如何使用循环将分类数据更改为二进制数据？

Question

I have a data set that compares improvement in student's performance to their studying method. The categorical parameter is the studying method and the three options are NO (no assistance), FULL (yes assistance), and CHECK (assistance as needed). To set up my regression, I need to set x1=1 if a student had FULL assistance else x1=0. I also need to set an x2=1 if a student was labeled as CHECK, else x2=0. I will also need to do the same for NO help (x3=1 if ASSIST='NO' else x3=0). Here's what I'm trying to do:

if (ACCHW$ASSIST<-"FULL") { x1=1
} else {
x1=0
}

But I get the ERROR: Error in if (ACCHW$ASSIST <- "FULL") {: argument is not interpretable as logical.

I'm new to creating loops in R, advice is greatly appreciated!

Answer 1

It would be enough to create a new data frame comparing ASSIST to each of it's values, and cbind that to the origiginal data frame. Using the following data:

ACCHW <- data.frame(ASSIST = c("NO", "FULL", "CHECK"), stringsAsFactors = F)

Try this:

cbind(ACCHW,
      data.frame(x1 = ACCHW$ASSIST == "NO",
                 x2 = ACCHW$ASSIST == "FULL",
                 x3 = ACCHW$ASSIST == "CHECK"
                 )
      )

#### OUTPUT ####

  ASSIST    x1    x2    x3
1     NO  TRUE FALSE FALSE
2   FULL FALSE  TRUE FALSE
3  CHECK FALSE FALSE  TRUE

Remember that TRUE == 1 and FALSE == 0 . If you really want 0s and 1s do the same as above, but with as.integer :

cbind(ACCHW,
      data.frame(x1 = as.integer(ACCHW$ASSIST == "NO"),
                 x2 = as.integer(ACCHW$ASSIST == "FULL"),
                 x3 = as.integer(ACCHW$ASSIST == "CHECK")
                 )
      )

Answer 2

Here is a dplyr alternative!

library(dplyr)

df <- tibble(ASSIST = c("NO", "FULL", "CHECK"))

df %>% 
  mutate(x1 = ASSIST == "NO",
         x2 = ASSIST == "FULL",
         x3 = ASSIST == "CHECK") %>% 
  mutate_at(c("x1", "x2", "x3"), as.integer) # Converts to binary

Answer 3

Base R Solution:

cbind(ASSIST = ACCHW$ASSIST, data.frame(lapply(data.frame(setNames(
  sapply(ACCHW$ASSIST, `==`, ACCHW$ASSIST),
         c(names(ACCHW), ACCHW$ASSIST))), as.integer)))

Data thanks (@ gersht):

ACCHW <- data.frame(ASSIST = c("NO", "FULL", "CHECK"), stringsAsFactors = FALSE)