insight 0.9.0

New supported model classes

• BGGM (BGGM), metaplus (metaplus), glht (multcomp), glmm (glmm), improved support for manova (stats)

New functions

• Value formatting functions format_bf(), format_pd(), format_p(), format_rope() and format_number() were moved from package parameters to insight.

Changes to functions

• get_variance() now also returns the corrlation among random slopes.
• get_variance() now also (partially) supports brmsfit models.
• get_parameters() for models that return (posterior or simulated) samples of model parameters gains a summary-argument, which - if TRUE - returns a point-estimate (mean of samples) instead of the full samples.
• format_p() returns "> .999" for p-values equal to or greater than 0.999.

Bug fixes

• Fixed issue in find_formula() that did not properly work for models with random effects in formula (in lme4 notation), when random effects were in between fixed effects parts.
• get_variance() did not return variance components for random effects for null-models with random slopes.
• Fixed issue with get_variance() for lme-models with categorical random slope.
• Fixed issue that occured since R 4.0.0 in find_weights() when function call had no weights-argument.
• Fixed issue in get_data() for models with cbind()-response variables and matrix-like variables in the model frame (e.g. when using poly()).
• Fixed issues with PROreg::BBmm(), due to changes in latest package update.

insight 0.8.5

New supported model classes

• robmixglm (robmixglm), betamfx, logitmfx, poissonmfx, probitmfx, negbinmfx, betaor, logitor, poissonirr, negbinirr (mfx), partial support for emmGrid, stanfit and bayesQR.

Changes to functions

• get_varcov.glmmTMB() now also returns the variance-covariance matrix for the dispersion model.
• model_info() returns $is_dispersion = TRUE for glmmTMB objects with dispersion model.
• clean_names() now also removes mathematical operations (like 100 * log(x), which will return "x").
• format_ci() gains a missing argument, as option how to print missing values.
• format_value() now uses NA_character_ as missing if missing = NA.
• format_value() also converts small numbers with many decimals into scientific notation.

insight 0.8.4

General

• HRQoL was removed from suggested packages, as it was removed from CRAN.
• Better support for dispersion models in glmmTMB.

Changes to functions

• null_model() now also works for non-mixed models.
• get_variance() now also computes variance components for models (from mixed models packages) without random effects.
• Improved support for afex_aov and aovlist (i.e. Anova with error term).

Bug fixes

• Fixed some issues with deparsings NULL strings under R 4.0.0.
• Fixed accuracy in get_variance() for models from Gamma family.
• Fixed edge case in clean_names().
• Fixed issues with find_formula.lme() under R 4.0.0.
• Fixed issues with examples from clean_names() under R-devel.

insight 0.8.3

General

• The function to calculate null-models for mixed effects models is now exported (null_model().)

New supported model classes

• arima (stats), averaging (MuMIn)

Changes to functions

• Improve family detection in model_info(), link_inverse() and link_function() for MCMCglmm.
• Minor revisions to meet changes in mlogit package.
• Improve support for bayesx and BBmm models.

Bug fixes

• Fixed issue in find_parameters() and clean_parameters() for brmsfit models with specific variable name patterns.
• Fixed issue in format_ci() when confidence interval only contained NAs and width was set to "auto".
• Fixed issue in find_formula() for mixed models when formula contained parentheses in the non-random parts, around a certain set of predictors.
• Fixed issue in get_priors.BFBayesFactor() for BFMetat class.
• Fixed issue in clean_parameters.BFBayesFactor() when model contained interaction terms and these were assigned to the “extra” component.

insight 0.8.2

Breaking changes

• model_info() now only returns TRUE for $is_ordinal, when model is an ordinal or cumulative link model. In past versions, $is_ordinal was also TRUE for multinomial models.

New supported model classes

• bife (bife), bcplm and zcpglm (cplm)

General

• Improved support for clogit-models.

Bug fixes

• Fixed issue in find_weights() for merMod models.
• Fixed issue in get_data() for models with weights, when weights also contained missing data.
• Fixed issue in get_data() for mixed models with complex offset-terms.
• Fixed issue in get_statistic() for zeroinfl models with theta-coefficients.
• Fixed issue in get_statistic() for lmerModLmerTest models with.
• Fixed issue in find_parameters() for brmsfit models for rare situations where a specific pattern of variables names, when used as random effects, did not properly separate fixed from random effects in the return value.
• Fixed issue related to CRAN checks.

insight 0.8.1

New supported model classes

• cglm (cglm), DirichletRegModel (DirichletReg).

General

• Improved efficiency of find_parameters() and get_parameters() for mixed models with large samples and many random effects, and only fixed effects where requested.

Changes to functions

• model_info() now returns $is_multinomial for multinomial (but not ordinal or cumulative) link models.
• format_value() gets an as_percent argument to format values as percentages.

Bug fixes

• Fixed issue in get_data() for clmm2-models.
• Fixed issue in get_data() for models that used the lspline()-function.
• Fixed issue in get_statistic() for multinom models.
• Fixed issue in get_priors() for stanreg models with flat intercept priors.
• Fixed tests that failed due to latest fixest update.

insight 0.8.0

New supported model classes

• brglm (brglm), cgam, cgamm (cgam), cpglm, cpglmm (cplm), feglm (apaca), glmmadmb (glmmADMB), glmx (glmx), partial support for mcmc (coda), mixor (mixor), MANOVA, RM (MANOVA.RM).

General

• Better handling of clm2, clmm2 and rqss models.

New functions

• format_ci() (re-implemented and slightly enhanced from parameters), to format confidence/credible intervals.

Changes to functions

• find_parameters() now also works for BFBayesFactor objects.
• Suppress non-informative warning in get_data() for model data with weights.
• format_value() automatically uses scientific notation for very large numbers (> 1e+5). Furthermore, the check for integer values was made more robust, to avoid warnings when checking very large numbers for integer type.
• Improved find_parameters(), get_parameters() and clean_parameters() for BFBayesFactor-objects.
• get_priors() now works for stanmvreg objects.
• Other minor improvements.

Bug fixes

• Better detect Tweedie-models in model_info().
• Fixed issue in find_random_slopes() for panelr-models with multiple random-effect parts.
• Fixed issues with zerotrunc models.
• Fixed issues with brmsfit models with correlated random effects.
• Fixed issue with edge-cases in clean_names().
• Fixed issue with breaking changes with latest brms-update.
• Further minor bug fixes.

insight 0.7.1

New supported model classes

• complmrob (complmrob), fixest (fixest), mclogit and mmclogit (mclogit).

Bug fixes

• Fixed bug in find_formula() for mixed models, when random effects are written before any fixed effects terms (like social ~ (1|school) + open + extro).
• Fixed bug in model_info() for VGAM models, where logit-link was not always correctly identified.
• Fixed issue in get_priors() for brmsfit models, where parameters of conditional and zero-inflated model components had identical names. This caused errors in bayestestR::simulate_prior().
• Fixed CRAN check issue.

insight 0.7.0

Breaking changes

• In order to unify column names across easystats-packages, get_parameters() and get_priors() now return column names according to our naming conventions (i.e. capitalized column names).
• model_info() returned both $is_zeroinf and $is_zero_inflated for zero-inflated models. Now $is_zeroinf is softly deprecated, so model_info() will return $is_zero_inflated only in future updates.

New supported model classes

• aareg (survival), brmultinom and bracl (brglm2), and wbgee (panelr). Furthermore, for different model-types from panelr models (within-between, between, etc.) are now also supported.
• Preliminary support for rma models (metafor).

Changes to functions

• get_statistic() supports multinom models (nnet).
• link_inverse() gets a what-argument, to return the link-inverse function for specific distribution parameters from gamls models.

Bug fixes

• Fixed edge case for find_weights().
• Fixed bug in get_statistic() for glmmTMB models that won’t return any data.

insight 0.6.0

New supported model classes

• bayesx (R2BayesX), bamlss (bamlss) and flexsurvreg (flexsurv). Note that support for these models is still somewhat experimental.
• Support for lavaan and blavaan was added, but only applies to some of the functions: get_data(), get_parameters(), find_parameters(), clean_parameters(), find_algorithm() and get_priors() (the two latter only for blavaan).

New functions

• get_statistic() to return the test statistic of model estimates.
• get_varcov() to return the variance-covariance matrix for models.
• supported_models() to print a list of supported models.

Changes to functions

• model_info() now returns the element is_survival for survival models.
• model_info() now returns the element is_truncated for truncated regression, or brmsfit models with trunc() as additional response part.
• model_info() now recognizes beta and beta inflated families from package gamlss.
• Better support for nonlinear quantile regression (quantreg::nlrq()).
• Better support for nonlinear mixed models (lme4::nlmer()). Note that model-specification requires the random term to be written in parentheses, i.e. (slope | group).

Bug fixes

• Fixed issues in get_data(), find_parameters() and get_parameters() for gamlss models.
• Fixed issue in get_data() for plm models, where the index-argument was used in the plm()-function call.
• Fixed issue in get_data(), find_predictors() and find_variables() for brmsfit multi-membership-models.
• is_model() did not recognize objects of class anova and manova.
• model_info() now correctly recognizes censored regression models from brmsfit.
• Fixed issues in find_parameters() and get_parameters() with multinom models.
• Fixed issues in clean_names() for cases where variable transformations where made in specific patterns, like log(test/10).

insight 0.5.0

Breaking Changes

• The previous is_model() function has been renamed to is_model_supported() since it was unclear if the function checked the entered object was a model or a supported model in insight. The new is_model() function checks if the entered object is a model object, while is_model_supported() checks if a supported model object.

New functions

• find_statistic() to return the test statistic of a regression model.
• format_value() and format_table() as utility-functions to format (model) output, especially for tabular output.
• color_if() as utility-function to add color formatting to values, depending on certain conditions.

General

• Make extraction of model family information more stable for gam-objects.

Changes to functions

• find_parameters() and get_parameters() now also support objects of class sim and sim.merMod (from arm::sim()).
• get_variance() now also supports models of class clmm.
• find_predictors() and find_variables() now include the Euclidean distance matrix for spatial models from glmmTMB (returned as random effects element, or more precise, as random slope).

Bug fixes

• find_formula() now extracts group factors of random effects for gamlss models.
• find_parameters() and get_parameters() no longer show NA coefficients from group factors of random effects for gamlss models.
• find_parameters() and get_parameters() did not work for multivariate response models of class brmsfit when argument parameters was specified.
• get_data() dropped value and variable label attributes, when model frame contained matrix variables (like splines).
• get_priors() swapped column names location and scale for brmsfit -objects.
• get_parameters() did not work for glmmTMB models without zero-inflation component.
• find_predictors() did not remove parentheses from terms in multiple nested random effects.
• Better support for gam models (package mgcv) with ziplss or mvn families.

insight 0.4.1

Changes to functions

• get_variance() now supports models with Gamma-family.
• get_weights() and find_weights() now work for brms-models.

Bug fixes

• Fix CRAN-check issues due to recent update from the panelr-package.