Hello, here is an issue. I have a survey with lots of conditions. Entire groups with 30-50 questions can be hidden - based on previous answers. For instance: first a question "are you currently working? [Y/N]" and then 30 work related questions (all conditional).

One would expect all the columns in the hidden section to be blank. If someone has no job then all data in the "jobs questions" should be blank. But if there are any checkbox based questions in the "work section", they are ALWAYS mapped to many, many NO's recorded in the database.

This is very confusing for anyone who is to analyse such data. Any particular reason for such design?

Thanks,
Luke

Hey!

Nobody wants to start a discussion?

Hi, Luke. I would like to start a discussion.

I'd say, if a question hasn't been shown, it would be good to have something like "na" or something. On the other hand, no is better than yes, isn't it?

However, it should be a problem for analysis, because those question should be analysed only for those, that have seen the question, otherwise the base of responses is wrong anyway.

Or why not just have blanks for all unchecked checkboxes? A blank means no action and that's exactly what the user did - he took no action by not having checked checkbox fields!

Well, if you just have blanks for all unchecked boxes, you have the same problem, right? You can't tell whether the person has seen the checkbox and didn't check it or whether the person didn't see the checkbox and thus couldn't check it.

IMHO there should be blanks in the database if the question was hidden.
It might be just an export issue. Please open a bug report.
Thank you!

Please post the link to the bugtracker here afterwards.