Is there a way to perform proper counterbalancing?


so in my studies i use a lot of counterbalancing. let me walk you through an example.

let’s say i have 32 conditions, and i run 32 participants.

as participants land at my study, i allocate them to conditions until all 32 conditions are exhausted.

now let’s say participant 5 decides they don’t want to do the study any more … so they “return that task” … now prolific allocates another person to my experiment … and a 33rd participant shows up. but what condition do i allocate them to? how do i know that the 33rd participant should be assigned to the 5th condition?

there are theoretical ways that i could figure this out. for example, my experiment might have a “heartbeat” mechanism that allows me to determine when someone has quit my study … but this requires time, (30 seconds? a minute?) to wait before deciding they’ve genuinely left, and not having computer troubles. in that time, prolific can send another participant before i know which person has left.

asking around the lab i work in, people usually do super dodgey stuff, like creating separate experiments for each condition, etc. or allocating people to conditions randomly, and then trying to even things out at the end with a second experiment targeting a particular condition.

i’ve done a lot of work with mturk, and one of the things i can do with mturk is assign a unique ID to each HIT. when participant 33 shows up at my experiment, there’s an ID attached to it which i can map to the earlier condition 5. this means i always end up with a neatly counterbalanced set of data.

is there a neat way to achieve this sort of counterbalancing with prolific? if not, can this be considered a feature request.

with thanks