Tidy Estimation of Heterogeneous Treatment Effects.

tidyhte

tidyhte provides tidy semantics for estimation of heterogeneous treatment effects through the use of Kennedy’s (n.d.) doubly-robust learner.

The goal of tidyhte is to use a sort of “recipe” design. This should (hopefully) make it extremely easy to scale an analysis of HTE from the common single-outcome / single-moderator case to many outcomes and many moderators. The configuration of tidyhte should make it extremely easy to perform the same analysis across many outcomes and for a wide-array of moderators. It’s written to be fairly easy to extend to different models and to add additional diagnostics and ways to output information from a set of HTE estimates.

The best place to start for learning how to use tidyhte are the vignettes which runs through example analyses from start to finish: vignette("experimental_analysis") and vignette("observational_analysis"). There is also a writeup summarizing the method and implementation in vignette("methodological_details").

Installation

You will be able to install the released version of tidyhte from CRAN with:

install.packages("tidyhte")

But this does not yet exist. In the meantime, install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("ddimmery/tidyhte")

Setting up a configuration

To set up a simple configuration, it’s straightforward to use the Recipe API:

library(tidyhte)
library(dplyr)

basic_config() %>%
    add_propensity_score_model("SL.glmnet") %>%
    add_outcome_model("SL.glmnet") %>%
    add_moderator("Stratified", x1, x2) %>%
    add_moderator("KernelSmooth", x3) %>%
    add_vimp(sample_splitting = FALSE) -> hte_cfg

The basic_config includes a number of defaults: it starts off the SuperLearner ensembles for both treatment and outcome with linear models ("SL.glm")

Running an Analysis

data %>%
    attach_config(hte_cfg) %>%
    make_splits(userid, .num_splits = 12) %>%
    produce_plugin_estimates(
        outcome_variable,
        treatment_variable,
        covariate1, covariate2, covariate3, covariate4, covariate5, covariate6
    ) %>%
    construct_pseudo_outcomes(outcome_variable, treatment_variable) -> data

data %>%
    estimate_QoI(covariate1, covariate2) -> results

To get information on estimate CATEs for a moderator not included previously would just require rerunning the final line:

data %>%
    estimate_QoI(covariate3) -> results

Replicating this on a new outcome would be as simple as running the following, with no reconfiguration necessary.

data %>%
    attach_config(hte_cfg) %>%
    produce_plugin_estimates(
        second_outcome_variable,
        treatment_variable,
        covariate1, covariate2, covariate3, covariate4, covariate5, covariate6
    ) %>%
    construct_pseudo_outcomes(second_outcome_variable, treatment_variable) %>%
    estimate_QoI(covariate1, covariate2) -> results

This leads to the ability to easily chain together analyses across many outcomes in an easy way:

library("foreach")

data %>%
    attach_config(hte_cfg) %>%
    make_splits(userid, .num_splits = 12) -> data

foreach(outcome = list_of_outcomes, .combine = "bind_rows") %do% {
    data %>%
    produce_plugin_estimates(
        outcome,
        treatment_variable,
        covariate1, covariate2, covariate3, covariate4, covariate5, covariate6
    ) %>%
    construct_pseudo_outcomes(outcome, treatment_variable) %>%
    estimate_QoI(covariate1, covariate2) %>%
    mutate(outcome = rlang::as_string(outcome))
}

The function estimate_QoI returns results in a tibble format which makes it easy to manipulate or plot results.

Getting help

There are two main ways to get help:

GitHub Issues

If you have a problem, feel free to open an issue on GitHub. Please try to provide a minimal reproducible example. If that isn’t possible, explain as clearly and simply why that is, along with all of the relevant debugging steps you’ve already taken.

Discord

Support for the package will also be provided in the Experimentation Community Discord:

You are welcome to come in and get support for your usage in the tidyhte channel. Keep in mind that everyone is volunteering their time to help, so try to come prepared with the debugging steps you’ve already taken.