generating random locations

Francisco Tornay ftornay at ugr.es
Wed Oct 20 11:09:56 UTC 2004


At 20:16 19/10/04, you wrote:
>HI,
>We are currently trying to program a visual search paradigm (Find a
>rotated T among rotated Ls). Our problem is this particular version of the
>task requires you to repeat some of the configurations while generating
>new random configurations on all other occations. There are 720 trials, and
>360 of these are entirely new, the others are the repetitions of 12
>previously generated displays. WE want to look at whether people do learn the
>"repeated configurations".
>
>The question is how can we generate random configurations in 360 trials
>(for each trial we need to choose 12 locations out of a possible 48
>locations where the target and the distractors appear and draw rotated Ts
>and Ls on these locations). If you have any suggestions, please let us
>know.

Hi, Aysecan. What you want to do is aleatorize a matrix of locations. The 
details, of course, depend on the way you present the stimuli, i. e. 
whether each location is a separate object in a slide or you paint the 
stimuli in different parts of a canvas object or you use some other system.

I have tried to write a couple of code lines for you but there are too many 
possibilities. Could you please elaborate on your procedure? I would need 
to know:

a)  How do you represent locations?: as integer indices (1 to 48), as 
different objects in a slide, as different display objects that get 
overlapped? If you use objects, how do you name them? If you haven't 
decided this yet, I can send you an experiment of mine when I do something 
similar by painting in a canvas object.
b) How do you choose the to-be-repeated frames? Is it a fixed set at the 
beginning of the experiment or do they have to meet some constraints, as to 
have been presented two or three or whatever trials ago? Or maybe you don't 
care which they are as long as they get repeated? Anyway, how many 
different frames must be repeated and how many times each?

Cheers,

Francisco J. Tornay
Universidad de Granada
España



More information about the Eprime mailing list