Millisecond Forums

random selection from within sub-groups/categories

https://forums.millisecond.com/Topic15952.aspx

By GeeP - 4/9/2015

Hi,

I am trying to set up a Lexical Decision Task (LDT) such that 4 pictures/images (e.g., im1, im2, im3, im4) will be used as primes that need to be presented randomly with 4 categories of words (e.g., A, B, C, & D)  and matched nonwords (e.g., a, b, c, & d). The issue is that there are two sets of these words and nonwords with each set consisting of 32 words and 32 nonwords with each category in each set having 8 words and 8 nonwords. Now the idea is to counterbalance the sets between subjects such that some subjects get set 1 and others get set 2 and within each set, all 64 words and nonwords need to be randomly presented with the 4 images (64x4=256 trials). How do I accomplish this? I know that using the <list> function I can randomly present the 4 images with the 64 words and nonwords, but I am not sure how to randomly present subjects with one of two sets of words and nonwords.

Any help with resolving this would be greatly appreciated! Thank you very much in advance!
Best,
-G
By Dave - 4/9/2015

There are many ways to do this. The most basic one is to have separate elements (e.g. two sets of <picture>s, <list>s, <trial>s) representing your two sets, and assign subjects to those two between-subjects conditions in the standard way: Via two <expt> elements and their /subjects attributes.
By GeeP - 4/10/2015

Thanks Dave!

I will try this with the list function after creating 4 different sets of the words and see if it will work. Thanks again!
-G
By GeeP - 4/10/2015

Hi Dave,

For properly counterbalancing,I am still not getting this right.
So, in our LDT - we have 4 sets:

Positive list1 (P1), Positive list2 (P2), Negative list1 (N1) and Negative list2 (N2)

We have created a combination of of these sets as follows:
1. P1N1 always presented with images A and B. P2N2 always presented with images C and D. This will then be counterbalanced.
2. P2N2 with A and B. P1N2 with C and D. This is also counterbalanced
3. P1N2 with A and B. P2N1 with C and D. Again, this is counterbalanced
4. P2N2 with A and B. P1N1 with C and D. Again, this is counterbalanced. Therefore, we will have 8 different conditions.

Moreover, each subject will complete 4 blocks:
block 1: they are conditioned with 4 images A, B, C and D
block 2: they complete LDT with one of the above combination of the lists with these images.
block 3: they are conditioned with another different set of images a, b, c and d
block 4: they complete LDT again with another combination of the list from the above with these images (a, b, c and d).

I am really stuck on how to NOT repeat the same combination of LDT words between the two blocks! I seem to have a programming block! If there is anything like that! Would you please suggest a simple way of doing this, please? The only way that comes to my mind, right now, is to create several blocks with all these combinations and present them in different experiments but that will create too many. Since this is a within-subjects study, we need only 34 subjects and doing that will mean that we won't run all the experiments! Please help!
Thanks a ton!!!!
-G
By Dave - 4/10/2015

> I am really stuck on how to NOT repeat the same combination of LDT words between the two blocks!

I don't understand why. Suppose you have two LDT blocks, <block ldt_a> and <block ldt_b>. Why can't you do

/ blocks = [...; 2,4 = noreplace(ldt_a, ldt_b); ...]

?