Dot probe - Randomizing question (case 1768)
Hello,
I am creating a dot probe task in DirectRT, and would like several of my fields to be randomized (location for stimuli pairs, time for display of stim, and location of subsequent dot probe). For example, I want 30 trials to be presented with each time duration (250ms, 500ms, and 1250ms), but I want these durations to be random. The user manual has been very helpful in brainstorming several ways to accomplish what I want, since I want several fields to be random in this way.
I also checked out this thread: [URL]http://www.empirisoft.com/support/showthread.php?t=1632[/URL] as I'm thinking about using the rand(250,500,1250) function for each stim pair. I'm wondering if there is any way to specify or guarantee that I'd like an equal number of trials to be 250ms, and 500ms, and 1250ms? Or, using this rand() function, will it just be "about equal" but no way to guarantee that 30 trials of each will be presented?
Hope this makes sense. Thank you!
Using WGR for Dot Probe Task
Hopefully this will help some others...
I believe there is a simple (if not tedious) way to GUARANTEE you will have a certain number of trials where a stimulus appears on the Left or Right as well as if the trial is 250, 500, etc. milliseconds.
The only way I know to do this involves using the same WGR value for each trial. Basically, you need to create the number of trials you will need along each dimension (e.g. 30 trials at 250 ms, 30 trials at 500 ms). Then, make half of those trials have a dot appear on the Left, and half have the dot appear on the Right. By having each possibility represented as a permanent trial in your input file, you will know it will be displayed. Moreover, randomization using the WGR column will ensure the presentation is random (and thus order in the input file does not matter).
Additionally, in order to BOTH randomize whether a stimulus appears on the Left or the Right side of the screen, AND to ensure that you will know if the participant responded correctly or not (using the Time column), you will need to code each trial ahead of time and WGR randomize the trials.