Skip to contents

semlbci 0.11.2

CRAN release: 2024-06-06

New Feature

  • Added the method "ur" for forming the LBCI. It uses root finding without derivatives. It is inefficient (very slow) because a model with an equality constraint is fitted in each iteration. However, it may be able to find the bound for a parameter when the "wn" method cannot. It also supports robust LBCIs using Satorra (2000) chi-square difference test. (

  • Add get_cibound_status_not_0(). It checks the status of each bound in a semlbci object, and returns as a list the cibound objects of bounds with status not equal to zero. For diagnostic purpose. (

  • When calling semlbci(), users can use lavaan model syntax operators to select parameters. Supported operators are "~", "~~", "=~", and ":=". (

(Possibly) Breaking Changes

  • Disabled some further attempts in the wn method by default as they rarely help but they usually increase the processing time unnecessarily. If necessary, one of the set of attempts, successively lowering the lower limits of variances can be enabled again by setting try_lb to TRUE when calling semlbci() (

  • Added the timeout argument to ci_bound_wn_i(), default to 300 (300 seconds or 5 minutes). Can be used in semlbci() too if the method is "wn" (the default). (


  • Added the option to use load balancing when calling semlbci() with use_pbapply set to TRUE. Enabled by default. (

  • Added support for dynamic scheduling in semlbci() when parallel is FALSE. (

  • Added fit_lb and fit_ub arguments to ci_bound_wn_i() for setting the bounds. Can also be used in semlbci() if the method is "wn" (the defalt). (,

  • In print.semlbci(), added suggestions on what to do if some bounds could not be found (

  • Use whitespace instead of tab to align the output of print.cibound(). (

Bug Fixes

  • Revised ci_bound_wn_i() and ci_bound_ur_i() to make sure the bound is a numeric object, even if NA. (

  • Fixed a bug in print.cibound() when the call is of the form xxx::yyy(). (

  • Fixed a bug with std_lav() for models with only one factor. (

  • Fixed a bug with semlbci(). Intercepts are now automatically skipped if standardized is TRUE, as stated in the help page. (

semlbci 0.10.4

CRAN release: 2023-10-31

New Feature


  • Updated for CRAN release. Identical to the CRAN release in code. (
  • Added an R CMD check for noSuggests. (
  • Revised an example of syntax_to_i() for a coming changes to the parser of lavaan (
  • Updated a few tests (,
  • Finalized to 0.10.4.

semlbci 0.10.3

CRAN release: 2023-05-07

  • First submission to CRAN. (0.10.2)
  • Fixed the description in DESCRIPTION. (0.10.3)
  • Uncommented some lines of code in examples. (0.10.3)
  • Make sure that cat() calls can be suppressed. (0.10.3)

semlbci to

  • First public release. (
  • Added R CMD Check GitHub actions. (
  • Revised a test. (
  • WIP fixed to the documentations. (
  • Simplified the examples of loglike_compare() and helpers. (
  • Used precomputed vignettes. (
  • Added a vignette for loglike_compare(). (
  • Made the examples for loglike_compare(), plot.loglike_compare(), and nearby_levels() run faster. (
  • Fixed a bug that loglike_compare() does not pass use_pbapply to the helper functions. (
  • Added the semlbci_out argument to loglike_compare(). (
  • Added a few arguments to plot.loglike_compare() and fixed the problems with labels. (
  • Added rlang to Imports to avoid the need to set global variables (
  • Updated the vignette for loglike_compare(). (
  • Set try_k_more to 0 such that the examples of loglike_compare() and plot.loglike_compare() can run faster. (
  • Fixed details in documentation. (
  • Fixed some examples in documentation. (
  • Proofread the vignettes and recompute them. (
  • Fixed the pkgdown site. (
  • Used bibentry() in CITATION. (
  • Updated doc. (
  • Added simplified versions of the technical appendices as vignettes. (
  • Updated vignettes. (
  • Changed a technical appendix from a vignette to an article. (
  • Proofread the technical articles. (
  • Updated the sources. (
  • Fixed some typos and tests. (