I use Limesurvey already for a number of surveys and want to extend it's use.

I am interested in doing surveys using the Harvard Implicit Association style testing.
implicit.harvard.edu/implicit/

This requires split second accuracy on timing.

Does anyone know how the "save timings" is implemented?

Specifically, is it client-side or server-side, ie. is it sensitive to lag and transit time from a client to server?
Or measured client side so the recorded time is not sensitive to lag?

In case you want more about what "Implicit" measurement is about...
implicit.harvard.edu/implicit/demo/background/index.jsp

My solution for IATs has been to use array-text questions with JavaScript to present the terms and record the results. As this is all client-side, there is no noticeable lag.

Here's a screenshot of the concept (the results array would, of course, be hidden in a live survey):

Anyone else have a solution to using Limesurvey for IAT style testing?

Anyone have experience implementing IAT style surveys on Limesurvey?

Not sure if i understand or if i answer to question but :

Did you try www.limesurvey.org/manual/LimeSurvey_Glossary#Timings ?

Think timing is set in PHP, but not sure.
(have a js timing, but validate in PHP)

Thanks, I am not sure that information is detailed enough.

In the case of timings measured at the 100th to 1000th of a second, then knowing when the timing starts and when it stops is critical.

Case 1: php timing. Timing starts from when the server sends the page, and then stops when the response is received by the server. This does NOT work for IAT style assessments

Case 2: js timing. Timing starts from when the page is displayed to the user and stops from the moment the person clicks the answer.

While for most survey situations, the difference between Case 1 and 2 is negligible. However, for IAT style testing, the differences are significant.

Does anyone know which timing Limesurvey uses?

I have added a workaround for an Implicit Association Test (IAT) - www.limesurvey.org/manual/Workarounds:_M...ation_Test_.28IAT.29

WOW!

Thank you very much for making that work!

That is way more than I was hoping for and a great example of what makes the limesurvey community great.

Tony, that link seems to go nowhere, can you put that somewhere I can access it?

Nevermind, found a way to get to it:
www.limesurvey.org/manual/Workarounds:_M...ation_Test_.28IAT.29