4
$\begingroup$

There are a few open-sources tools available to run psychology experiments. However, some are working in a browser. Do we know how precisely they can record response times? Also, can we present a visual frame for a precise amount of time? Considering refresh rates and wait_top signals, this is not trivial. The few tools I have in mind are the following, but the answer may include other platforms as well:

  • jsPsych (browser-based)
  • PsychoPy
  • OpenSesame
  • PEBL

If good reviews of these software have been published, please share reference.

$\endgroup$
2
$\begingroup$

Short answer
"Precision" and "accuracy" are relative parameters and the interpretation depends on your demands; their exact values depend on your hardware and software.

Background
RT measurements are finicky, even on a single pc simple basic USB polling latencies can add unacceptable latencies (8 ms or so) to RT estimates. Graphics latencies in visual tasks, or syncing issues in auditory-visual tasks can seriously mess up the stimulus timing. First-time calling of Matlab commands may add unacceptable latencies to the first trial. Etcetera etcetera. The Psychophysics Toolbox site has a wealth of information on these things.

Depending on why you wish to use a browser environment, but if you are attempting to do remote experiments over the internet, and assuming the software is stand-alone on the pc on the other end, you cannot control above-mentioned parameters at all. Further, if your pc is driving the experiment, things can get a lot worse, because a server or two will process the data in-between and there's the pc on the other end that you can't control.

But everything depends on the type of reaction time you are after. Are you interested in millisecond-range responses (standard react as-fast-as-you-can tasks), or do a few seconds more or less not matter (tasks where subjects get the time to make a cognitively evaluated choice)?

Also - if you are interested in latency differences (e.g. before and after drinking a few beers), then systematic latency deviations due to hardware issues do not matter anymore, as they are contained in both conditions and will be subtracted.

So it all depends on the experiment you are to use, the hardware that will be deployed and most of all, the type of response you are testing.

| improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ Hello @AliceD. I do look for milisecond RTs, in absolute value, not a difference with another condition. Thanks for the link. $\endgroup$ – Denis Cousineau Feb 3 '17 at 12:39
  • 1
    $\begingroup$ Quite honestly I do not have experience with online RT measurements, but given my experience with offline measures I reckon that absolute RTs in the millisecond range are going to be pretty hard using regular hardware. $\endgroup$ – AliceD Feb 3 '17 at 12:41
  • $\begingroup$ jsPsych can be used over the internet, but it doesn't have to be. The OP never said anything about it being used online, only that it was being used in a browser. Browsers can also be used to run local software. $\endgroup$ – TheBlackCat Feb 3 '17 at 21:22
  • $\begingroup$ @TheBlackCat - good point. I'll adapt the answer. $\endgroup$ – AliceD Feb 3 '17 at 22:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.