Hi everyone,
I am a LimeSurvey starter and would very much appreciate your support and advice, here we go:

I have a question section that consits of three parts:

(1) Matrix with multiple questions and each question can be assessed by "rather unimportant"; "rather important"; "Important"; "very important".

(2) Follow-up question that is asked openly so respondents can add own values which can be assed  (assessment only taking place if values are added by respondent).

(3) If values are added => assessment of respondent's values by "rather unimportant"; "rather important"; "Important"; "very important".

And here is where im stuck:

(4) Goal of this questions is to have a top 5 ranking of all questions and respondent's values (from (2)) that have been prior assessed with "very important".

I have tried it with conditional coding and have also looked up other topics in the forum dealing with dynamic lists and ranking. I found a similar entry , but the JavaScript idea is not working for me. I use LS version 3.22.2+200204.

Happy on any advice, thanks!

Please, send a lss export of these questions that we see how you implemented (2
In my opinion this all can be done with Expression Manager.
Will be the usual way to insert a hidden multiple question which will be used as array filter for the ranking.

Joffm

So, you want to have something like this?
I put your points (1) - (3) into one question.





Joffm

Thanks a lot! Actually an interesting idea. I will think about the way suggested by you. But currently it looks like in the lss export file: questions A2b - A5 where A5 is the question asking for a ranking. Currently a ranking of all items asked before is implemented. But I would like to limit the ranking only on the items assessed by "very relevant".

Kind regards,
zepmaya

 

Hi,
first some remarks.
You have SQ016 and SQ017 in all four questions, but with different options.
And the items of the ranking question seem to be different from the others.

Your answer options of the array are more or less a scale.
So you should use numeric codes. Later you want to do some statistical tests, not only frequencies or  CHI2

In my opinion A4 shouldn't be randomized. The participant will be confused if the items are displayed in different order than A3

Well.
To be able to show something

• I recoded everything
• Used the same items in A2-A4 and the ranking.
• Made the display in the ranking question dynamically. ("Please rank at least the five most" sounds a bit silly if there are only three items "very important")

Have a look at the group "Calc" there is all you have to do.

After you understood the method you may recode it again.



Joffm

And you shouldn't use this "condition builder". It is really outdated. Now we use Expression Manger. It is easier and much more flexible.

And here a shorter prototype with some explanation.
The main thing is the hidden inserted multiple question that is used as array filter of the ranking question..
In this hidden multiple questions the relevant items are "checked" by a floowing eqation.

 

Joffm