does somebody know how to set up a simple, standard experiment with one prime and one target in each trial? (whatever the modality: auditory-auditory, auditory-visual, etc.)
The issue is: how to set up a REAL expt. of, say, 200 trials from a text file (or an excel file or else) containing one line (row) per trial with a prime-target pair in each line.
I cannot find relevant info in the forums for ANY experiment with several stimuli per trial (e.g., AXB discrimination: one guy says there are several ways to do it: just give me one!)
Maybe SuperLab works fine, but it’s a mystery to me that no one (in more than 50 “countries”) has faced the simple problem of setting up an expt. from a list of trials in text format such as below:
rel paper pencil
filler book worm
unrel tree pencil
…
Or, for some AXB discrimination expt., something like:
is that in Superlab, you can’t deal directly with complex stimulus events, only simple ones, like individual text string events. However, if you have a script that sets things up like you said, “code prime target” in three columns, then you can make another simple script to split each column into a separate file of only one column. You then read each of the files into a separate text stimulus list, and set up “trials” that present: the next code for your intertrial interval in all black letters (assuming black is the background color); the next prime for the soa; and then target until the subject responds. Superlab will read through the three stimulus lists as the trials progress. The reason for the all black “event” is so you can read the trial code in the logfile.
Obviously there is a piece missing: how to randomize. This is a pretty big gap in superlab: you can’t randomize multi-trial chunks of trials. It would be great if there were another randomizable level “chunk” between what are currently called “trials” and “blocks”, but there isn’t. Probably the easiest way to handle this is to use a script to generate a manageable number of pre-randomized sequences and import the resulting stimulus lists into several copies of the experimental scenario; then randomly assign subjects to the scenarios.
Thanks Greg, your explanation is quite clear and helpful.
Well, yes, it’s a bit of a disappointment that multiple-event trials can’t be specified more easily. And a surprise: there are so many priming experiments, or discrimination (AX, ABX, AXB, etc.) that require multiple events. So, at least it seems possible to do it with several event lists. I hope the synchronization cannot get mixed up! The bad news is that trial randomization on the fly becomes impossible. I’ll have to go back to the old scheme: one random order prepared off-line and its mirror-order counterpart. Or I will try to use PsyScope for OS X. I remember the brave old PsyScope (for OS 7-9) could do that and more.
That’s something we might want Cedrus to consider. The present limitation to single-event trial lists is indeed hardly acceptable.
