Error: Reaction Times are Measured Incorrectly?

Plus d'informations
il y a 4 mois 2 semaines - il y a 4 mois 2 semaines #179994 par ultimatte
Hi everyone,

I have been successfully (imho) using Limesurvey for three years and I love it. Thank you for your great work.
However, now, I have encountered an issue with which, please, maybe you could help me:

I used Limesurvey Version 3.15.6+190108 locally on XAMPP (Windows 10, 64-bit) to design a survey.
Limesurvey, of course, measures, reaction time for each page that is answered.
In my survey, I created several pages with text that forward automatically to the next page after 3 seconds. There is no option to interact with these pages.
I was hoping to measure a general latency of the hardware and internet setup of future online respondents to my survey.
Unfortunately, when I activated and tested the survey I realized that the reaction time measures vary considerably for these pages.
I tested three times and the reaction times varied between 3.69 and 4.91 seconds, although, of course, I tested the survey locally with the exact same setup and the pages forwarded each after 3 seconds.

Do you know if this is an error in Limesurvey?
Is there a way to prevent this, i.e. to suppress this variation of reaction time measurement?

Thank you so much,
Daniel
Dernière édition: il y a 4 mois 2 semaines par ultimatte.

Connexion ou Créer un compte pour participer à la conversation.

LimeSurvey Partners
Plus d'informations
il y a 4 mois 2 semaines #179996 par holch
Webpages never load within the same time. There are so many factors that influence that (Browser, webserver, database server, OS, etc.).

I find 3.69 - 4.91 actually estonishingly good for your setup (even if you are running things locally, there are so many things that play into this.

I think if you want to run tests with the level of accuracy that you expect, you probably can only do it within a native application (not web based) and after a lot of testing.

I would always take the times in Limesurvey (or any other webbased survey system) with a grain of salt and I would definitely not rely on this if you need accurace of 1 second or less.

I'm not a LimeSurvey GmbH member. I answer at the LimeSurvey forum in my spare time. No support via private message.
Some helpful links: Manual (EN) | Question Types | Workarounds

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 2 semaines #179997 par holch
By the way: LS does not measure "reaction" time. But the time between when the page is loaded to when the respondend moves on from the page.

I'm not a LimeSurvey GmbH member. I answer at the LimeSurvey forum in my spare time. No support via private message.
Some helpful links: Manual (EN) | Question Types | Workarounds

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 2 semaines #180000 par ultimatte
Thank you very much for your reply.
I realize that reaction times are an issue with online experiments. However, I was not expecting the maximum delay to be almost 200% higher than the lowest if the study is run locally with browser, server, OS, etc. all the same (and nothing else running in background). That's a bummer :/
Studies seem to suggest that they are more reliable (cf. URL: peerj.com/articles/1058/ )
Thanks again,
Daniel

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 2 semaines #180001 par holch
I am not an expert on this, so wait until some more experienced users (like Tpartner & Denis) will skip in. However, in my experience, everything that has to do with web technology is generally always quite "unstable" when it comes to response time. You still have http and database connections, that might take sometimes longer, some times are quicker.

I guess when experiments are time sensitive, they should not be dependend on database connections, etc. You might be able to run them purely in the browser (e.g. via Javascript, Flash, etc) and only at the end transfer the timings etc. This is probably better when treating time sensitive experiments. But even then, with Javascript, at least a couple of years ago there were significant differences between the speek of how browsers rendered javascript code (if I remember right, Chrome was especially quick in doing so), etc.

Of course, with the same machine and browser, etc. this shouldn't impact too much, but I just wouldn't want to rely on it.

I'm not a LimeSurvey GmbH member. I answer at the LimeSurvey forum in my spare time. No support via private message.
Some helpful links: Manual (EN) | Question Types | Workarounds

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180011 par DenisChenu
For showing a page during X seconds,

I think i use a
- Hide page during load
- Use learn.jquery.com/using-jquery-core/document-ready/ to show the page and start the counter
- Click next (and hide the page) when counter get 0


For body loading/loaded see
javascript : gitlab.com/SondagesPro/SurveyThemes/skel...s/skelvanilla.js#L85
css : gitlab.com/SondagesPro/SurveyThemes/skel.../skelvanilla.css#L31
html : gitlab.com/SondagesPro/SurveyThemes/skel...yout_global.twig#L25

Assistance on LimeSurvey forum and LimeSurvey core development are on my free time.
I'm not a LimeSurvey GmbH member, professional service on demand .
An error happen ? Before make a new topic : remind the Debug mode .

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180017 par ultimatte
Hi, thank you for your reply. However, I'm sorry, I can't follow.
Do you suggest that I should not use the timer provided by Limesurvey but jquery?
Unfortunately, I'm not familiar with jquery. I have to write the code in the question's source code, right?
I would appreciate a lot if you could elaborate.
Thank you so much

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180018 par ultimatte
I couldn't find an option in Limesurvey to 'hide the during loading' :/ (only 'hide question'?)

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180020 par DenisChenu

ultimatte écrit: …
Do you suggest that I should not use the timer provided by Limesurvey but jquery?

Sorry, i didn't know you use LimeSurvey timer :).

Maybe you can test importing skelvanilla theme and test with it.

gitlab.com/SondagesPro/SurveyThemes/skelvanilla
dl.sondages.pro/skelvanilla.zip

Assistance on LimeSurvey forum and LimeSurvey core development are on my free time.
I'm not a LimeSurvey GmbH member, professional service on demand .
An error happen ? Before make a new topic : remind the Debug mode .
Les utilisateur(s) suivant ont remercié: ultimatte

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180022 par ultimatte
Thank you!
I tested it. Unfortunately, I still get a considerable variance in page load times with your theme although your theme seems slightly better: 3.78-4.82s with Vanilla; 4.2-5.2s with Skelvanilla.
Please, do you have another suggestion?

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180023 par ultimatte
Apart from this, your theme looks great and I would like to use it. However, I need the grey background from Vanilla as I present images with white background. Could you please tell me how to get the grey background back in your theme?
Thanks a lot

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180028 par DenisChenu
Loook at theme option : you have whole boostswatch included.
And here : you can choose "well" for question wrapper.

And you can make your own theme with gitlab.com/SondagesPro/SurveyThemes/less-skelvanilla (but need css/less compentency)

Assistance on LimeSurvey forum and LimeSurvey core development are on my free time.
I'm not a LimeSurvey GmbH member, professional service on demand .
An error happen ? Before make a new topic : remind the Debug mode .

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180044 par ultimatte
Thank you! That worked <3

Do you or someone else maybe have another suggestion for the response time measurement?

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 1 semaine #180046 par DenisChenu
Maybe this need to be reported : response time must be done after HTML is shown ?

Didn't know …

Assistance on LimeSurvey forum and LimeSurvey core development are on my free time.
I'm not a LimeSurvey GmbH member, professional service on demand .
An error happen ? Before make a new topic : remind the Debug mode .

Connexion ou Créer un compte pour participer à la conversation.

Commencez dès maintenant !

Créez simplement un compte et commencez à utiliser LimeSurvey dès aujourd'hui.

Inscrivez-vous maintenant

Inscrivez-vous à notre Newsletter!