I'm running a study in which each participant will be presented with stimuli that have been randomized at two different levels: blocks (3 unique blocks) and trials (4 unique trials per block) within blocks. So I am trying to create a data frame with a pre-randomized stimulus presentation list by:
- Randomly order the trials within each block
- Randomly order the blocks within participant.
--
df <- df %>%
group_by(id, block) %>%
mutate(trial = trial[sample(row_number())])
The trials randomize within block (and participant) but how do I then randomize the blocks within participant?
The data frame looks like this:
id |
block |
trial |
|---|---|---|
n1 |
b |
1 |
n1 |
b |
2 |
n1 |
b |
3 |
n1 |
b |
4 |
n1 |
p |
1 |
n1 |
p |
2 |
n1 |
p |
3 |
n1 |
p |
4 |
n1 |
s |
1 |
n1 |
s |
2 |
n1 |
s |
3 |
n1 |
s |
4 |
n2 |
b |
1 |
n2 |
b |
2 |
n2 |
b |
3 |
n2 |
b |
4 |
n2 |
p |
1 |
n2 |
p |
2 |
n2 |
p |
3 |
n2 |
p |
4 |
n2 |
s |
1 |
n2 |
s |
2 |
n2 |
s |
3 |
n2 |
s |
4 |
n3 |
b |
1 |
n3 |
b |
2 |
n3 |
b |
3 |
n3 |
b |
4 |
n3 |
p |
1 |
n3 |
p |
2 |
n3 |
p |
3 |
n3 |
p |
4 |
n3 |
s |
1 |
n3 |
s |
2 |
n3 |
s |
3 |
n3 |
s |
4 |
CodePudding user response：
How about something like this:
dat <- structure(list(id = c("n1", "n1", "n1", "n1", "n1", "n1", "n1",
"n1", "n1", "n1", "n1", "n1", "n2", "n2", "n2", "n2", "n2", "n2",
"n2", "n2", "n2", "n2", "n2", "n2", "n3", "n3", "n3", "n3", "n3",
"n3", "n3", "n3", "n3", "n3", "n3", "n3"), block = c("b", "b",
"b", "b", "p", "p", "p", "p", "s", "s", "s", "s", "b", "b", "b",
"b", "p", "p", "p", "p", "s", "s", "s", "s", "b", "b", "b", "b",
"p", "p", "p", "p", "s", "s", "s", "s"), trial = c(1L, 2L, 3L,
4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L,
4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L,
4L)), row.names = c(NA, 36L), class = "data.frame")
library(dplyr)
library(tidyr)
dat %>% group_by(id, block) %>%
summarise(trial = list(sample(trial, n(), replace=FALSE))) %>%
group_by(id) %>%
slice_sample(n=3) %>%
unnest(trial)
#> `summarise()` has grouped output by 'id'. You can override using the `.groups`
#> argument.
#> # A tibble: 36 × 3
#> # Groups: id [3]
#> id block trial
#> <chr> <chr> <int>
#> 1 n1 b 4
#> 2 n1 b 1
#> 3 n1 b 2
#> 4 n1 b 3
#> 5 n1 s 2
#> 6 n1 s 4
#> 7 n1 s 3
#> 8 n1 s 1
#> 9 n1 p 4
#> 10 n1 p 3
#> # … with 26 more rows
Created on 2022-05-13 by the reprex package (v2.0.1)