Hi, I am trying to pilot an experiment on Pavlovia. For many-many days, the incomplete csv data was saved locally, so I could check things (e.g. which trial generated the error etc). This stopped yesterday, despite the Save incomplete data is ON. I have the problem both when using @wakecarter’s web app to generate consecutive participant number and when not using it…so I guess it is not related. The demographic survey data is saved (json file) but that’s it, no csv. any idea what might be the cause? Another issue is that the exp stops running when it gets to a certain image, saying unknown resource. The image is among the resources in the Experiment Settings tab, it is in the images folder and so far, there was absolutely no issue with it.
This isn’t another issue. Your experiment is crashing, which is why it isn’t saving data.
Is it always the same image? There might be a space in the filename, or the file could be corrupt.
Hi, Interesting. So far I had data saved locally whenever errors popped up…I found the root cause (one of the many conditions files, one parameter of the image was somehow missing)
Locally data saves incrementally, so you get data on a crash. Online it saves at close, so a crash can’t save anything (unless you have periodic saving turned on, which I haven’t explored)
I had data when piloting online. The periodic setting was turned on. Anyways, I found the cause, so hopefully it will be okay. I know this is an unrelated problem but I don’t wan’t to start another thread: Did anyone find any solution for the blurriness of the images and texts?
Here is the thread about what we currently know about online blurriness. Feel free to add a new post there.
Hi @wakecarter, No datafile (csv) is saved in piloting mode, even if I complete the whole test (no crashing). I get the survey results (they are automatically downloaded) but nothing else, though previously this worked. Save incomplete results and Save periodically are on. I would really need to look into the datafile (to see what is saved, how precise the timing, mouse pos registration etc is online before I turn it into running mode but now I don’t know how). Do I have to put sth special in the code (e.g. saving at the end?)
I would turn off periodic saving and try again. Do you have any code in End Experiment?
Hi, There is another short pavlovia survey (a debriefing one). And, maybe that is the issue! Previously I had a ‘thanks and goodbye’ routine at the end and nothing else. I will try to re-add this routine after the survey and run the experiment again this way (with and without periodic saving turned on) tomorrow.
Hi @wakecarter, So It does not save data in pilot mode, whatever I try. The experiment is “completed”. It is listed as “submitted” but only the survey is saved locally at the end (json file). This happens both when the Save periodically option is on and when it is off. Save incomplete data is on (it was also on before when data was saved automatically). I am covered by a departmental license. The structure is the following: consent screen, demographic survey (pavlovia survey) then a large loop contaning the actual experiment and finally a nonconsent goodbye screen in a loop (rep turns 1 if the person did not consent at the beginning).
Try adding a short routine (1 second blank screen or “Please wait” should be plenty) in between the end of the last embedded survey and the end of the experiment.
You could try outside the loop – to the far right of the flow.
If the experiment can be shortened to a test version I’d be happy to take a look (username Wake).
Save incomplete shouldn’t be an issue. What I previously discovered (but am not 100% sure about) was ending on an embedded survey and then immediately redirecting to a new Pavlovia Survey (which I was using for post submission debriefs).
Hi, placing the “end expe screen” routine outside the loop solved the problem! It would be good to understand though, why having it inside the loop did not help…Thanks for the suggestion.
It might be that the survey data gets written to the data file at the end of the loop, rather than immediately after the routine. I’m guessing here, but I am glad it solved the issue.
Hi @wakecarter, So finally I could check the datafile and noticed that certain image attributes that are included in the conditions files do not show up in the datafile (the header is there but the cells are empty). These attributes are read in and successfully used as variables during the experiment. Specifically, the size and position of the images are missing. These are defined as arrays, e.g.: [0.15, -0.25]. I do not save any information included is in the conditions file explicitly, and for all the other type of information included (image paths etc.), this is not an issue. Do array type of data work differently? If yes, do I simply have to add the the data with psychoJS.experiment.addData(??), using the respective conditions file header as ‘name’?
That sounds like the square brackets are confusing PsychoJS. I usually use separate columns for x and y.
I’ve just done some investigating and discovered the following.
Locally you can use square or curved brackets in your spreadsheet (or use separate columns). Online, curved brackets fail to be interpreted as numbers and square brackets work but the entry isn’t saved in the spreadsheet. You can, however, manually save the square bracket entry using thisExp.addData, but only if you give the saved version a different name. The numbers in the brackets then have quotes added.
Hi @wakecarter, Unfortuntely, the original issue is back. Data is not saved in pilot mode (at least when using your app). This is independent of whether save periodically is off/on. The internet connection is stable. Any idea what might cause this? There is a routine at the end and the only thing I changed is how the img pos and size is given in the conditions file (there are 2 variables instead of an array).There is no error message, just no data at all at the end, other than the first survey’s responses in json.
I am somewhat unclear about the exact issue you are experiencing.
Are you saying that you can finish an experiment run on Pavlovia in pilot mode and it doesn’t download a csv file at all? I suspect that there must be an error… perhaps check the developer console.