We have been running the above linked task for the past few months for ~50 participants with no issues, but upon running the last batch of 15 participants, we discovered that none of their data files had been saved. We have not made any changes to the task script. We ask our participants to run this task on Google Chrome and when our task ends, they are directed to press ‘space’ to submit their work and then they are automatically redirected to a Qualtrics survey. A couple of them had reported seeing an error message at the very end of the task but according to them it disappeared before they could even read it and were redirected to the next part with no other issues. I have not been able to replicate this-- when I make it to the end, I do not see an error message but my datafile is still not saved when I run it.
Any ideas on what could be causing this? Any and all suggestions would be greatly appreciated.
Hi all,
Just wondering if anyone has seen this message and has any thoughts? We are completely stumped as to what has caused this change in our ability to save data collected on pavlovia for this task. It is impacting our ability to complete this project for a student’s dissertation, and given that nothing has changed in our code since we started collecting data successfully until now, we don’t know where to turn. Tagging some of the developers/‘leaders’ on the forums here; I hope that is ok; I’m not trying to spam your inbox, but just hoping for some thoughts/guidance @jon@Michael@wakecarter
I’m afraid this has been sitting unanswered because, well, I don’t know the answer either. Without a way to reproduce it, and with not much information it’s really hard for us to know what could cause such a problem.
Have you inspected the end of your experiment for any errors in the code (like an if-statement that doesn’t get executed correctly in certain rare cases)?
The server was under a lot of pressure from people uploading spam projects to point to there sales websites in the last few weeks but the symptoms of that were usually on gitlab being sluggish, not on any failure to receive of data files, so I don’t think that’s the reason but I don’t have a better answer I’m afraid
You could try using the VESPR Study Portal. Participation is confirmed after the study is completed. I haven’t yet experienced cases where Pavlovia data was missing for confirmed participants.
After looking at this task at every possible angle and running it on our end with no data files still being saved, we decided to just clone the experiment repo without making any changes and run it off of that clone as a last resort to see if anything happens. And lo and behold, we got a datafile! We ran one participant as a test and it worked again–their data was saved. Any ideas as to what may have “broke” the first copy, or why it might only be working on the cloned version?
I am very glad that the problem disappeared with your cloning the experiment!
I had a look at our extensive logs to try and understand what might have happened.
Something odd I noticed is that on a number of occasions, a piloting token was present even though the experiment was in running mode. Could it be that the URL communicated to some of your participants was not the one mentioned on the experiment page but one copy pasted from the URL bar, on a previous occasion? Would that make sense?
I also had a quick look at the experiment code. It appears that the choice of resources to be loaded depends on the participant id. That would most certainl explained why those participants that did not enter their name could not run the experiment.
