Selecting the option chosen in a discrete-choice experiment in R-CodePudding

I have the following dataset, from a choice experiment:

   ID task opt q1 q2 q3 q4 q5 q6 q7 q8
1   1    1   1  2  3  1  2  2  1  2  1
2   1    1   2  2  3  1  2  2  1  2  1
3   1    1   3  2  3  1  2  2  1  2  1
4   1    2   1  2  3  1  2  2  1  2  1
5   1    2   2  2  3  1  2  2  1  2  1
6   1    2   3  2  3  1  2  2  1  2  1
7   1    3   1  2  3  1  2  2  1  2  1
8   1    3   2  2  3  1  2  2  1  2  1
9   1    3   3  2  3  1  2  2  1  2  1
10  1    4   1  2  3  1  2  2  1  2  1
11  1    4   2  2  3  1  2  2  1  2  1
12  1    4   3  2  3  1  2  2  1  2  1
13  1    5   1  2  3  1  2  2  1  2  1
14  1    5   2  2  3  1  2  2  1  2  1
15  1    5   3  2  3  1  2  2  1  2  1
16  1    6   1  2  3  1  2  2  1  2  1
17  1    6   2  2  3  1  2  2  1  2  1
18  1    6   3  2  3  1  2  2  1  2  1
19  1    7   1  2  3  1  2  2  1  2  1
20  1    7   2  2  3  1  2  2  1  2  1
21  1    7   3  2  3  1  2  2  1  2  1
22  1    8   1  2  3  1  2  2  1  2  1
23  1    8   2  2  3  1  2  2  1  2  1
24  1    8   3  2  3  1  2  2  1  2  1

Meaning individual 1 (ID) was asked 8 questions (task), each of which had 3 options (opt) she could choose from. The remaining columns (q1...q8) reflect the option she chose in each of the choice tasks. So, she chose option 2 on question 1 and, again, option 2 on question 7.

I'm looking to turn it into a dataset that looks like this instead:

   ID task opt chosen
1   1    1   1  FALSE
2   1    1   2  TRUE 
3   1    1   3  FALSE
4   1    2   1  FALSE
5   1    2   2  FALSE
6   1    2   3  TRUE 
7   1    3   1  TRUE 
8   1    3   2  FALSE
9   1    3   3  FALSE
...
22  1    8   1  TRUE
23  1    8   2  FALSE
24  1    8   3  FALSE

In other words, a new column (chosen) reflects TRUE if that particular option was chosen in that particular question, and FALSE otherwise.

I've tried all forms of melt and reshape and cast but I've come up empty.

Thank you!

CodePudding user response：

You can first pivot_longer columns q1 to q8, then filter for "q" number that matches the task column. Then set the chosen column to TRUE if the value column matches the opt column. Finally removes the newly created columns.

Update: Credit to @Sotos for improving the answer. Now we set the logical column chosen directly from the comparison of value to opt without an if_else.

library(dplyr)
library(tidyr)

df %>% pivot_longer(q1:q8, names_to = "Q") %>% 
  filter(task == gsub("q", "", Q)) %>% 
  mutate(chosen = value == opt) %>%
  select(-Q, -value)

   ID task opt chosen
1   1    1   1  FALSE
2   1    1   2   TRUE
3   1    1   3  FALSE
4   1    2   1  FALSE
5   1    2   2  FALSE
6   1    2   3   TRUE
7   1    3   1   TRUE
8   1    3   2  FALSE
9   1    3   3  FALSE
10  1    4   1  FALSE
11  1    4   2   TRUE
12  1    4   3  FALSE
13  1    5   1  FALSE
14  1    5   2   TRUE
15  1    5   3  FALSE
16  1    6   1   TRUE
17  1    6   2  FALSE
18  1    6   3  FALSE
19  1    7   1  FALSE
20  1    7   2   TRUE
21  1    7   3  FALSE
22  1    8   1   TRUE
23  1    8   2  FALSE
24  1    8   3  FALSE

CodePudding user response：

Here is a dplyr solution with rowwise() and c_across(), where you don't even need to transform data to long form.

library(dplyr)

df %>%
  rowwise() %>%
  mutate(chosen = c_across(q1:q8)[task] == opt) %>%
  ungroup() %>%
  select(!q1:q8)

# A tibble: 24 × 4
      ID  task   opt chosen
   <int> <int> <int> <lgl> 
 1     1     1     1 FALSE 
 2     1     1     2 TRUE  
 3     1     1     3 FALSE 
 4     1     2     1 FALSE 
 5     1     2     2 FALSE 
 6     1     2     3 TRUE  
 7     1     3     1 TRUE  
 8     1     3     2 FALSE 
 9     1     3     3 FALSE 
10     1     4     1 FALSE 
11     1     4     2 TRUE  
12     1     4     3 FALSE 
13     1     5     1 FALSE 
14     1     5     2 TRUE  
15     1     5     3 FALSE 
16     1     6     1 TRUE  
17     1     6     2 FALSE 
18     1     6     3 FALSE 
19     1     7     1 FALSE 
20     1     7     2 TRUE  
21     1     7     3 FALSE 
22     1     8     1 TRUE  
23     1     8     2 FALSE 
24     1     8     3 FALSE