Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Logging Responses Within The Same Loop in the Response Column

Hi everyone,

I am creating an experiment for JATOS and using OSWeb to do this. The problem is, as you could see in the attachment, I have different keyboard responses in the same loop and they are not stored in the "response" column in the log file. Logger saves the responses in different columns (for example, as response_ai_q1, response_ai_q2, etc) and response column remains empty for that part of the experiment (or sometimes logs the last response only). I cannot put logger after each keyboard response, experiment just stops without giving an error. Is there any way to log each response within the response column?

Thanks for your help in advance.

Best,

Jonah


Comments

  • Hi @wishiknewprogramming ,

    There is a known issue with OSWeb where multiple linked logger items (and certain other items) in the same sequence cause a crash. To work around this, you can simply put the logger as the only item in a sequence , and then use multiple linked copies of this sequence item.

    Like so:

    Hope this helps!

    — Sebastiaan

  • edited November 2020

    Hi @sebastiaan ,


    Thanks for the quick response! If I put the sequence you suggested both in the beginning and in the following responses in the same loop, experiment works without crashing. However, it still doesn't log responses of each keyboard response in the "response" column in the log file. It just logs all the responses in seperate columns with their names (e.g. response_ai_q1_response, response_ai_q2_response, etc.) in the same row. I also tried to put an inline java script after each keyboard response, which is:

    vars.response = vars.response_ai_q1_response

    vars.response = vars.response_ai_q2_response

    etc. so I thought that if there would be multiple variables with the same name, it might store it in the same column, one after another. However, this doesn't work either. I also changed the name of the keyboard response, i.e., used ai_response for the keyboard responses and used multiple copies of the same keyboard response item for the each response. It didn't work either. Am I doing something wrong, or is not there any way to do that?

    Thanks for your help.

    Best,


    Jonah

  • Hi @wishiknewprogramming ,

    Why don't you the attach the experiment to this discussion, or post a screenshot of the overview area?

    — Sebastiaan

  • Hi @sebastiaan,

    I finally changed the structure and created loops for questions, rather than putting every question in a different sketchpad in the same loop. Now it is working! Thanks for your help!

    Best regards,

    Jonah

  • Hi,

    I'd like to follow up on your disccusion. I had a similar problem and now tried to solve it with the logger_sequence. I ended up with a new error that occurs when testing the experiment in the browser due to the compatibility check:

    "The prepare phase for item logger_sequence is called multiple times in a row The prepare phase for item logger_sequence is called multiple times in a row The run phase for item logger_sequence is called multiple times in a row The run phase for item logger_sequence is called multiple times in a row The prepare phase for item logger_sequence is called multiple times in a row The run phase for item logger_sequence is called multiple times in a row"

    I'm running an online experiment via OSWeb. Necessarily I have some nested loops, that is, the outer loop contains several logger-items, but at different levels (in the nested loops, if necessary). Could this be a problem? 

    OR: Can this error be ignored?

    Thanks in advance!

    Carla

  • Hi @CarlaC,

    It is generally not a good idea to multiply the loggers, but it's not really possible to address the issue in your specific cae without seeing exactly how you've implemented the logger. Could you post your experiment to this forum (or a stripped down version of it that generates the errors you described?).

    Best,

    Fabrice.

    Buy Me A Coffee

  • Hi @Fab

    sorry for the late reply, I had to restructure my experiment due to technical issues. 

    I solved the problem by giving each logger sequence a unique name (while the loggers themselves are linked). 

    Best,

    Carla

Sign In or Register to comment.