Reaction times for key presses are < 1 sometimes

Online experiments
1

Hi,

SOLVED: I looked at the data with LibreOffice and did not realize that it recognized the . as thousands separator…

I tried to set up a manikin task inspired experiment, and after I had some problems with Builder I edited the JavaScript code manually. Now the source code is probably more chaotic but everything seems to work on Firefox (probably not in chrome). However, I have one problem which make me think that I might have done something wrong:

  1. On the first page the participant has to press space, and then either “up” or “down” after an image appears.

If I just press one key for testing purpose (e.g.up) all the time, reaction times are between 1 and 4 (which is what I expected). However, if I actually do the experiment more or less seriously, reaction times mostly range between 400 and 1000 (which is also what I expected), but some reaction times are below 1 (which is unexpected and I can’t find an explanation for this behavior).

Link to Project: https://pavlovia.org/run/kaja/manikintask4/html/?avatar=manikin&control=control

Do I have to make the gitlab repository public or do you have access to it somehow? If yes, please note that I only edited the .js file, not any .py file or the nomodule.js file, so you cannot open the project in Builder.

Best regards,
kaja

2

@kaja, have you tried using a code component to insert your custom code into the JavaScript? There is now a dual view code component, where you can select the code type for inserting Python or JavaScript code, or you can use “Both” to view both code panels simultaneously. Creating your experiment using the Builder interface, rather than changing the JavaScript code directly, would be a better option to help ensure your response timings are correct.

3

I’m so sorry, I just realized that I was too stupid to use LibreOffice when looking at the data. It recognized the dot as thousands separator, so 1.200 became 1200. Logically, a reaction time of 0.5 was still displayed 0.5. So this definitely wasn’t a PsychoPy error.

@dvbridges: Thank you for your reply anyway, I did not realize that you could enter code that way, I will try to change the code for this when I have time for it.

1 Like