[solved] reaction time query

MarkG

Hi all

Sorry if this has been raised before - but I couldn't find anything about it upon searching (and please move it if its in the wrong area). I'm a bit confused about some of my results upon opening my output file. Sorry if this is a bit simple but i'm only at an early level in my psych career!

I've built a dot probe experiment for my research project, looking at reaction times between socially threatening, physically threatening and neutral words. They have to press the Z key if the corresponding probe appears on the left, and the M key if it has to appear on the right. The experiment seems to work correctly.

However what i'm finding is that a lot of my reaction times are coming up as 0ms when i'm running the programme - meaning that they are reacting as soon as the stimulus presents, which i've been told by my supervisor is impossible because humans cant react that fast to visual stimuli. This also means that a lot of my data is going to be screened out when it comes to writing it up, as we typically discard any RT under 200ms and im wondering if there is something that im doing wrong or havent taken into account from a software point of view?

Ive run the programme on myself several times and am also hitting 0ms by responding as quickly as I can to the probe (so its not just participants hitting the key prior to the probe, as they are also getting it right!)

Thanks in advance - look forward to hearing from you

Mark

Edwin

Hi Mark,

Couple of questions, to help us figure out what the problem is:

1) Could you upload the experiment to either a code sharing or cloud storage website (e.g. Google Drive or Dropbox), so that it is accessible to us? Alternatively, if you don't feel comfortable sharing the experiment, uploading a shorter example that demonstrates the same problem would be great.

2) Could you upload an example output file? Make sure it's the original CSV, without it being edited!

3) How are you calculating the RTs?

RTs of 0 ms are impossible indeed, and should not occur normally. But I'm sure we'll be able to figure out the problem!

Cheers,

Edwin

MarkG

Hi Edwin

Thanks for getting back to me so quickly - I hope these links work:

My programme is available at:
https://www.dropbox.com/s/v9ejv4o6isg15x4/ThesisDotProbeFINAL.tar.opensesame.tar.gz?dl=0

And an example CSV file is available at:
https://www.dropbox.com/s/f4zsy4wn1ywjy36/subject-7.csv?dl=0

thanks in advance

Mark

MarkG

Sorry - just spotted your third point!

Reaction times measured for this experiment is the time between the probe (X) appearing (following the stimulus word) and the response button being clicked (either Z or M)

n the output file im looking at: response_time as the measure of the reaction times

thanks!

Edwin

Hi Mark,

Just took a look at your experiment, and I think I've spotted the problem. The solution consists of two simple steps!

1) Set the duration of your DotProbe sketchpad to 0 rather than 500 ms. This will make sure that the keyboard_response following it will become active straight away, instead of only after 500 ms have already passed.

2) Make sure to check the box labelled Flush pending keypresses in your keyboard_response item. This will make sure are keypresses that occurred before the onset of the probe are ignored. Without checking this box, a premature keypress will be picked up. This was happening in your experiment too: people were probably responding within the 500 ms that the probe was visible and the keyboard_response was not yet active, which was then picked up by the non-flushing keyboard_response when it became active (i.e. after the 500 ms delay).

Does that fix your issue?

Other than this, your experiment looks well constructed! You could consider contributing it to our standard tasks database (I think a lot of social psychologists could benefit from a ready-made dot probe task!).

Thanks!