---
title: "Release history of causal.decomp"
author: "Suyeon Kang, Soojin Park"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Release history of causal.decomp}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE, eval = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r, message = FALSE, include = FALSE}
options(width = 200)
```

The following changes have been made since the initial release of `causal.decomp` 0.0.1.

## Changes in `causal.decomp` 0.1.0.

* The function `sensitivity` is added, which implements the sensitivity analysis for the causal decomposition analysis. As of version 0.1.0, the argument `boot.res` of `sensitivity` must be an object generated by `smi` with a single mediator. The object generated by `sensitivity` can be visualized in contour plots with robustness values using the `plot()` method.
* If non-NULL weights are used in fitting `fit.m` and `fit.y`, the weights are incorporated in the estimation by the `smi`, `mmi`, or `pocr` function.
* New data `sMIDUS` is added, which is synthetic data containing variables from actual Midlife Development in the U.S. (MIDUS) data used in Park et al. (2023). As the actual data is not publicly available due to confidentiality concerns, `sMIDUS` is not directly derived from the actual data but artificially generated to mimic the actual MIDUS data.

## References

* Park, S., Qin, X., & Lee, C. (2020). Estimation and sensitivity analysis for causal decomposition in health disparity research. Sociological Methods & Research, 00491241211067516.
* Park, S., Kang, S., & Lee, C. (2021+). Choosing an optimal method for causal decomposition analysis: A better practice for identifying contributing factors to health disparities. arXiv preprint arXiv:2109.06940.
* Park, S., Kang, S., Lee, C., & Ma, S. (2023). Sensitivity analysis for causal decomposition analysis: Assessing robustness toward omitted variable bias, *Journal of Causal Inference*. Forthcoming.