Compare two rows in one column R

If there are more than two elements per group, do a group by, and create a condition with ifelse/case_when to create the 'new_var'

library(dplyr)
df1 %>% 
  group_by(id) %>% 
  mutate(new_var = case_when(value == max(value) ~ "higher",
      TRUE ~ "lower")) %>% 
  ungroup
# A tibble: 8 × 3
  value    id new_var
  <int> <int> <chr>  
1     1     1 lower  
2     2     1 higher 
3     5     2 higher 
4     3     2 lower  
5    10     3 lower  
6    11     3 higher 
7     2     4 higher 
8     1     4 lower  

An option is also to arrange and then create a new column in the order with rep (assuming 2 elements per id)

df1 %>%
  arrange(id, value) %>% 
  mutate(new_var = rep(c("lower", "higher"), length.out = n()))

data

df1 <- structure(list(value = c(1L, 2L, 5L, 3L, 10L, 11L, 2L, 1L), id = c(1L, 
1L, 2L, 2L, 3L, 3L, 4L, 4L)), row.names = c(NA, -8L), class = "data.frame")

We could mutate a new_var with an ifelse statement conditional on min of value:

library(dplyr)
df %>% 
  group_by(id) %>% 
  mutate(new_var = ifelse(value == min(value), "lower", NA_character_)) %>%
  ungroup()

    value    id new_var
  <int> <int> <chr>  
1     1     1 lower  
2     2     1 NA     
3     5     2 NA     
4     3     2 lower  
5    10     3 lower  
6    11     3 NA     
7     2     4 NA     
8     1     4 lower