```{r setup, include=FALSE} knitr::opts_chunk$set(echo = TRUE) ``` ## Prerequisite Please read the `voted_na.csv` dataset and assign it as `data` object. ```{r} data <- read.csv("data/voted_na.csv") ``` ## Check every step you do; don’t run a big block of codes at once Let's say, I create a new variable called `HispanOld` that takes the value of `TRUE` if a respondent is **Hispanic** (*racename*) and has an **age** equal or above of 60 years. And I want to know if there are differential **gender** (*female*) effects on **voting** across respondents who are `HispanOld` or not: ```{r} data$HispanOld <- ifelse(data$racename == "Hispanic" & data$age >= 60, TRUE, FALSE) hoT.data <- subset(data, HispanOld == 1 & female == 1) hoC.data <- subset(data, HispanOld == 1 & female == 0) nhoT.data <- subset(data, HispanOld == 0 & female == 1) nhoC.data <- subset(data, HispanOld == 0 & female == 0) ho_diff <- mean(hoT.data$voted, na.rm=T) - mean(hoC.data$voted, na.rm=T) nho_diff <- mean(nhoT.data$voted, na.rm=T) - mean(nhoC.data$voted, na.rm=T) print(ho_diff) print(nho_diff) ``` Why is there an error? Why is there a `NA`? It is crucial to isolate at which exact step that the error or unexpected result first appears. ## Re-estimate the differences in means (`ho_diff` and `nho_diff`) using `tapply()`. Recall that `tapply()` requires at least three arguments: a **numeric** vector, a vector containing the **index or groups**, and the **function** (`FUN`, e.g. `sum`) to apply in each group. *Hint:* in some argument, you will need to use the function `list()`. ```{r} ## Calculate the average first difference by treatment group and by whiteConserv status first_diffs <- tapply(data$voted, list(data$female, data$HispanOld), mean, na.rm = TRUE) first_diffs[2,] - first_diffs[1,] ``` ## Knit to `PDF` or ask for help in Slack. If you successfully compute all the expected quantities, **knit** the document into a `PDF`. If you encounter any issues, create a **minimal reproducible example** by copying the minimum necessary code to reproduce the error or the part where you need help into an `R script` file. Use the `write.csv()` function to save the data file, ensuring to include the correct extension `.csv`. Once you have either the `PDF` or the **minimal reproducible example**, send me those files via **Slack** in a **private channel** (to me, Ramses). Alongside these files, include either a PDF with the solutions or an R script file + the saved data required for replication. Additionally, mention in the Slack message which **section** you are in (AC or AD). 