Support Forums : How can I export values from a Hidden matrix table

Hi,

I have created surveys which I use scripting commands such as ’sgapiSetValue’ to set an array of string values to a hidden matrix questions (one column/multiple rows, each row allowing open-ended). For example,

Price (Matrix with one column allowing open-ended text)
price1 (open-box) - question id 64
price2 (open-box) - question id 65
…

Then I piped these string values for 64/65/… in the following questions’ text part. It is working fine. In those questions’ text part, correct hidden values from 64/65… are shown. So that is not my problem.

My trouble is with data exporting, although I can clearly see this item “Price” shown as “table” in data selection field, but I cannot see it in exported csv or Excel data file. I know the data must be in system’s database otherwise the later questions will not show correctly. But how can I get these data stored in “Price” as hidden values exported to a data file?

I am on Professional version, will Enterprise version be different on this or this is something currently cannot be done? It is quite important to be able to export these values, because they are the conditions for the responses for later questions, so definitely they are part of analysis.

Thanks

Scott

Hi Mario,

Yes, that’s correct, and other surveys we do are also similar.

So in this case, we need to export both choice answers for scenario 1 to 8, we also need to export those string values from the hidden matrix.

Thanks

Scott

Hi Mario,

I have tested the way you suggested, I cannot export the data yet. However, you mentioned that in SurveyGizmo2 this will not be allowed has raised lot more concerns to me. Because even if the hidden values only stored for the session and I can’t export them, I still can merge data with my pre-designed matrix file. So even I don’t have a way to export data for hidden values, as long as I know they match with my pre-designed file, I still have alternative methods.

But if I am unable to do in the way I am using now, it will raise lot more problems which means this type of work may not be doable. Pre-population via URL and login is probably one of the least preferred options for researchers. It will be important to assign values within a respondent’s survey not between respondents’ surveys, only based on a version/profile number given randomly by system at the point of survey, not pre-assigned to sample. We do not know what biases it is going to bring to the results especially if sample is out out your control (like from an online panel company).

Using many hidden value fields also does not sound like an option, because at the moment I can efficiently assign lots of value following different conditions into one matrix question only, if I use hidden value field, how many hidden value questions do I have to use? Given the complex patterns that survey needs to follow, it is going to make this kind of task impossible.

I definitely hope the current way will still work in new version even data only stored for the session. That is also the reason one of my clients just sign up an enterprise version to handle this type of work, and several other researchers are also considering.

I hope this can be regarded as an important feature that SurveyGizmo can handle and not a feature should be dropped. Many thanks

Scott

Hi Mario,

Thanks a lot for your help!  The suggestion you made to developers was exactly what I would like to have, which is to allow each type of question to have a hidden property. Looking forward to hear more on this development.

Scott