Hello there! I had a query about updating the Pavlovia shelf to run additional participants, particularly only adding additional slots to some conditions, and not others.
I recently completed data collection for 200 participants via Prolific, 4 groups of 50, and this all worked great! I have some data quality checks I ran and 4 participants failed these (random responding), and I am planning to resample. I will need 1 extra in the first condition, 3 extra in the second, but none in the other two.
Is it just a case of updating the ‘target’ in the shelf to 51 and 53? Or do I need to also update the Counterbalance component in PsychoPy too? I wasn’t sure what to write in the number of slots for the latter, since it’s different per condition (it’s now essentially 51, 53, 50, 50).
While I understand your motives, be aware that this is actually a form of p-hacking. It is better to include checks during the experiment and stop it when certain quality criteria are not met. If this happens, do not save the data.
Thank you for answering the question; I’ll mark this as solved shortly
Jens, I see your view, absolutely; I will definitely keep in mind doing checks during the experiment in future. In my particular case it was quite hard to do so as there’s a big range of values that would be considered acceptable, but I can be smarter about this, I think!
Just for context, the four participants in question essentially spent most of the experiment messing around and typing in random predictions (e.g., 3043633273, 2832774634, 438643664) that were way larger than any feasible true answer (largest possible true answer was 60024) meaning their data was essentially unusable - especially as many key analyses are focused on accuracy!
The data quality check and resampling was included in my pre-registration for this experiment, but in light of your concerns I am going to hold off and contact my supervisors just to check how they want me to proceed.
