Hello we are evaluating using LimeSurvey for both online and paper surveys. I think this is possible with LimeSurvey but I need to get some feedback from users who have used it with Paper surveys and perhaps some paid tech support for implementing this.

Please let me know if you have experience with this and/ or can support us in implementing this.

Thanks!

Maybe you can outline your requirements in detail?

Do you just need some data entry possibility for paper surveys and the option to create a printable version of the online survey?
In this case it should all be there at Limesurvey.

Hi Mazi,

We want to be able create paper surveys in LimeSurvey.

Have the paper surveys filled out.

scan the paper surveys so that the data is transferred to LimeSurvey.

You want this:

quexf.sourceforge.net/

Which integrates very nicely with LimeSurvey.

The UI is a bit quirky, and I haven't been able to turned a scanned survey into data, BUT I can see that's it's close, and either some bugs will be fixed or I'll figure out what I was doing wrong. But anyway, this is the exact package you're looking for.

Please let us know how it works out for you!

Tac

Oh, I should mentioned that the step before QueXF is QueXML:

quexml.sourceforge.net/

LimeSurvey exports to QueXML, which you can then import (including the banding information, basically the coordinates of the boxes) into QueXF.

Tac

Thanks for your feedback, Tac.

Petrus, once you have successfully set this up maybe you can report back and write a short tutorial for other users having the same requirements!?

Mazi wrote: Petrus, once you have successfully set this up maybe you can report back and write a short tutorial for other users having the same requirements!?

Oh yes !

I think it can be a great contribution to our documentation !

Has anything come of this? I'd love to here how the testing turned out. We're looking at setting up LimeSurvey to replace our SurveyMonkey account and used this sort of solution to administer and process a 20,000 user paper based survey. We're a school district and about 50% of our families don't have Internet access. This is something we could never look at doing with SM.

Thanks!

There are two issues: the technology and the workflow.

First, code your survey into LS, and then export it as a QueXML file. You'll want the survey to have very few open-ended (text) questions, since they'll probably need to be coded by hand (the OCR for handwriting doesn't work).

Then you'll want to install QueXML, and make sure that the printer and scanner you're using work for recording the answers. If you're happy with that, you need to evaluate if the propose QueXML workflow works for your 20,000 responses.

If the technology works, you can hire someone to re-do the workflow.

When I tested this 2 years ago, I couldn't get the accuracy high enough, and the workflow drove me crazy. But if you have an IT person that can spend a few hours to a couple of days configuring this, it might work for you.

First (and easiest) step is coding up the survey and printing it out in the QueXF PDF file. If nothing else, it looks very sharp.

The other approach, which might sound crazy when you first hear it, is to "Turk" the paper survey into LS (or SM). Amazon Turk ( www.mturk.com ) is an amazing tool for doing things that humans do best, like reading handwriting and transcribing survey results. We're using it now for image tagging, although it's not yet integrated into LS, something I intend to do if I can find a client to fund the development.

Tac

Hey, thanks for the great and quick reply!

We were mainly interested in the OMR capability. Our current system doesn't handle hand written text so that was going to be a benefit if it even worked.

I don't know much about the workflow so could you describe this in a little more detail? We have a testing vendor that we scan Scantron type tests into that processed test scores. Our staff is fairly used to scanning the documents into a system that them processes them en masse.

One the form is created in LimeSurvey and exported as a queXML is scanning into and processing by queFX really a pain?

Thanks!

Maybe it's not that bad -- I didn't care for it, but I'm picky about some things. The way Limesurvey stores survey results in the database is like fingernails on a chalkboard to me, I cringe in pain when having to deal with it. The fact that survey_ids are non-sequential and lack 0's causes me heart palpitations. So I might not be a good judge.

The good thing is, it's easy to find out if it'll work for you, and the LS integration is pretty nice. So give it a try, and report back and let us know how well it work. If you're already used to the whole process with Scantron, this might be an easy migration.

Tac

Sounds good. I'll post how it goes.

I too despise poor database design (which some only think consists of table structure alone and could careless about names or IDs). We already work with a fair number, all fitting somewhere between a pleasure (only one of those) to down-right horrible.

I guess I'm just saying if the db design leaves much to be desired I won't be that surprised.

Thanks!