Data doesn't save

Description of the problem: Hello,
I have started collecting data, but since then, two of my participants (out of like 13 or so) reported that the page returned to the participant prompt screen in the middle of the experiment. When I check the database, I see empty csv files, although I have the “save incomplete results” option on, so those data are not saved at all. Needless to say, the experiment is on “running” mode rather than pilotting.

Does this issue stem from the fire in the server building? Does anybody have a clue?

thank you in advance, tutku

Hi @tutku,

It’s very unlikely that this issue is related to the fire. But it’s strange though. The only scenario that comes to mind in which this would happen, is when participants reload the web-page half-way, for instance by pressing “F5” on the keyboard. If this is done relatively early in the experiment, it’s likely not data has been saved yet.

Maybe ask your participants a bit more about the when & what? Where exactly did it happen? What did they do just before it happened?

Best, Thomas

thank you so much for your reply, Thomas. one participant reported that the issue occured around like 15 minutes ahead of her start to the experiment (the experiment takes around 20 minutes or so in total to complete). But I’ll dig into it. indeed she might have accidentally pressed the f5 button, thats possible (although she didnt mention anything about it).

best, tutku

1 Like

You’re welcome! If it’s of any use, I once wrote a manual for getting detailed info from participant (which in turn can be useful for you and me do diagnose the problem). The manual is over here: https://thomaspronk.com/pdf/Tech_Support_and_Bug_Report_Guidelines%20v1.3.pdf

hey there again, another subject reported the same issue. I asked her if she had an internet connection or she accidentally pressed a button. she said she didnt. the thing is, all participants reporting the issue said having the issue in the middle of the experiment (i.e., where I put a block break for 60 seconds - they said the experiment didnt move forward and went back to the participant prompt screen).
This is the experiment link: https://run.pavlovia.org/oztel/mainyoking_full/html

though, i use https://moryscarter.com/vespr/ for the consequtive participant ID assignment. would that be the problem?

thanks a lot, tutku

Günaydın!

If I click the link I immediately get an error (TypeError: Cannot read property ‘0’ of undefined). I guess that’s because I haven’t filled in a valid participant and/or katilimciNo). Regardless, thanks for collecting the info; this can help pinpoint what the problem is. Could you share your repo with me? Then I’ll take a peek at the code (tpronk, make me a delevoper). This guide explains how: Searching for experiments on Pavlovia — PsychoPy v2021.1

Cheers T

Günaydın to you too @thomas_pronk (assuming you are somewhere in Europe)! :slight_smile: I just added you as a developer to the project.
and yes, the experiment needs at least a valid value for “participant” section, could be something in between 1 and 46. the katilimciNo isnt really necessary for the code to work, but I do some condition assignments according to “participant”.
Thank you so much for your help, tutku

Yep, about 3 hours before you I guess :). I don’t see anything strange at all in your experiment, so I can’t explain it. I’ll ask my team if I missed anything about experiments reloading. Meanwhile, maybe you could ask your participants which OS and browser they were using?

thank you so much for all your help, @thomas_pronk much appreciated. I know, its weird, I pilotted the experiment for maybe a thousands of times before running it and encountered many problems, but none as such that what a few participants (not all of them) are currently reporting now.
The participants reporting the issue were using Safari and Chrome, but didnt ask for OS. I’ll do so as well.
If your team figures out anything wrong, can you please let me know?

cheerio, tutku

You’re welcome! And yes, I’ll do so.

Arrrghh, same issue in a different experiment of mine. it is getting really frustrating that I am losing data and dont know why… I collected massive amounts of data from pavlovia before, never encountered such a thing. seriously, could it be related to the fire in any way? It doesnt even save the data that is half through although the participant stated that she was doing the experiment for like 75 minutes straight. and the experiment window just randomly closed.

the link of the experiment is: https://run.pavlovia.org/oztel/dualtask_letter2/html

@thomas_pronk , in case you have time to take a look, I can make you developer on this one as well.

best, tutku

Gonna be monday before I’ll take a look, but sure, hook me up

done, added you as a developer :slight_smile: thanks a lot @thomas_pronk

My experiments are used for merging tests at the moment, so I’ll need to hold off diving into yours for a while. Update soon!

thank you so much for the notice, @thomas_pronk :slight_smile:

Took a look. I see that you’ve been editing straight in JS (instead of via the builder) and I notice that the version of PsychoJS being used is quite an old one (v2020.1, while we’re now at 2021.1.2). I see some code in there that doesn’t actually do anything, such as order_num.toString(); on line 3726.

I’m sorry to say that I can’t really debug this code; it’s way too complex to get a grasp of. Taking a general look I’d say: consider rebuilding it in the most recent version of PsychoPy builder, don’t edit the generated JS.

lol thats a sad news indeed, but thank you so much for giving it a look :slight_smile:

1 Like