Responses and response times from an experiment in which three participants were asked to decide whether the overall brightness of pixel arrays displayed on a computer monitor was "high" or "low". In addition, instruction manipulated speed and accuracy between blocks.
data.frame with 24,358 obs. and 12 variables:
participant id, factor with three levels
session number, integer
block number, integer
trial number within a block, integer
factor with two levels:
"accuracy" for blocks with accuracy instructions;
"speed" for blocks with speed instruction
factor with two levels: distribution strength was drawn from,
proportion of white to black pixels were varied by 33 equally spaced proportions from zero (all 1,024 pixels were black) to 1 (all 1,024 pixels were white). with 0 darkest and 32 lightest. Integer.
factor with two levels:
numeric response variable such that
boolean indicating whether or not
source==response. (Does not seem to be used in the original analysis.)
response time in seconds
boolean indicating whether or not the response was considered an outlier by Ratcliff and Rouder (1998), i.e., RTs outside of (200ms, 2500ms)
The Experiment is described in the following by Ratcliff and Rouder (1998, pp. 349):
In Experiment 1, subjects were asked to decide whether the overall brightness of pixel arrays displayed on a computer monitor was "high" or "low". The brightness of a display was controlled by the proportion of the pixels that were white. For each trial, the proportion of white pixels was chosen from one of two distributions, a high distribution [i.e., light] or a low [i.e., dark] distribution, each with fixed mean and standard deviation. Feedback was given after each trial to tell the subject whether his or her decision had correctly indicated the distribution from which the stimulus had been chosen. Other than this feedback, a subject had no information about the distributions. Because the distributions overlapped substantially, a subject could not be highly accurate. A display with 50 from the high distribution on one trial and the low distribution on another.
The stimulus display for Experiment 1 was a square that was 64 pixels on each side and subtended 3.8 degree of visual angle on a PC-VGA monitor. [...] In each square, 3,072 randomly chosen pixels were neutral gray, like the background, and the remaining 1,024 pixels were either black or white; the proportion of white to black pixels provided the brightness manipulation. There were 33 equally spaced proportions from zero (all 1,024 pixels were black) to 1 (all 1,024 pixels were white). The two distributions from which the bright and dark stimuli were chosen were centered at .375 (low brightness) and .625 (high brightness), and they each had a standard deviation of .1875.
A subject's task was to decide, on each trial, from which distribution, high or low brightness in Experiment 1, the observed stimulus (stimuli) had been sampled. Subjects made their decision by pressing one of two response keys. On each trial, a 500-ms foreperiod, during which the display consisted solely of neutral gray, was followed by presentation of the stimulus; presentation was terminated by the subject's response. In Experiment 1, speed-versus-accuracy instructions were manipulated. For some blocks of trials, subjects were instructed to respond as quickly as possible, and a "too slow" message followed every response longer than 550 ms. For other blocks of trials, subjects were instructed to be as accurate as possible, and a "bad error" message followed incorrect responses to stimuli from the extreme ends of the distributions. Experiment 1 had ten 35-min sessions, and Experiments 2 and 3 had four sessions. In Experiment 1, subjects switched from emphasis on speed to emphasis on accuracy every 204 trials. Each session consisted of eight blocks of 102 trials per block, for a total of 8,160 trials per subject. Each session consisted of eight blocks of 102 trials, for a total of 3,264 trials per subject in each experiment. For all trials in each experiment, subjects were instructed to maintain a high level of accuracy while responding quickly, and an "error" message indicated incorrect responses. Responses were followed by a 300-ms blank interval, and the error message was displayed for 300 ms after the blank interval.
The data is already prepared following Ratcliff and Rouder (1998) by removing the following trials:
the first session for each participant
the first 20 trials of each session
the first trial of each block (each change in speed accuracy starts a new block)
To fully replicate the data used by Ratcliff and Rouder (1998) one only needs to remove the trials that are
TRUE in column
outlier (i.e., RTs outside of (200ms, 2500ms)). The full raw data is also available as part of this package, see:
system.file("extdata", "rr98-data", package = "rtdists") and
system.file("extdata", "rr98-data.codes", package = "rtdists")
Ratcliff, R., & Rouder, J. N. (1998). Modeling Response Times for Two-Choice Decisions. Psychological Science, 9(5), 347-356. http://doi.org/10.1111/1467-9280.00067
1 2 3 4 5 6 7 8 9 10 11 12
data(rr98) rr98 <- rr98[!rr98$outlier,] #remove outliers head(rr98) # id session block trial instruction source strength response response_num correct rt outlier # 1 jf 2 1 21 accuracy dark 8 dark 1 TRUE 0.801 FALSE # 2 jf 2 1 22 accuracy dark 7 dark 1 TRUE 0.680 FALSE # 3 jf 2 1 23 accuracy light 19 light 2 TRUE 0.694 FALSE # 4 jf 2 1 24 accuracy dark 21 light 2 FALSE 0.582 FALSE # 5 jf 2 1 25 accuracy light 19 dark 1 FALSE 0.925 FALSE # 6 jf 2 1 26 accuracy dark 10 dark 1 TRUE 0.605 FALSE ## See vignette for more examples.