Random Without Replacement

Hi Guys!

I have two experiments I am programming. Both contain a set of five practice frames prior to the actual presentation. I’ve used expressions to program a random 5 from the same list used in the main presentation. However, I would not like those 5 that were randomly presented in the practice trial to present themselves in the main trial. They do use the same lists and the main presentaion is 25 from the same list using expressions as well. Is this possible?

Thankyou so much for your continued help. It is so appreciated.


Separate lists

For what it’s worth, I always set up a separate practice list that is used for all subjects. I don’t usually randomize it, instead, I try to avoid repetitions and present the the practice trials in a fixed order. In Superlab, this could be done in a separate block, to be presented before the real block(s). This will insure that all subjects have the same amount and type of practice before real trial 1.

Your approach of randomly selecting several trials from the real trial list for each subject will also mean that each subject will receive a different set of real trials, which would complicate item-level analysis.

Greg Shenaut

I think the easiest way to prevent trials from being re-presented would be to use code values on the trials to mark which have already been presented.

First, you create a code value, “presented,” with values, “yes” and “no,” to be used on trials. Then, you set the value on your trial(s) to “no.” Next, you create a feedback that runs somewhere in the trial every time it’s presented that changes this code value to “yes.” Lastly, you set up an expression to skip the next trial if its code value is “yes.” Since this is dependent on traits specific to the next trial, this can only be done at the trial level–that is, in the trial editor.

I don’t remember if it was 4.0.6 or 4.0.7 that added this functionality, but it’s relatively new to SuperLab 4.

Thankyou! I’ll be trying that shortly and I’m sure it will work smoothly.
Some faces in my experiment are meant to be attractive while others are unattractive. It is important I know, without the participant knowing, which faces (attractive or unattractive) appeared and in what order. Do I use codes in this instance as well? Or descriptions? It would be helpful if it would appear on the results print out or somewhere I could easily compare results with the order of faces that appeared.
What do you suggest?

Thankyou again for your always helpful and prompt replies!!!