|
GeeP
|
|
|
Group: Forum Members
Posts: 81,
Visits: 118
|
Hi,
I am designing a backward conditioning experiment in which 4 faces are presented (in a within subject design) wherein 2 faces (one male one female) are paired with a threat stimulus (e.g., snake) and the other two faces (one male one female) are paired with a neutral stimulus. Each face is presented for 20 trials with the total trials being 80 (4 faces X 20 trials). However, as a cover story, participants are told to ignore the faces or any images by respond only to the last image that will be either a 'star' or a 'triangle'. They are basically instructed to respond as soon as possible using one of two keys assigned if they saw a star or a triangle on each trial. Of the 20 trials associated with each face, 10 trials will be "star" trials and the remaining 10 will be "triangle" trials. I am just not sure how to set up all these various conditions using the stimulustimes attribute. For one condition I was able to set it up as follows: / stimulustimes = [0=Fixation; 2000=Threat; 2014=Mask; 2200=Blank; 3000=Female1; 6000=Triangle; 6014=Blank]
But I am not sure how to set it up for the following: Group 1:female1 (threat) = 20 trials (10=star, 10=triangle) male 1 (threat) = 20 trials (10=star, 10=triangle)female 2 (neutral) = 20 trials (10=star, 10=triangle)male 2 (neutral) = 20 trials (10=star, 10=triangle)
Group 2:female 2 (threat) = 20 trials (10=star, 10=triangle) male 2 (threat) = 20 trials (10=star, 10=triangle)female 1 (neutral) = 20 trials (10=star, 10=triangle)male 1 (neutral) = 20 trials (10=star, 10=triangle)
For each group of participants, the 80 trials must be randomly presented. How do I accomplish this? Sorry, I tried but failed! :/ Any help would be greatly appreciated! Thanks a ton in advance! - GPandey
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 109K
|
There are many ways to do this, but the most efficient approach (IMO) is using <list> elements to encode your factor combinations in the amounts needed. Then sample from those combinations. See e.g. http://www.millisecond.com/forums/FindPost13987.aspx for an example.
|
|
|
|
|
GeeP
|
|
|
Group: Forum Members
Posts: 81,
Visits: 118
|
Thanks for the quick reply Dave!
I looked at the example but ended up having a few other questions: As explained in my previous request, for half the participants, 2 faces cannot be paired with anything else except the threat image and 2 other faces need to be paired with only the neutral image. For the other half of the participants, this will change (vice versa), I am not sure if listing the combination will allow this restriction.
While waiting for a reply to my post, this is what I did:
<block test_A> / bgstim = (pleasant_Alabel, unpleasant_Alabel) / trials = [1-80 = noreplace(F1_Threat_Star, F1_Threat_Tria, M1_Threat_Star, M1_Threat_Tria, F2_Neutral_Star, F2_Neutral_Tria, M2_Neutral_Star, M2_Neutral_Tria,)] / recorddata = true </block>
I added the 8 combinations as separate trials for being presented an equal number of times until 80 trials are up. When I run the script, it runs fine but is this a bad way to do this? In other ways, since RT is my most important DV, am I hurting myself by doing this? Thanks again for your patience and all the help! - GeeP
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 109K
|
> As explained in my previous request, for half the participants, 2 faces cannot be paired with anything else except the threat image and 2 other faces need to be paired with only the neutral image. For the other half of the participants, this will change (vice versa), I am not sure if listing the combination will allow this restriction.
Perfectly possible to do using the approach I outlined. List the combinations applicable to group 1, list the combinations applicable to group 2, draw from the appropriate one.
> I added the 8 combinations as separate trials for being presented an equal number of times until 80 trials are up. When I run the script, it runs fine but i s this a bad way to do this?
It's not a "bad" way to do this. It's just a very verbose / somewhat inefficient way to do it, as you get lots of essentially redundant code (e.g. all <trial> elements are structurally identical; a single one would thus suffice in the solution I proposed, whereas you need 8).
|
|
|
|
|
GeeP
|
|
|
Group: Forum Members
Posts: 81,
Visits: 118
|
Thanks Dave! I understand the difference now. However, to have 2 faces always show up with the threat and the other two faces show up with neutral stimuli, I might need 2 lists for each group I believe, am I right? In your example, you state 3 stimuli (A, B, C) to be presented in 2 positions (left, right) at two different rates (slower, faster) with the combination 3X2X2 = 12 combinations, but in my case, I have 2 faces (male1, female1), always presented with a threat NEVER with a neutral stimulus but ending with 2 objects (star, triangle). For the other two faces (male2, female2) they are never combined with the threat but always with the neutral, so to accomplish this, I am thinking I'll need two lists for each group: Group 1: Threat: 2(faces) X 2(objects) Neutral: 2(faces) X 2(objects)
Group 2: same as above with faces swapped. Am I thinking this right?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 109K
|
I don't see why this would require separate lists for threat and neutral. Number your face stimuli 1-4 (1=m1; 2=f1; 3=m2; 4=f2) and simply never combine 1 & 2 with neutral, and 3 & 4 never with threat.
|
|
|
|
|
GeeP
|
|
|
Group: Forum Members
Posts: 81,
Visits: 118
|
Ah! OK, I see it now! Thanks again Dave! You are always so helpful!
|
|
|
|