Post by verena on Apr 7, 2020 19:38:35 GMT
Hello everybody,
I hope you are all fine!
I have a really troubling timing issue in a quite simple experiment, and I hope you can help us.
We want our subjects to rate voice recordings. We present a total of 186 stimuli (N = 40 male speakers and 53 female speakers, each comes with two sentence stimuli S1 and S2; everything in .wav format). Stimuli are blocked by voice gender. After each voice presentation the subject should rate the voice. We want to measure reaction times from voice offset on.
The problem is, that the rating scale appears already very often (but not always) during voice presentation.
We figured out, that this is the case more often for female stimuli, and more often for S1 sentence stimuli. S1 stimuli are slightly longer than S2 stimuli (M = 2386 ms, vs. 1876 ms). But female stimuli are of comparable length compared to male stimuli.
we run several tests (each with 28 trials) on various Operating systems (Windows 8, 10, Chrome OS, Mac OS Catalina, High Sierra) and using different Browser (Firefox, IE, Chrome, Safari). The problem occurred on each machine and browser.
Why does this happen and what can we do about it?
Thank you very much. I appreciate your help!
Best wishes
Verena
This is the trial code:
part runTrial
set &TrialCnt increase
set &TrialCntBreak increase
keys 1 2 3 4 5 6 ? ; . ! kp1 kp2 kp3 kp4 kp5 kp6
show bitmap slideBlack. #a black screen for 1000 ms
delay 1000
show bitmap slideLoudspeaker. #a loudspeaker symbol to announce the test voice (for 200 ms)
delay 200
sound @1 #the test voice
delay @8 #the duration of the test voice in ms
show bitmap &ratingScaleSlide #the rating scale
readkey @9 3000
if STATUS == 3 ##no reaction
show bitmap slideTooSlow #feedback if a particiapnt is too slow (for 500 ms)
delay 500
fi
if STATUS != 3 ##reaction was fast enough
show bitmap slideBlack #a black screen for 500 ms, if the participant was fast enough
delay 500
fi
save &ratingScale &TrialCnt @1 @2 @3 @4 @5 @6 @7 @8 RT KEY STATUS BLOCKNAME TABLEROW BLOCKORDER
I hope you are all fine!
I have a really troubling timing issue in a quite simple experiment, and I hope you can help us.
We want our subjects to rate voice recordings. We present a total of 186 stimuli (N = 40 male speakers and 53 female speakers, each comes with two sentence stimuli S1 and S2; everything in .wav format). Stimuli are blocked by voice gender. After each voice presentation the subject should rate the voice. We want to measure reaction times from voice offset on.
The problem is, that the rating scale appears already very often (but not always) during voice presentation.
We figured out, that this is the case more often for female stimuli, and more often for S1 sentence stimuli. S1 stimuli are slightly longer than S2 stimuli (M = 2386 ms, vs. 1876 ms). But female stimuli are of comparable length compared to male stimuli.
we run several tests (each with 28 trials) on various Operating systems (Windows 8, 10, Chrome OS, Mac OS Catalina, High Sierra) and using different Browser (Firefox, IE, Chrome, Safari). The problem occurred on each machine and browser.
Why does this happen and what can we do about it?
Thank you very much. I appreciate your help!
Best wishes
Verena
This is the trial code:
part runTrial
set &TrialCnt increase
set &TrialCntBreak increase
keys 1 2 3 4 5 6 ? ; . ! kp1 kp2 kp3 kp4 kp5 kp6
show bitmap slideBlack. #a black screen for 1000 ms
delay 1000
show bitmap slideLoudspeaker. #a loudspeaker symbol to announce the test voice (for 200 ms)
delay 200
sound @1 #the test voice
delay @8 #the duration of the test voice in ms
show bitmap &ratingScaleSlide #the rating scale
readkey @9 3000
if STATUS == 3 ##no reaction
show bitmap slideTooSlow #feedback if a particiapnt is too slow (for 500 ms)
delay 500
fi
if STATUS != 3 ##reaction was fast enough
show bitmap slideBlack #a black screen for 500 ms, if the participant was fast enough
delay 500
fi
save &ratingScale &TrialCnt @1 @2 @3 @4 @5 @6 @7 @8 RT KEY STATUS BLOCKNAME TABLEROW BLOCKORDER
