Creating a variable that illustrate change in another variable-CodePudding

I'm trying to create a variable that illustrate change in dicthomous variable. In other words, I have a dummy variable that illustrate whether something is active during a specific year. I need the new variable to account for change in the dummy variable while accounting for year and country. I have tried to illustrate the dummy I'm trying to create in the variable NewChangeVariable below.

read.table(
text =
"Country, Year, Dummy, NewChangeVariable,
US, 1, 1, 0,
US, 2, 0, 0,
US, 3, 1, 1,
US, 4, 1, 0,
UK, 1, 0, 0,
UK, 2, 1, 1, 
UK, 3, 1, 0, 
UK, 4, 1, 0, ", sep = ",", header = TRUE)

The NewChangeVariable is suppose to react to the first year a positive change in the dummy takes places. Again, dependent on country.

Please dont hesitate to ask if anything is unclear

CodePudding user response：

Here is a tidyverse approach. First, you can group_by Country. Then, for each Country, get the first row number where the prior Dummy was 0, but that row's Dummy is 1. Those rows meeting the criteria will be set to 1 (otherwise 0).

library(tidyverse)

df %>%
  group_by(Country) %>%
  mutate(NewChangeVariable2 =  (row_number() == which(lag(Dummy) == 0 & Dummy == 1)[1]))

Output

  Country  Year Dummy NewChangeVariable NewChangeVariable2
  <chr>   <int> <int>             <dbl>              <int>
1 US          1     1                 0                  0
2 US          2     0                 0                  0
3 US          3     1                 1                  1
4 US          4     1                 0                  0
5 UK          1     0                 0                  0
6 UK          2     1                 1                  1
7 UK          3     1                 0                  0
8 UK          4     1                 0                  0