Fast reactions not being collected
Hello,
I'm looping stimuli in quick succession (~2 Hz) until the participant responds, and then another loop begins. For some reason, when the response comes (almost) immediately with the next stimulus in the same sequence - it is not being collected (essentially ignored and the trial continues as if nothing happened). This happens both on the normal OpenSesame platform and on OSWeb (in which the experiment is intended to run on when collecting data).
Apart from a short black screen and a fixation point, each trial involves a stimulus follow by an inter-trial-interval. The participant is able to respond during both of these. Both durations are set to '0' followed by a keyboard component.
It feels like I might be missing something very trivial, but I don't know what exactly and would appreciate your help.
I can't upload the experiment directly since the file is too large but it is available through the following link:
https://file.io/KtwNItxs451F (Excuse me for the Hebrew. Press "space" to continue and "s" or "l" during trials to make a response).
Thanks,
Omer
Comments
Hi Omer,
the file is not available anymore. Could you re-upload it? If it is too big, there is always the option of emptying the file pool somewhat and/or simpliyfing the experiment, e.g. by removing parts of the experiment that you know is not related to the problem (that would also make it easier for us to focus on the problem).
Once I can look at the experiment, I can try to find out what is happening here.
Eduard
Hey Eduard,
Thanks for replying. Here's another link for an even more simplified version of my experiment.
Waiting for your feedback!
Omer
Hi Omer,
Could you explain again what behaviour you expect and what behaviour you observe? I just tried it locally and pressing keys seems to do something, i dont know though whether it is what you want.
Regardless, running it in the browser doesn't work (the experiment crashes).
Hi Eduard,
Thank you for your time.
Yes, the keys do work, but no always (at least when I try it). The unexpected behavior I observe is that when a new number in the sequence appears and I respond immediately - the response is not being collected and the sequence continues as if nothing happened. The expected behavior is that responses will always be collected throughout the sequence's presentation (after fixation, of course).
I don't worry too much about running it in the browser at the moment (works on my end, by the way). Once I can fix the response collection issue I'll try it over the full version of the experiment.
Thanks again,
Omer
Hi Omer,
the response is not being collected and the sequence continues as if nothing happened
What is supposed to happen? You don't do anything with the response as far as I can see? The only thing that happens depending on the response is the execution of the second sketchpad and keyboard response item for 20 ms. Or is there anything else?
I find it near impossible to see whether things go as expected in your experiment. I modified it a little (see attachment), when trying it, I dont see anything weird. It behaves as I would expect it.
Hope this helps,
Eduard
Hi again Eduard,
I'm sorry my explanation is not clear enough. I'll try to explain using a scheme of a trial in my experiment:
After fixation, a sequence of numbers is presented. The participant is supposed to make a self-paced response. Because the sequence is composed of random numbers, at times the same number will appear twice in a row (an intended behavior). In order to make it clear for the participant that those are two distinct members of the sequence, we use a short "blink" (a black screen appearing for 20 ms). But then, we also wanted the participant to be able to make a response whenever s/he liked. Therefore, we enabled responses to the sketchpad with the number or the "blink" sketchpad. While it works mostly as intended (green 'v' icons), if a response is made when a new number in the sequence is presented - it will not be collected (red 'x' icons).
* This does not only happen when a number appears twice in a row. That's for illustration purposes and happens for any number.
Hope I was able to make it clearer now.
Thank you,
Omer
Brilliant! Very clear explanation, why not right away? :)
Ok, that's probably as easy fixed as unticking "Flush pending key events" in the keyboard_response item. I am not 100% that will work exactly as you intend. For example, it might be possible that one response would trigger multiple trials to end -- unlikely though.
For more flexibility and more control you probably would need to use an inline_script, but it seems like this study is supposed to run online, so never mind..
My apologies :)
I have just tried to untick "Flush pending key events" now, but sadly I still experience the same behavior (early responses to stimuli not collected, while later responses are collected correctly). It did not make anything else go wrong, though.
Thank you for trying... Any other suggestions?
O.K. I got it now... While eduard was right about unticking "Flush pending key events". What was missing was also unticking "Flush pending key presses at sequence start" under the sequence tabs. It seems to work as intended now.
Thank you for your help!
Hi Omer,
Good to hear that you figured it out (and sorry for not having gotten back to you earlier).
Cheers!
Eduard