removing rows based on conditions in R - r

Question

For my research, I have to remove certain trials to limit contamination in the data. Here are the rules:
Remove the first trial.
For RT calculation, remove error trials (trials with 0, trials with 1 are corrects trials) and trials following an error. Say, we have 20 trials and 3 errors, we have to remove 6 trials from the final data, i.e., mean RT will be calculated from 14 trials. If the errors are in a row, say 3 errors in a row, RT will be calculated based on 17 trials.
For error rate calculation, remove trials following an error. Say, we have 20 trials and 3 errors, we have to remove 3 trials following errors from the final data, i.e., error rate will be calculated by 3/17. If the errors are in a row, say 3 errors in a row, only the first error retains and the next two errors are excluded, so the error rate will be calculated by 1/18.
I am new to R. I hope someone can help me with the script. Thanx in advance
trial_no correct
1 1
2 1
3 1
4 1
5 1
6 1
7 1
8 1
9 1
10 1
11 1
14 0
15 1
16 1
17 1
18 1
19 0
20 1
21 1
22 1
23 0
24 0
25 1
26 1
27 1
28 1
29 1
30 1
31 1
32 0
33 0
34 1
35 1
36 1
37 1
38 1
39 0
40 1

Answer 1

If I understand your correctly the two rules work as follows:
A trial is excluded from RT calculation if this trial or the preceding trial has correct==0.
A trial is excluded from RT calculation if this trial has correct==0 and the preceding trial has correct==0.
If this is how you want things done, consider the two methods below.
Reproducing your example data:
df <- data.frame(trial_no = 1:40,
correct = rep(1, 40))
df$correct[c(13, 19, 23, 24, 32, 33, 39)] <- 0
Solution using base R only:
# usage for RT
df$use_rt <- df$correct==1
df$use_rt[2:nrow(df)][df$correct[-nrow(df)]==0] <- FALSE
# usage for error rate
df$use_er <- TRUE
df$use_er[2:nrow(df)] <- df$correct[-nrow(df)]==1 | df$correct[-1]==1
# calculate error rate
1 - mean(df$correct[df$use_er])
Solution using dplyr:
library(dplyr)
# usage for RT and error rate
df <- df %>%
mutate(use_rt = correct==1 & lag(correct)==1,
use_er = correct==1 | lag(correct)==1)
df$use_rt[1] <- df$correct[1]==1 # (without this you will have an NA-value in the first row)
# calculate error rate
df %>%
filter(use_er) %>%
summarize(error_rate = 1 - mean(correct))
The error rate is 0.1315789 (5 / 38).
For calculating reaction times you can now filter by use_rt, i.e. df[df$use_rt,].
df resulting from either method:
trial_no correct use_rt use_er
1 1 TRUE TRUE
2 1 TRUE TRUE
3 1 TRUE TRUE
4 1 TRUE TRUE
5 1 TRUE TRUE
6 1 TRUE TRUE
7 1 TRUE TRUE
8 1 TRUE TRUE
9 1 TRUE TRUE
10 1 TRUE TRUE
11 1 TRUE TRUE
12 1 TRUE TRUE
13 0 FALSE TRUE
14 1 FALSE TRUE
15 1 TRUE TRUE
16 1 TRUE TRUE
17 1 TRUE TRUE
18 1 TRUE TRUE
19 0 FALSE TRUE
20 1 FALSE TRUE
21 1 TRUE TRUE
22 1 TRUE TRUE
23 0 FALSE TRUE
24 0 FALSE FALSE
25 1 FALSE TRUE
26 1 TRUE TRUE
27 1 TRUE TRUE
28 1 TRUE TRUE
29 1 TRUE TRUE
30 1 TRUE TRUE
31 1 TRUE TRUE
32 0 FALSE TRUE
33 0 FALSE FALSE
34 1 FALSE TRUE
35 1 TRUE TRUE
36 1 TRUE TRUE
37 1 TRUE TRUE
38 1 TRUE TRUE
39 0 FALSE TRUE
40 1 FALSE TRUE