Pseudorandomization without repeating trials of a category within a block

[kdroon]

I have 6 blocks that contain 29 trials each. Block order is randomized (from a list of 6 rows in a file called block_order.xlsx) by a loop (works fine). The trial order for each block (block1.xlsx, etc.) is randomized (works fine). Each trial belongs to a category, indicated in a ‘category’ column. Every category has at least two trials in each block. The problem I have is that I want to constrain the randomization so that two consecutive trials must not be of the same category. I realize this requires custom code to implement. I have tried this and failed. I think I have two specific problems:

1. Most of the examples I have seen (example below) provide code that goes in “Begin Exp.” My problem is that I want this constraint applied within each block, not across the entire experiment. “Begin Exp.” does not seem to me to be the right place to put this code.

I assume that means that the code should run at the beginning of a routine that is at the beginning of my trial loop, and that it should run when thisN == 0.

Randomisation without consecutive presentations

@Mirjam provides a general solution here: In your specific case, something like this might work: Insert a code component on your trial routine (from the “Custom” component panel). In its “begin experiment” tab, put something like this ( we need to run this at the start of the experiment since the code is non-deterministic: we don’t know how long it will take to run (realistically though, it will be a fraction of a second for only 12 entries): stimulus_list = ['v', 'b', 'n', 'm'] * 3 shuffl…

2. I was able to change the example code above so that I could get the trialList for a block and loop through it to check for duplicate categories. What I have not been able to figure out is how to update trialList to a reshuffled order (e.g., the code in the example topic doesn’t update anything other than the array it made for demonstration purposes. I have tried setting trials.trialList to the reshuffled list, but it doesn’t work.

[wakecarter]

You can set use rows for a loop to a custom ordered list and set the loop type to sequential, but you would need to do this before the loop starts, so use concentric loops with the outer loop for the blocks.

There are some examples that might be helpful here:

PsychoPy Online Demos

If you have routines or experiments that are likely to be useful to others, please create a post in this thread, editing your post to add new demos so there is one post per creator. Ideally set your demos so they can be tested without consuming a credit (if you don’t have a site license) by adding a final routine that can’t end, disabling save incomplete results and adding a small number of credits. Please do not discuss the demos in this thread or create more than one post unless you have a la…

Complex randomisation and randomisation without repetition show two different approaches.

[kdroon]

Thanks. I have thought of that, but then the order of stimuli for each participant will be the same. I would ideally like the within-block trial order to be different–but within my constraints–for each participant.

[wakecarter]

I think you’ve misunderstood me. A sequential loop is still random if you randomise the list of indices.

[kdroon]

I’m afraid I do not understand. I also looked at your code examples (including complex-randomisation.psyexp). The problem is that I do not really know Python well enough to follow what you’re doing. If this question has a simple answer, then maybe I can figure it out: is it possible in a routine at the start of a trial loop to shuffle and replace the trialList?

I have tried

currentTrialList = trials.trialList (then do a bunch of checking, which seems to work fine)

shuffle(currentTrialList)

trials.trialList = currentTrialList

This doesn’t work.

I really appreciate your taking the time to try to help, but I think I’m going to have to try a different tack. I am sure this is largely due to my lack of Python knowledge, which I do not have time to address before running this experiment.

[JensBoelte]

Hello @kdroon

Do you mind posting your condition-file? I would give it a try.

Best wishes Jens

[kdroon]

Thank you for your offer to help, Jens. I am going to address this issue another way.

To anyone else who comes across this, please don’t worry about it. I would delete the topic, but I can’t.