RoBMA (version 4.0.0)

rstandard.brma: Internally Standardized Residuals for brma Objects

Description

Computes internally standardized residuals from a fitted brma object using the hat matrix. Returns a data frame with raw residuals, standard errors, and standardized residuals (z-values). Available for normal outcome models only.

Usage

# S3 method for brma
rstandard(model, unit = "estimate", conditioning_depth = "marginal", ...)

Value

A data frame with columns:

• resid: Raw residuals

• se: Standard errors of the residuals

• z: Standardized residuals

Arguments

model

a fitted brma object.

unit

output unit. Only "estimate" is implemented currently.

conditioning_depth

conditioning depth. Options are:

• "marginal" (default): Residuals from fixed effects predictions (\(observed - X\beta\)).

• "cluster": Residuals from cluster-level predictions (observed - (\(X\beta + gamma\))). Only available for multilevel (3-level) models.

• "estimate": Residuals from BLUPs, i.e., deviations of the observed effect sizes from the best linear unbiased predictions of the estimate-specific true effects (observed - theta).

...

additional arguments (currently ignored)

Details

This function returns a data frame with three columns matching the output of metafor::rstandard:

• resid: Raw residuals (observed - fitted values)

• se: Standard errors of the residuals

• z: Standardized residuals (resid / se)

Internally standardized residuals divide the observed residuals by their corresponding standard errors computed using the hat matrix. For a correctly specified model, these residuals should approximately follow a standard normal distribution.

This function is only available for normal outcome models without selection (weightfunction) bias adjustment. For other model types, use rstudent.brma which uses LOO-PIT.

See Also

rstudent.brma(), residuals.brma(), blup.brma(), predict.brma()

Examples

if (FALSE) {
if (requireNamespace("metadat", quietly = TRUE)) {
  data(dat.lehmann2018, package = "metadat")
  fit <- brma(yi = yi, vi = vi, data = dat.lehmann2018, measure = "SMD")

  # marginal internally standardized residuals (default)
  rstandard(fit)

  # estimate-level (BLUP-based) residuals
  rstandard(fit, conditioning_depth = "estimate")
}
}