Different feedback offline vs on Pavlovia

URL of experiment: Sign in · GitLab

Description of the problem: In a simple bandit task, choosing a winning picture out of 4 will show a winning feedback image, otherwise showing a losing feedback image. When running offline (macOS Catalina), the feedback display was correct given the correct answers, but not on Pavlovian when running on tablet (macOS safari and iPad, tested the same on both Chrome and Safari).

End of routine in trial:

# check if correct (either mouse or keyboard)
if key_resp.keys:
    corr = key_resp.corr
    rt = key_resp.rt
    rt = touch_resp.time[0]  # annoyingly mouse is a list of rts
    if (('upper_left' in touch_resp.clicked_name[0] and CorrAns=='u') or
        ('upper_right' in touch_resp.clicked_name[0] and CorrAns=='i') or
        ('lower_left' in touch_resp.clicked_name[0] and CorrAns=='j') or 
        ('lower_right' in touch_resp.clicked_name[0] and CorrAns=='k')):
        corr = 1
        corr = 0

thisExp.addData('rt',  rt)
thisExp.addData('corr', corr)

And begin of routine in feedback

feedback_images = ['lose.png','win.png']
if corr==0:
    msg="You lost"
    msg="You won"
# Set the image to the new image

Tablet recorded csv (CorrAns - correct answer, touch_resp.clicked_name - selected response, corr - selection based feedback)

Offline recorded csv (CorrAns the same as above, not shown, selected response the same, but corr are all 0)

Question: are there any special adjustments needed for Pavlovia mode? There are two additional rows in tablet mode recording responses to instructions, could it be the reason for this difference?

Thanks in advance.

Could you share the GitLab repo of your experiment? This page explains how. My username is: tpronk

Thanks Thomas, I’ve figured it out already, I’ll delete this post.

1 Like