Hi All,

I am running a survey where respondents (undergraduate students) have to select what room they took for a course. The way I have this set up is that the respondents first select the subject they took (89 subjects total), then the course they took (~15 courses per subject, 1400 courses total), then the room that they took the course in (since many courses run multiple classes). The way that I have this working is by creating a separate question for each course to give the user a list of allowable possible rooms and using conditional logic to display the right question. (I wrote a C++ program to translate schedule database entries into questions in the tab-delimited format and then upload them.) The problem is, since courses could be scheduled in multiple rooms, I had to create multiple copies of each room question, so I am currently sitting at 3097 questions. Since whatever database we are using (I think it is mySQL) is limited to like 1000 columns, when I go to activate the survey it is deeply unhappy. So I need to reduce the number of variables that are stored. I would really rather not break the survey into multiple parts, but If that is the only way to fix this, so be it. However, before I do that:

1. Is it possible to use relevance equations (or something similar) at the answer level instead of at the question level? That would trim my survey down to like 500 questions.
2. Is it possible to somehow tell LimeSurvey NOT to store certain variables in the database?
3. How exactly would I go about conditionally changing the end URL to direct the respondent to whichever of the surveys applies to them?
4. Is there any better way to do what I am doing?

I have attached a zip file with the .txt of the survey. It works in preview mode. It takes about 10 minutes to parse on my server (running 2.06+ build 151205), but it does parse successfully. (Should probably have a warning when you try to create more than the allowable number of questions?)

Hope that someone can help me on this!

Yours,

Chris Litfin
McMaster Univerisity
Hamilton, Ontario, Canada

I think you're going to have to use multiple surveys. The first survey could ask the Subject and then redirect to one of 89 "secondary" surveys.

The easiest way to handle the redirection would be to use the "secondary" survey IDs as answer codes in the Subject question. Then your end URL would be something like this (assuming the Subject question has code "subject"):