MindProbe: missing trial data and unexpectedly early study finish
Hi all,
I am trying to figure out once issue with a jsPsych experiment hosted on MindProbe, and I'd like to know if there's anything I'm missing here.
A participant reports completing the entire study and receiving the Prolific completion code, but the saved results show that they only completed the first few trials. The script saves data every trial, so I already found it strange that they completed the full study (~ 30 minutes), when only ~7 minutes are saved (we have already tested quite some people, and it's the first time we've encountered this). When inspecting the study logs from JATOS, what I find is the study starting and a "Finished study run" 7 minutes later: the finish event occurs too soon to be a fully completed study. Additionally, the message on JATOS is "FAIL It's not allowed to reload this component". However, when asked, the participant can report the full experiment (even tasks that were not described in the main description because they are surprise tasks).
Is there anything I might be missing that could explain this discrepancy? Thank you so much for your help!
Comments
Hi all,
Just to add a bit of context to this: I just realised this is a Mac issue. I was recording a participant live using the same script, and on MindProbe it showed the same error, no data stored- but the experiment continued on their end without showing any error.
Hi,
It’s possible that something in the study’s JavaScript—or a participant refreshing the page—triggered an attempt to reload the component. If reloading is disabled, that could explain the error message.
The experiment might then continue in the participant’s browser even though JATOS is no longer accepting its data. Receiving the Prolific completion code wouldn’t necessarily confirm that the data was saved successfully; that depends on how the study handles saving and completion.
You could send us the study’s code, ideally as a JATOS study export, and we can have a look at what might be triggering this and how the script handles failed data submissions. It would also help to know which browser the affected participants used on their Macs.
Best,
Kristian
Hi Kristian,
Thank you so much for your reply! I have already changed reloading to enabled, but yes, having you look at the code would be good to understand what's happening. Where should I send it?
Both participants were using Firefox because I had issues with Safari and the microphone before.