|
|
hi,
in my experiment i have 2 types of trials, A-trials and B-trials. both trial types should all be randomized within each other (i.e they are all treated as one group for wgr). BUT i also need to make sure that B type trials do not follow each other. if a B trial was presented, the next trial must be of type A. is it possible to implement this?
thanks!
So it should always be ABABABABABA... etc? If so, just assign all the A's a WGR of 1 and all the B's a WGR of 2. Yes?
the proportion is 4 A-type trials to 1 B-type trial. so it can be AAAAAAABABA etc.
I've been working on a nice simple notation system for creating constraints like that, but I haven't programmed it. Might be working by the next version. The only way I can think of to do this in v2006 would be to treat the BA's as BGR pairs and draw the stimuli randomly from stimlists as follows:
BGR WGR STIM
1 0 $A
2 0 $A
3 0 $A
4 0 $A
5 0 $A
6 0 $A
7 0 $B
7 0 $A
8 0 $B
8 0 $A
because i'm using compound screens, drawing the stimuli from a stimlist wouldn't work in my case.
after some thinking & consulting, i implemented the following solution:
since there are exactly 4 different A-type stimuli and only one B-type stimulus, i created 4 compound B trials: BA1, BA2, BA3, BA4. then i randomized these trials with the regular A trials:
WGR stim
1 A1
1 A2
1 A3
1 A4
1 BA1
1 A1
1 A2
1 A3
1 A4
1 BA2
and so on.
so, i've ensured that after each B trial there would be an A trial, but the stimulus for the particular A trial is randomized.
i hope this is clear.
michal.
So you stuck two trials together into one? i.e., whenever you have a B trial, you automatically get an A trial because it's actually stuck onto the end of that B trial?
that's exactly what i did. what do you think?
Clever! :)
