---
title: "Morocco"
author: "Bastiaan Quast"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Morocco}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, echo = FALSE, message = FALSE}
knitr::opts_chunk$set(collapse = T, comment = "#>")
```

we use the data from the Initiative Nationale du Development Humaine (INDH) a development project in Morocco.
The data is included with the `rddtools` package under the name `indh`.

We start by loading the package and the dataset.

```{r, message=FALSE}
library(rddtools)
data("indh")
```

Now that we have loading the data we can briefly inspect the structure of the data

```{r}
str(indh)
```

The `indh` object is a `data.frame` containing 720 observations (representing individuals) of two variables:

- `choice_pg` 
- `poverty`

The variable of interest is `choice_pg`, which represent the decision to contibute to a public good or not.
The observations are individuals choosing to contribute or not, these individuals are clustered by the variable `poverty` which is the municiple structure at which funding was distributed as part of the INDH project.
The forcing variable is `poverty` which represents the number of households in a commune living below the poverty threshold.
As part of the INDH, commune with a proportion of household below the poverty threshhold greater than 30% were allowed to distribute the funding using a **Community Driven Development** scheme.
The cutoff point for our analysis is therefore `30`.

We can now transform the `data.frame` to a special `rdd_data` `data.frame` using the `rdd_data()` function.

```{r}
rdd_dat_indh <- rdd_data(y=choice_pg,
                         x=poverty,
                         data=indh,
                         cutpoint=30 )
```

The structure is similar but contains some additional information.

```{r}
str(rdd_dat_indh)
```

In order to best understand our data, we start with an exploratory data analysis using tables...

```{r}
summary(rdd_dat_indh)
```

...and plots.

```{r}
plot(rdd_dat_indh[1:715,])
```

We can now continue with a standard Regression Discontinuity Design (RDD) estimation.

```{r}
(reg_para <- rdd_reg_lm(rdd_dat_indh, order=4))
```

In addition to the parametric estimation, we can also perform a non-parametric estimation.

```{r}
bw_ik <- rdd_bw_ik(rdd_dat_indh)
(reg_nonpara <- rdd_reg_np(rdd_object=rdd_dat_indh, bw=bw_ik))
```

Sensitity tests.

```{r}
plotSensi(reg_nonpara, from=0.05, to=1, by=0.1)
```