Delay betweent local reaction time and OSWeb reaction time

goran123

Hi! We are hosting the same experiment both online and in person on a local device. We want to know whether there is a delay between reaction times in experiments done online and done on a local device, and if there is, how great it is. And if the delay is great, how we can reduce it.

kri

It'd say it depends. So, if you run your experiment both times (online and local) on JATOS with the same code the big difference is the lag caused by the internet when you transfer data with the JATOS server. And here it depends on your experiment design: If you only transfer data at the end of your experiment it should have no difference. But if you do a lot of sending result data or batch session accesses, it can add some lag.

Best,

Kristian

elisa

Hi Goran,

As Kristian said it depends on the code. But if you're coding things properly, then you should be collecting reaction times in a way that is independent of your internet connection. That is: you have your stimuli ready, or whatever thing it is that you're asking people to react, and then collect the time between the stimulus shown and the key press. That means that everything happens, in effect, locally, as there is no communication between server and local browser within your reaction-time measurement window.

What is more important is the system (hardware, software) that people are using remotely. Those can greatly impact your RT measurements. I recommend you check out some papers published on this (or somewhat related topics), like:

https://psyarxiv.com/jfeca/

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7512138/

https://www.degruyter.com/document/doi/10.1515/psych-2022-0003/html?lang=en

Hope this helps

Elisa