A survey can hold roughly 1,300-1,400 variables. We are planning to do a survey with the following parameters:

- 6 evaluation criteria
- 4 available choices for objects to evaluate
- 50 categories in total
- 20 different regions

In total I come up with 6*4*50*20 = 24,000 variables

This number of variables is extremely high and would take a lot of surveys to build this structure. I was thinking of reformatting the variables as such:

If I limit the number of categories I can choose to a maximum of 5 categories for example. Then I limit the number of evaluation criteria to 6 by 5 categories rather than 6 by 50. How would I format the survey so that the evaluation criteria only show up for the maximum 5 that are chosen? That would be 6*4*5*20 = 2,400 variables which is much easier and doable than the first idea written.

Hi,
but what is your question?

so that the evaluation criteria only show up for the maximum 5 that are chosen?

But there are still 50? You only display 5? That will not lower your amount of variables.

As long as we do not know your survey, we can only guess.
Maybe it is possible to use "micro-tayloring"?

- 20 different regions

Do the regions have totally different questions? Otherwise you have just one question for regions or it is already in the token attributes.

- 4 available choices for objects to evaluate

What is that? Single punch? Or what.

You see, a lot of questions.
So you really should show us a bit more of the structure. A rough overview, what are
"evaluation criteria"
"objects"
"categories"
"regions"

Joffm

My question is towards the end of the survey structure:

20 regions -> 50 Categories -> Up to 4 brands to Choose -> 6 Evaluation statements (evaluated from 1-10)
The evaluation statements are questions for example: Please evaluate the level of talent on a scale from 1-10 = 1 variable
I am trying to limit the number of variables in this structure. How it stands now, the respondent could technically answer up to 1,200 questions, since they could choose 4 brands for every category and would have to evaluate 6 statements (scale of 1-10) for each brand.
I can limit the number of variables by tying the number of evaluation statements to a fixed amount of categories. Rather than having 6 evaluations for every possible category scenario, I only apply this to 5 categories.
This should give me 6*4*5= 120 Variables
Would I need to have evaluation statements for every scenario for this to function? The idea is not to include evaluation statements for every category, but only for the ones chosen.
How else with this basic structure could I limit the number of variables in question?

Hi,

Would I need to have evaluation statements for every scenario for this to function? The idea is not to include evaluation statements for every category, but only for the ones chosen.

As far as I understand the questions are the same in each scenario, don't they?
And you say "up to 50 categories", but only the chosen ones.
How many can be chosen? All?

So in my opinion it is sufficient to create as many groups as the maximum of chosen categories and pipe the rest. Buzzword: Micro-Tayloring.

But without more input ...

This should give me 6*4*5= 120 Variables

So you really should send a sample (lss export) with a 2*2*2 approach, that we get a roughr idea of your design.

Joffm

Hi Joffm,

Micro-tayloring sounds to be the right term. I will create an example structure to get a better idea of how the LimeSurvey will be structured.

The idea is definitely to only have the standardized evaluation for the chosen/clicked answers. Therefore if you had 50 categories or 100 categories, only the 5 clicked would be affected, and those that are not do not flood the survey with unneeded variables.

Thank you for the input so far. I will get back with an example.