I am in middle of a data collection and a participant contacted me with this error message:
Unfortunately we encountered the following error:
when getting the name of a group, using a counterbalanced design with key: ["counterbalance","@designer","@experiment"]
unable to access the record for the given key: either it does not exist, or it is locked, possibly by another experiment session
Try to run the experiment again. If the error persists, contact the experiment designer.
What is happening? What is settings this lock and how am I removing it?
It worked when I unlocked it twice it had two lock ids. Is there any documentation how they these can be locked? The variable keeps on getting locked again.
Are you using Prolific or similar? There’s a setting to limit the number of participants starting at the same time which might help. However, if it’s locking straight away if you test yourself then that’s not the issue.
I am running it on prolific but only very few slots are open. Now I paused it because it’s keep locking up.
Edit: I am unlocking, trying it out myself, which is immediately locking it.
Edit2: The counterbalancing also completely failed. The group assignment is 32:11.
I honestly cannot recommend using shelf variables and the counterbalancing routine. It’s beyond frustrating and I am currently not sure how I can finish my ongoing data collection.
I removed group1 from the counterbalancing variable (the counterbalancing is completely botched anyway). Now, I am just trying it out myself and it works once and then locks back.
Next I tried changing the slots
=
slot
10
*
nb reps
10
This seem to work for now. I wish I would have used @wakecarter’s VESPR tool like I did in the past or used something like JATOS. Both are much more robust handling group assignments and giving running participant numbers.
Did you previously have 1 per slot with 100 reps to make the 100 slots? I think most of my bad experiences with the counterbalance tool have been related to having a large number of reps of a comparatively small tranche (slots per rep). I also failed when I tried to remove two groups in the hope that future participants would be allocated to the other two and ended up going with random assignment – it was a student project with a short deadline but unequal numbers didn’t really matter)
If you want to use the VESPR Study Portal to complete your study then I’d be happy to insert some fake participants so that the initial participants get sent to group 2 until the groups are equal. It’s also possible to have an automatic launch rather than hosting the PI on the tool.
Yes, I think this might be what happened. I think I had 1 per slot and 100 reps.
I am currently just starting and stopping the study to make sure that it’s not overwhelmed. For the moment this seemed to work. Luckily, the counter balancing ‘failure’ is not as bad as I thought because the issues created by this also meant many of those participants didn’t finish anyway.
