Dif-in-dif

Carregar as bibliotecas necessárias

library(dplyr)

## 
## Anexando pacote: 'dplyr'

## Os seguintes objetos são mascarados por 'package:stats':
## 
##     filter, lag

## Os seguintes objetos são mascarados por 'package:base':
## 
##     intersect, setdiff, setequal, union

library(broom)
library(ggplot2)

Criar um conjunto de dados simulado

set.seed(123)
data <- data.frame(
  city = rep(c("Treatment", "Control"), each = 100),
  year = rep(c(0, 1), times = 100),
  salary = c(rnorm(100, mean = 50, sd = 10), rnorm(100, mean = 55, sd = 10),
             rnorm(100, mean = 50, sd = 10), rnorm(100, mean = 50, sd = 10))
)

Ajustar o modelo Dif-in-Dif

model <- lm(salary ~ year * city, data = data)

Resumo do modelo

summary(model)

## 
## Call:
## lm(formula = salary ~ year * city, data = data)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -26.532  -6.443  -0.392   6.425  35.721 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        51.87305    0.97802  53.039   <2e-16 ***
## year               -0.18379    1.38314  -0.133    0.894    
## cityTreatment       0.08774    1.38314   0.063    0.949    
## year:cityTreatment -1.62907    1.95605  -0.833    0.405    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 9.78 on 396 degrees of freedom
## Multiple R-squared:  0.005744,   Adjusted R-squared:  -0.001788 
## F-statistic: 0.7626 on 3 and 396 DF,  p-value: 0.5156

Visualizar os resultados

data %>%
  group_by(city, year) %>%
  summarise(mean_salary = mean(salary)) %>%
  ggplot(aes(x = year, y = mean_salary, color = city, group = city)) +
  geom_line() +
  geom_point() +
  labs(title = "Diferenças em Diferenças",
       x = "Ano",
       y = "Salário Médio") +
  theme_minimal()

## `summarise()` has grouped output by 'city'. You can override using the
## `.groups` argument.