CRAN release version 1.0.16

What's Changed

• Secondary pruning, and efficient contributions calculation by @Kss2k in #115
• Add pure forward difference mode in fdHESS() by @Kss2k in #116
• Add standardized argument to modsem_inspect_da() by @Kss2k in #120
• Automatically handle ovs in modsem_da() by @Kss2k in #119

Full Changelog: v1.0.15...v1.0.16

CRAN release version 1.0.15

What's Changed

• Dynamically set H0 argument in summary()
• Add plot.jn.points argument to plot_jn()
• Fix wrong axis labels in plot_surface()
• Don't standardize sampling weights with standardize.data=TRUE
• Correctly handle tibble input.
• Check for RCS models in bootstrap_modsem()
• Pass additional argument to FUN in bootstrap_modsem.lavaan()

Full Changelog: v1.0.14...v1.0.15

CRAN release version 1.0.14

Release of version 1.0.14 on CRAN

What's Changed

• Handle NA-modifiers by @Kss2k in #97
• Multigroup models in modsem_da by @Kss2k in #100
• Allow overlapping indicators between etas and xis in LMS by @Kss2k in #104
• Add greyscale option to plot_interaction() and plot_jn() by @Kss2k in #107
• Sampling weights by @Kss2k in #108

Full Changelog: v1.0.13...v1.0.14

CRAN release version 1.0.13

Release of version 1.0.13 on CRAN

What's Changed

• Handle higher order variables in relcorr_single_item() and modsem_da() by @Kss2k in #78
• Include quadratic effects in plotting/margins by @Kss2k in #80
• Fix handling of lavaan labels for interaction terms by @Kss2k in #81
• Improve input validation in plotting functions by @Kss2k in #82
• Add options for setting theme in plot_surface() by @Kss2k in #83
• Handle residual covariances between indicators in x and y by @Kss2k in #84
• Improve handling of residual covariances in relcorr_single_item() by @Kss2k in #85
• Upgrade to C++17 by @Kss2k in #86
• Better optimization of starting parameters of latent mean structures by @Kss2k in #87
• Cleaner calculation of Binv in src/equations_lms.cpp by @Kss2k in #88
• Better scaling of Hessian calculation by @Kss2k in #89
• Add clustered bootstrap by @Kss2k in #90
• Mplus constraints by @Kss2k in #91

Full Changelog: v1.0.12...v1.0.13

CRAN release version 1.0.12

Release of version 1.0.12 on CRAN

Changes

• Fixed error in original equations from QML paper (Klein & Muthén 2007).
• Added cluster robust standard errors (equivalent to TYPE=COMPLEX in Mplus)
• Improved handling of ordinal variables
• Fixed bug in plot_jn
• Improved error handling in modsem_da
• More fit measures available for modsem_da
• Better multithreading in LMS approach
• More efficient calculation of standard errors in LMS approach

CRAN release version 1.0.11

Release of version 1.0.11 on CRAN

Changes

• Fixed typos and formatting in documentation.
• Added correction argument to standardized_estimates.modsem_pi for getting the correct standardized coefficients for interaction terms
• Added centered_estimates function
• Three way interactions are now available in modsem_mplus
• Improved handling of residual covariances between product indicators in the PI approaches, when there are multiple interaction terms in the model.
• Added reliability corrected single items (rcs argument, and relcorr_single_item function).
• Added twostep() function..
• Added auto.split.syntax argument to modsem_da().
• Added modsem_mimpute() function to perform multiple imputation using the DA approaches (modsem_da).
• Added bootstrap_modsem() function to bootstrap modsem models, and function.
• Improved efficiency of quasi-adaptive quadrature for higher dimensions.
• Made adaptive.quad = TRUE the default in modsem_da.

CRAN release version 1.0.10

Release of version 1.0.10 on CRAN

Changes

• Fixed bug in issue #18.

CRAN release version 1.0.9

Release of version 1.0.9 on CRAN

Changes

1. Implemented an Accelerated Expectation Maximum (EMA) algorithm for the LMS approach. Standard EM algorithm can still be used by setting algorithm="EM".
2. Implemented a quasi-adaptive quadrature for the LMS approach. Can be used by setting adaptive.quad=TRUE.
3. Imroved estimation of observed Fischer information when using the LMS approach. Observed Fischer information is now the default.
4. Added significance tests of slopes of x on y at +/- 1 SD of z, in the simple_slopes() function.
5. Added standardized argument to simple_slopes(), plot_interaction(), plot_surface and plot_jn().
6. Implemented compare_fit() and estimate_h0() for modsem_pi objects.
7. Added comparative fit statistics based on baseline model (model without interaction terms) to the summary for modsem_pi objects. Previously only done for modsem_da objects.
8. Added some initial support for multilvel models using the cluster argument, for modsem_pi objects.