Causal Contrast Marginal Effects Estimation — causal_contrast

This function estimates the average treatment effect (ATE) or average treatment effect on the treated (ATT) using generalized linear models (GLMs). It supports handling of continuous and categorical treatments, optional use of spline transformations, and adjustments for multiple imputation datasets.

Usage

causal_contrast_marginal(
  df,
  Y,
  X,
  baseline_vars = "1",
  treat_0,
  treat_1,
  estimand = c("ATE", "ATT"),
  type = c("RR", "RD"),
  nsims = 200,
  cores = parallel::detectCores(),
  family = "gaussian",
  weights = NULL,
  continuous_X = FALSE,
  splines = FALSE,
  vcov = "HC2",
  verbose = FALSE
)

Arguments

df: Data frame containing the data.
Y: The response variable in the data frame.
X: The treatment or exposure variable in the data frame.
baseline_vars: A vector of names of baseline covariates to adjust for in the model.
treat_0: The reference level of the treatment variable, corresponding to no treatment or control condition.
treat_1: The active level of the treatment variable, corresponding to receiving the treatment.
estimand: A character vector specifying the estimand; "ATE" for Average Treatment Effect or "ATT" for Average Treatment Effect on the Treated.
type: A character vector specifying the type of effect size; "RD" for Risk Difference or "RR" for Risk Ratio.
nsims: Number of simulations to perform, relevant when handling multiple imputation datasets.
cores: Number of cores to use for parallel processing.
family: The family of the GLM to be used (e.g., "gaussian" for linear models).
weights: The name of the weights variable in the data frame, or NULL if no weights are to be used.
continuous_X: Logical indicating whether the treatment variable X is continuous.
splines: Logical indicating whether to use spline transformations for the treatment variable X.
vcov: The method to use for variance-covariance matrix estimation.
verbose: Logical indicating whether to display detailed output during model fitting.

Value

Depending on the 'type' specified, it returns a summary object containing either risk differences or risk ratios along with additional statistics like confidence intervals.

Examples

# Assume that df is your dataset with variables 'outcome', 'treatment', 'age', and 'gender'
result <- causal_contrast_marginal(df = df, Y = "outcome", X = "treatment",
                                   baseline_vars = c("age", "gender"),
                                   treat_0 = "control", treat_1 = "exposed",
                                   estimand = "ATE", type = "RD", nsims = 100,
                                   cores = 2, family = "gaussian", weights = "weight_var",
                                   continuous_X = FALSE, splines = FALSE,
                                   vcov = "HC3", verbose = TRUE)
#> Error in model.frame.default(formula = as.formula(formula_str), data = imputed_data,     weights = weights, drop.unused.levels = TRUE): 'data' must be a data.frame, environment, or list