Generalized Linear Regression with Covariates — regress_with

This unitility function performs a generalized linear regression on a specified dataset using an outcome variable, an exposure variable, and a set of baseline covariates. By default, the function uses the Gaussian family (linear regression), but it allows for specifying other families for generalized linear models (GLM).

Usage

regress_with_covariates(
  data,
  outcome,
  exposure,
  baseline_vars,
  family = gaussian(),
  sample_weights = NULL
)

Arguments

data: A data frame containing the variables for the analysis.
outcome: A character string specifying the name of the outcome variable in the data frame.
exposure: A character string specifying the name of the main exposure variable in the data frame.
baseline_vars: A character vector specifying the names of baseline covariates to include in the model in addition to the exposure variable.
family: A family object or a character string naming the family (default is gaussian(), which performs linear regression). This parameter determines the error distribution and link function to be used in the model.

Value

An object of class glm representing the fitted model, which includes coefficients, residuals, and other model diagnostics. This object can be further analyzed using standard methods for GLM objects, such as summary() for model summaries or anova() for analysis of variance.

Details

The function constructs a model formula using the outcome, exposure, and baseline variables. It then fits a generalized linear model using this formula. The baseline variables are filtered to exclude the outcome and exposure variables before model fitting. The function prints the formula used for the regression analysis for verification.

Examples

if (FALSE) { # \dontrun{
# using `df_nz` is your data frame with "income" as the continuous outcome variable,
# "age" as an exposure variable, and other covariates
outcome_var <- "income"
exposure_var <- "age"
baseline_vars <- c("age", "education", "partner")
model <- regress_with_covariates(df_nz, outcome_var, exposure_var, baseline_vars, family =  gaussian())
summary(model)
} # }