Data not saved when using Safari

I have an online experiment on Pavlovia where the status is set to “running.” Everything is normal and the data can be downloaded until I have someone who completed the study using Safari (all the other ones were completed using Chrome)—neither can I see that he consumed one credit nor can I see his data.
I tried looking for this issue in the older posts but haven’t seen a super relevant one yet.

For PsychoPy version, I have the “Use PsychoPy version” under “Experiment Settings” set to blank.

If anyone knows a solution/something relevant to this issue, I would love to hear them. I really appreciate your help—thank you!

Hi Enting, MacOS Safari is indeed not very stable. My tests hint that PsychoJS does run on Safari + MacOS Catalina, but not on Safari with older MacOS versions. We don’t have a solution yet, we are first setting up automated testing pipelines to figure out where exactly the problem is.

In general, we recommend using MacOS + Chrome, not only for stability but also better timing. Here are two refs for that:

Thank you so much for the answer! After asking my participant, I realized that he was using the beta version of MacOS + Safari, and that might also add onto the problem. So I guess now we know a MacOS version too new is not ok either haha!

Oops…sorry that I did not click on the “Reply” button!

I think the issue that I raised in the first post is mostly resolved…except there are more new issues, I will post here just in case anyone knows the answer:

  • I have a friend who told me she got to the end of the study, hit the “OK” after the system popped up the window “thank you for your participation” or something like that. But after I went to look for her data, I did not see her data, which really should be properly saved when I download the csv files. She used Chrome. I’m really puzzled by this and haven’t found an answer yet.
  • I have my study linked to Qualtrics, and those are all friends whom I’m really sure at least 80% of them did some parts of the study, and at least should have some results saved (I switched “save incomplete results” on). The problem is, I can see that they completed the Qualtrics questionnaire, but there was no corresponding data on Pavlovia. Despite there’s a chance that they didn’t do the Pavlovia study at all, I’m suspicious that the data was not being properly saved, since I can see quite a few files that were saved, but upon opening them I only saw an empty “Participant” field and nothing more, no data. Any idea of why this is happening?

I would really appreciate any inputs, thoughts, discussions, past experiences. No need to be an answer! Just want to know if anyone else is running into the same problem as me.

Thank you guys!

The feedback reported by your friends does suggest the data got sent to Pavlovia; the system is quite careful with that.

What happens if you take part yourself? Any data then?

Yes, I did it myself and everything was fine.
I also had people who quit halfway (without even getting to the last page, let alone hitting the “OK”), but their data got saved fine.

I’m still going through some posts in the forum, but a few questions still linger:

  • The “OK” pop-up window is normally shown several seconds after the participant finish the last page and hit ESC, so if they do that and if the system works fine it should save the data. What if the participants did not wait until the pop-up window to show up, got to the last page, and directly closed their browser window—will the data still be saved? Will the data be reserved?
  • When a session has reserved credit, the documentation only says that indicates “the participant initiated the session.” However I’m not quite clear if that means they quit the study halfway, or just got to the first frame of the study and closed the browser window immediately.
  • What does it mean when the credit is “reserved”? A similar post mentioned that reserved credits will be released 24 hrs afterwards, but does it mean that the “incomplete” data will be made available to download in CSV format after 24 hours only? Does that possibly explain why I don’t see the data even though I know some people finished it?

A similar post is here, just for reference: Pavlovia credits reserved, no data

For your last two questions, I’m afraid I don’t know the answer. For the first, I got something to say though. I think data is sent to Pavlovia during the task, so if they would quit before seeing the last dialog box, I’d still expect some data being logged. You could check this by running a couple of trials, waiting a minute, then close the task. If there is no data at all, I’m inclined to suspect that no data is sent, most likely due to some bug in the task.

Something else you could check are the logs instead of the CSV files by the way

Thank you!
The experiment is set to “running”, and I can see that some people quit halfway through the task because I was able to see that the data was saved successfully and from the length of the trials that they did not finish, meaning that this was doing its job, but only for some people, not some others…
There are people who only have CSV files but not logs, what does that indicate?

Hello @ening,

To answer your questions:

  • Since you decided to save incomplete results, a credit will be consumed whenever a participant runs your study, even when he or she decides to stop half way through. We do check however that a little bit of data was collected, i.e. a credit is typically not consumed if the participant closed the tab after the first hello page.
  • Those participants for which you have .csv files but not .log files are typically those that interrupted the experiment by closing the tab in which it was running.
  • A credit is reserved at the start of the experiment, when a session is opened. It will stay reserved until the experiment is completed, or until the participant decides to abruptly go away, at which point it will be either consumed or released. In your case, it will be consumed when the partial results are saved.

I have a question for you: you mention that participants have to hit the Escape key at the end of the experiment. Why is that? Why not just let the experiment conclude?

Alain

Hi Alain,

Thank you for raising this very obvious but I haven’t thought about question. I now changed the ending to a limited duration where the experiment will conclude by itself right away after the subject reached the end page. I didn’t think about it clearly (or thought it’d be an issue when I went with the ESC key), but now it seems like a bad idea…Do you think this might be where the problem of not saving data is?

But if this is where the problem is, it doesn’t really explain those with CSV files but no logs, meaning that they probably did not press ESC but still managed to have their data saved.

About reserved credit and corresponding data:
I checked again today, after 24 hours have passed, and do not see any new data—although I’m not completely sure (definitely should keep check on that though) that I had 1 reserved credit yesterday, I know a friend who completed the study all the way, but her data didn’t show up when I downloaded saved data. So I yesterday I thought it had something to do with credit reserve and decided to wait for another day to see if I have anything new, but it turned out that I still do not have her data.

Emma

As I mentioned, those sessions with .csv files but no log files are those which the participants interrupted abruptly by closing the tab. And indeed data is saved immediately, there is no delay, so data will not appear after 24h, I’m afraid.

Alain

Your answer is really helpful. Thank you Alain!

Hi Ening, did you find a solution to this or explanation? I’ve just run in to the same exact issue.