Skip to contents

Residualize trial-wise responses against nuisance terms

Source: R/grouped_data.R
residualize_trials.Rd

Fits linear models within optional groups and returns response residuals in a new column. This provides a general trial-level nuisance-regression step for behavioral analyses without hard-coding any specific covariates.

Usage

residualize_trials(
  data,
  response,
  terms = NULL,
  by = NULL,
  family = "gaussian",
  type = c("response", "working", "deviance", "pearson"),
  out = NULL
)

Arguments

data

data.frame containing response and nuisance columns.

response

character scalar naming the response column.

terms

optional character vector of nuisance column names. If empty or NULL, the response is copied unchanged.

by

optional character vector of grouping columns. Separate models are fit within each group.

family

model family. May be either a character scalar such as "gaussian", "binomial", or "inverse.gaussian", or a family object returned by family.

type

residual type passed to residuals.glm.

out

optional output column name. Defaults to paste0(response, "_resid").

Value

data.frame with an added residual column.