---
title: "Session 16 — Diplomacy experiment (DIM)"
author: "Prof. Dae-Jin Lee (daelee@faculty.ie.edu)"
output: html_document
---

```{r setup, message=FALSE, warning=FALSE}
library(tidyverse)
exp <- read_csv("data/diplomacy_experiment.csv")
```

## Allocation and balance

```{r}
exp |> count(treatment_message)

exp |>
  group_by(treatment_message) |>
  summarise(
    mean_age = mean(age),
    mean_prior = mean(prior_support),
    n = n(),
    .groups = "drop"
  )
```

## Difference in means

```{r}
exp |>
  group_by(treatment_message) |>
  summarise(
    mean_y = mean(post_support_diplomatic_cooperation),
    n = n(),
    .groups = "drop"
  )
```

## OLS (= DIM for binary T)

```{r}
m <- lm(
  post_support_diplomatic_cooperation ~ treatment_message,
  data = exp
)
summary(m)
confint(m)
```

**Interpret:** What was randomized? How large is the estimated effect on the outcome scale? What does the CI tell you?