Regarding the first question, you need to use either the SGQA OR the question code. You have mixed them together.


[/code]

Hi,
this is very helpful...
Basically we have distinct subgroups. We want to present them the same questionaire and distinguishing between different subgroups of participants in the background. So they should not see which group they belong to.

My idea was to create a questionaire-url that has a parameter which preselects a value in a question that refers to the subgroup. With the parameter I mentioned in the first posting it already works (thanks for your help btw!). The prob is, that if we hide the question in the lime survey backend it never gets any values.

Can you show me a way how to assign the participants to a group invisibly to them ?

Thanks!

Best,
Svenja

I don't understand : how you know group ?
If you know group when create token : why not use Token attribute ?

Hi,

we know the group when create token. But we want to offer a similar link to everyone like for instance:
example.org/index.php/123456?token=moinmoin&lang=de

But this would lead to resultset, where each line would represent a participation trial. How can we know, to which group we can assign each trial ?

Sorry for the questions we are right at the beginning... Maybe we need a different way...

The prob is, that if we hide the question in the lime survey backend it never gets any values.

Hiding the question should not matter. What LS version are you using? How are you hiding the question?

svenjalell wrote:

Hi,

we know the group when create token. But we want to offer a similar link to everyone like for instance:
example.org/index.php/123456?token=moinmoin&lang=de

 

I still don't understand ?

You want to use token or not ? You want to use same token for all user ?

And again : Why not use participant attribute ?
 

Hi,
yes we want to use tokens. Each user receives his own token.

... not sure how to deal with the participant attribute and where to find it ... sorry...

Hi,
... managed to hide a question via the custom.css of the theme... works for me.

www.limesurvey.org/manual/Survey_participants#Manage_attributes
www.limesurvey.org/manual/Expression_Man...#Access_to_Variables

TOKEN:ATTRIBUTE_1 can allow you to choose group for each user for example.

Hi Denis,
thx a lot...
... came across this also... but this does not work for anonymous polls? Is it possible to put the to put the token in the result matrix ? So with the token added in each line you may see which group the participant belongs ?

Best,
Svenja

OK,
Right : no way with anonymous survey.

You can use it in email (to add the Q01={ATTRIBUTE_1}) for example.

ok thanks ...
... so it isn't possible to write the token somewhere in the resultset of a participation ? say, as the value of a hidden text field ?

Best,
Svenja

No, this would break anonymity. You need to choose: either you want an anonymous survey (as technically defined in Limesurvey) or you want to be able to pass attributes.

Because even if you would only use attributes that can not identify the respondent, if Limesurvey would offer the option, then the anonymous mode couldn't be considered anonymous anymore.

Even passing attributes/variables via the link is dodgy (in my opinion), when talking about "technically anonymous" surveys.

What we often do in market research is to promise that the data will be analyzed anonymously, while there is still an option to connect respondents' data with responses (e.g. for inviting them to a follow-up research project, to provide incentives/raffles, quality control, etc.).

In this case the anonymity is not based on technology, but rather based on research ethics. We promise to not analyze data together with personal information. After a while the personal information is disconnected from research data and later also deleted, when we are sure we don't need it anymore.

Now if you want or need to guarantee 100% anonymity based on technology, then you can't pass data from the token table to the response table or the survey, but I also think you shouldn't pass any variables via the survey link. In your case the data you pass might not break anonymity, but respondents don't know this. The variable you pass on could stand for anything (e.g. be an ID), etc.

Even passing attributes/variables via the link is doggy in my opinion, when talking about "technically anonymous" surveys.

Yes +1

What we do in market research often is to promise that the data will be analyzed anonymously, while there is still an option to connect respondents data with responses (e.g. for inviting them to a followup research project, to provide incentives/raffles, quality control, etc.). 
hhh

So !