-01 Tue> ** User visible change
  • add arguments df, cluster to glht2
  • argument display is now named print in summary

** Internal change

  • remove dependency on tcltk for displaying progress bar with parallel computation.
  • prepare for update of ggplot 2.3.0
  • Version 1.3.4 <2018-04-23 Mon> ** User visble changes
  • =calibrateType1= is now a method and support parallel computation
  • =summary2=, =compare2=, and =iid2= for =lvmfit= object can compute cluster robust standard error.
  • Version 1.3.3 ** User visble changes
  • add option Ftest to =calibrateType1=.
  • Version 1.3.2 ** Internal changes
  • fix bugs
  • Version 1.3.1 ** User visble changes
  • New feature: =calibrateType1= to perform simulation studies of the type 1 error rate for Wald test in LVM.
  • Version 1.3.0 <2018-04-04 Wed> ** User visble changes
  • =summary2= can be called directly.

** Internal changes

  • identification of the small sample bias of the variance-covariance parameters. This enable to correct the derivatives and not only the estimation of the conditional variance.
  • =sCorrect= is able to handle missing values
  • reorganisation of the structure of the functions in =sCorrect=. Only =sCorrect= do real computations and all the results are stored in the object (slot sCorrect). The other functions (=summary2=, =compare2=, =iid2=, =score2=, =residuals2=) only extract and combine information from the object.
  • Version 1.2.0 <2018-16-03 Fri>

** Internal changes

  • iterative estimation of the small sample bias
  • Version 1.1.0 <2018-02-01 Thu> ** User visible changes
  • The argument =numericDerivative= has been renamed =numeric.derivative=.
  • The argument =adjust.residuals= has been renamed =bias.correct=.
  • The method =dVcov2= has been renamed =sCorrect=.
  • The method =lTest= is now replaced by the method =compare2=. =compare2= is similar to =lava::compare= but with small sample correction. The argument =C= of =lTest= is now =contrast= in =compare2=.
  • The interaction with the multtest package has been re-organized. There is no more =mlf2= or =mmm2= function. When no small sample correction is needed use =glht=, otherwise =glht2=.
  • The function =createContrast= has been improved.

** Internal changes

  • When possible argument x has been converted to object.
  • Dependency on several packages has been moved from Import to Suggest. This leads to several requireNamespace in the code to check the presence of packages.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.