outcome_model function

outcome_model

This function will run the outcomes model for your analysis. Upon completeion, use the model object returned from this function and call ov_simgrid to check the sensitivity of your findings.

outcome_model(ps_object = NULL, stop.method=NULL, data, weights=NULL, treatment,
outcome, model_covariates, estimand = "ATE")

Arguments

• ps_object: A ps object exported from TWANG
• stop.method: If the user specifies ps_object, stop.method should be used to export the weights (e.g "ks.max")
• data: A data frame containing the data
• weights: A column name in data that represents the relevant weights
• treatment: A column name in data for the treatment indicator
• outcome: A column name in data indicating the outcome vector
• model_covariates: A vector of column names representing the covariates in your final outcome's model
• estimand: "ATE" or "ATT"

Returns

outcome_model returns a list containing the following components: - ps_object: The ps_object from TWANG specified in the function call. If ignored, this component will be NULL

• stop.method: The stop method, if applicable, specified in the function call

• data: the updated data frame

• weights: the original vector of weights

• tx: a character name in data indicating the treatment indicator

• y: a character name in data indicating the outcome

• outcome_mod_fmla: the final outcome model formula

• estimand: The estimand specified in the function call

• mod_results: an object of class "svyglm"

Examples

data(sud)
sud = data.frame(sud)
sud$treat = ifelse(sud$treat == "A", 1, 0)
sud$wts = sample(seq(1, 10, by=.01), size=nrow(sud), replace = TRUE)
outcome_mod = outcome_model(data = sud,
                            weights = "wts",
                            treatment = "treat",
                            outcome = "eps7p_3",
                            model_covariates = c("sfs8p_0"),
                            estimand = "ATE")

References

Lumley T (2020). “survey: analysis of complex survey samples.” R package version 4.0.