extract_random_coefs: Extract random coefficients and their variances

In m-clark/mixedup: Miscellaneous functions for mixed models

Extract random coefficients and their variances

Description

Fixed effect + random effects.

Usage

extract_random_coefs(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = NULL,
  ...
)

## S3 method for class 'merMod'
extract_random_coefs(model, re = NULL, ci_level = 0.95, digits = 3, ...)

## S3 method for class 'glmmTMB'
extract_random_coefs(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = "cond",
  ...
)

## S3 method for class 'lme'
extract_random_coefs(model, re = NULL, ci_level = NULL, digits = 3, ...)

## S3 method for class 'brmsfit'
extract_random_coefs(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = NULL,
  ...
)

## S3 method for class 'stanreg'
extract_random_coefs(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = NULL,
  ...
)

## S3 method for class 'gam'
extract_random_coefs(model, re = NULL, ci_level = 0.95, digits = 3, ...)

extract_coef(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = NULL,
  ...
)

extract_random_coefficients(
  model,
  re = NULL,
  ci_level = 0.95,
  digits = 3,
  component = NULL,
  ...
)

Arguments

model	A merMod, nlme, brms, or glmmTMB object
re	The name of the grouping variable for the random effects.
ci_level	Where possible, confidence level < 1, typically above 0.90. A value of 0 will not report it. Default is .95. Not applicable to nlme objects.
digits	Rounding. Default is 3.
component	For glmmTMB objects, which of the two components 'cond' or 'zi' to select. Default is 'cond'. For brmsfit objects, this can filter results to a certain part of the output, e.g. 'sigma' or 'zi' of distributional models, or a specific outcome of a multivariate model. In this case component is a regular expression that ends the name of the parameters of the output (e.g. '__component').
...	Other arguments specific to the method. For example add_group_N for extract_random_effects. Will not apply to brmsfit or stanreg models. Experimental.

Details

Returns a data frame with random coefficients, a.k.a. random intercepts and random slopes, and their standard errors. r Note that the standard errors assume independence of the conditional variance and the fixed-effects variance, thus the standard errors are the sum of variances for the respective fixed and random effects. See Bolker's demo here and additional discussion at the GLMM FAQ. As noted there, this assumption may not be, and likely is not, appropriate, and if you are really interested in an accurate uncertainty estimate you should probably use brms.

Please realize that this functionality is likely only appropriate for simpler GLMM type models, and is mostly just a shortcut for those settings. It may work for more complicated situations also, but I don't make any guarantees. For more complex models that include multiple outcomes/categories or have other anomalies, this function likely will not work even if the underlying extract_fixed_effects and extract_random_effects do, as naming conventions are not consistent enough within the relative packages to deal with this in a general way. I will continue to look into its feasibility, but don't expect much.

Value

A data frame of the random coefficients and their standard errors.

Note

The nlme package only provides the coefficients with no estimated variance, so this function doesn't add to what you get from basic functionality for those models. In addition, nlme adds all random effects to the fixed effects, whereas lme4 and others only add the effects requested.

For multicomponent glmmTMB models, e.g. zip, please specify the component argument.

extract_coef and extract_random_coefficients are aliases.

See Also

Other extract: extract_cor_structure(), extract_fixed_effects(), extract_het_var(), extract_model_data(), extract_random_effects(), extract_vc()

Examples

library(lme4)
library(mixedup)

lmer_1 <- lmer(Reaction ~ Days + (1 | Subject), data = sleepstudy)
extract_random_coefs(lmer_1, re = 'Subject')
library(glmmTMB)
tmb_1 <- glmmTMB(Reaction ~ Days + (1 | Subject), data = sleepstudy)
extract_random_coefs(tmb_1, re = 'Subject')

m-clark/mixedup documentation built on Oct. 15, 2022, 8:55 a.m.