Description of the problem: Today I ran a jsPsych experiment on TurkPrime with incomplete results turned off for Pavlovia. TurkPrime logged 293 complete responses but Pavlovia only logged data for 274 participants. This difference matches the 19 credits that are still reserved. However, participants only received the secret completion code after the task finished the connection with Pavlovia. I’m wondering if Pavlovia may have failed to connect, such that the credits remain reserved, the data is marked incomplete, and participants nonetheless received the code?
If so, is there any way to recover these data? Unfortunately this amounts to over $100 of lab funds spent on missing data. It seems the alternative is that we could have saved incomplete results, but then we would have spent credits on the ~125 participants who aborted the task. Either way we lose money… Would really appreciate any insights!
I am not as familiar with jsPsych as I am with PsychoJS but reading through the documentation on how to use it with Mechanical Turk, https://www.jspsych.org/overview/mturk/, and looking at your code I wonder whether it could be that since the secret code is the same for everyone it is possible that some participants got the code from a previous participant, accepted the Mechanical Turk HIT, started the experiment and then entered the code almost immediately. That would explain the discrepancy perhaps?
The jsPsych documentation talks about submitting data to MTurk but this is not your approach, hence my believing that your experiment is running in an iframe in MTurk. Does that make sense?
I’m actually having this issue as well. However from participant survey responses it seems like people are actually doing the task, but for whatever reason their data isn’t being submitted. Do you have any suggestions to solving this issue?