if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "data_set.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "data_set.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "data_set.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "data_set.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "dataset.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "dataset.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "dataset.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "dataset.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "dataset.dta"))
# Chunk 1: setup
knitr::opts_chunk$set(echo = TRUE)
# Chunk 2
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, here, haven, grf, corrplot, caret)
# Chunk 3
sink(here("results", "mylogfile.txt"))
suv_dta <- haven::read_dta(here("data", "data_set.dta"))
# Chunk 4
# Generate covariates vector
covariates <- c("soleproprietor", "local_sales", "small", "medium", "large", "lessthan10yrs", "man_lessthan10yrs", "manufactoring", "retail_whole", "sales_under100mil", "sales_100_500mil", "sales_over500mil", "services", "informal_comp", "auditor", "bank_account", "inspected", "gov_contract", "taxrates_obst", "taxadmin_obst", "java", "owner_manager", "female_manager", "foreign", "national_sales", "DGT_fair", "informal_payments")
suv_dta <- suv_dta %>%
filter(!is.na(IDa1))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa1),
W = as.numeric(as.character(suv_dta$treat1)),
W.hat = mean(as.numeric(as.character(suv_dta$treat1))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_1 <- variable_importance(train.forest)
rownames(important_1) <- names(suv_dta[,covariates])
print(important_1)
suv_dta <- suv_dta %>%
filter(!is.na(IDa2))
train.forest <- causal_forest(X = data.frame(suv_dta[,covariates]),
Y = as.numeric(suv_dta$IDa2),
W = as.numeric(as.character(suv_dta$treat2)),
W.hat = mean(as.numeric(as.character(suv_dta$treat2))),
min.node.size = 5,
seed = 12345,
num.threads = 1)
# Variable Importance
important_2 <- variable_importance(train.forest)
rownames(important_2) <- names(suv_dta[,covariates])
print(important_2)
stargazer::stargazer(important_1, type = "text", title = "Treatment 1", out = here("results", "important_1.doc"))
stargazer::stargazer(important_2, type = "text", title = "Treatment 2", out = here("results", "important_2.doc"))
sink()
