Hmmm, I am not sure if the Qualtrics feature would do the trick. Depends a little how those test responses are generated. If they are generated directly in the database, then they might not uncover any routing problems. I had only a quick read and I couldn't figure out how those test responses are generated.

For testing screening, routing, branching, jumping (yeah, I know...), conditions, etc., I think you need some kind of robot that goes through your questionnaire and fills it randomly, I guess.

With a solution using Random + relevance + validateSurvey (can be done in plugin).
You can have finally : the group where it's invalid + (unsure on this) the validation error. Or if survey seems valid …

"Show verbose output: Show more details concerning each test iteration, including which questions and blocks were tested, which order the questions were displayed in, and whether randomization was applied."

OK, this part indicates somehow, that the test feature really goes through the questionnaire and doesn't just create answers in the database.

holch wrote: OK, this part indicates somehow, that the test feature really goes through the questionnaire and doesn't just create answers in the database.

Questback has a similar toolsi nce the beginning. It's not just filling the database. But i don't see LimeSurvey codebase ready for such a tool now. Mantis is full with issues more basic.

Even if you sponsor for X EUR, the feature won't come. It's like decorating the penthouse, while the groundfloor is not finished.

I don't think it was available in Questback back in my days, when it was still called Globalpark. At least I have never seen it or I have never heard of the programming team using it.

holch wrote: back in my days, when it was still called Globalpark.

In the German version called "Projekttest" and at least available since 2004.

You got a nice routing statistic with percentage usage. Would help in LimeSurvey a lot, when you have random parts and might have forgotten to update ExpressionScript to distribute to all questions.

OK, then it existed back in my days, I just didn't run across it. Wasn't my direct department anyway. Was just an internal client...

A generic tool to test webapplications with focus on UX is e.g. en.wikipedia.org/wiki/Selenium_(software)

You can record via a browser extension
addons.mozilla.org/en-US/firefox/addon/selenium-ide/
chrome.google.com/webstore/detail/seleni...kldjjndioackbalphokd

It's still a lot of work, which might be better invested in just filling out the survey.
But if parts of a survey are reused, it might be worth a try.

See github.com/LimeSurvey/LimeSurvey/tree/master/tests

DenisChenu wrote: See github.com/LimeSurvey/LimeSurvey/tree/master/tests

How will that help to fill out a survey 100 times to test e.g. for relevance bugs?

jelo wrote:

DenisChenu wrote: See github.com/LimeSurvey/LimeSurvey/tree/master/tests

How will that help to fill out a survey 100 times to test e.g. for relevance bugs?

You speak for Selenium : test of LimeSurvey core functionnality use seleium too …

I mean bugs inside the survey (not PHP/JS bugs). E.g. having three questions groups but have a random equation which only generates 1 or 2. The survey designer forgot to change the function. When you run 100 respondents you see that non answered the third question group. That kind of tests would be easier with the selenium core support by LS?