Accuracy values rounded in OpenSesame 4.1.20: intended behavior or bug?

Comments

• sebastiaan

  2:17PM

  Dear @fotisfotiadis ,

  This is intended behavior. The rationale behind it is that the acc and avg_rt variables are to provide human-readable feedback to participants during the experiment (and not used for offline analysis), and for that reason you don't need decimals.

  Looking at the commit history, this was already implemented in 4.0. But indeed, before that time, these variables weren't rounded.

  — Sebastiaan

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• fotisfotiadis

  2:38PM

  Dear @sebastiaan

  Thank you for your prompt response.

  I would just like to add that the accuracy variable can also be useful for ending training once a predefined accuracy criterion has been reached.

  If I may offer a suggestion, it would be helpful to have both a rounded accuracy variable for feedback and an unrounded version for finer control over such criteria.

  Thank you again for your help.

  Best,

  Fotis
• sebastiaan

  2:49PM

  I would just like to add that the accuracy variable can also be useful for ending training once a predefined accuracy criterion has been reached.

  If you need an unrounded value for this, then you can also use the total_responses and total_correct variables (which I realize now are not documented). Or use the responses object:

  • https://osdoc.cogsci.nl/4.1/manual/python/responses/

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!