The Design of Software (CLOSED)

A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.

The "Design of Software" discussion group has been merged with the main Joel on Software discussion group.

The archives will remain online indefinitely.

Logging responses – question or questionnaire.

I am working on an interface where an operator will log responses from completed paper questionnaires.

What I am wondering is, what would work better:

Prompting the user with a question at a time for the response to be entered and then a “Next” button to advance to the next question


Displaying the whole questionnaire allowing the operator log responses at will.

Will the number of questionnaires to be logged make a difference to the optimal approach?

Am keen to hear your opinions and viewpoints on either approach.
Sunday, April 22, 2007
I'd recommend something in the middle. 3-5 questions per screen. This way, when the operator is transcribing entries from the paper version, they are less likely to skip/duplicate an entry, resulting in a spare/missing answer at the bottom to try to determine, and possibly a lot of rework.

If you're displaying the questions to subjects, displaying them all at once will raise the abandonment rate, as the user says something like "I'm not wasting my time on this" and clicking cancel. This is why almost all questionaires do one question per screen.
Sunday, April 22, 2007
If people "have" to fill it in, then do whatever works - they won't abandon it. That applies to employees and paid research subjects.

On the internet, keep it short and let people know the actual time commitment, otherwise you'll find people never get past question 9 of your 10 question survey because they get bored and think they'll be there all day.

While you're at it, read up on the statistical validity of self-selected survey participants.  ;)

Sunday, April 22, 2007
Thanks for your feedback so far. 

I should clarify, this is for a desktop application and it is for post event surveys. Participants will fill in paper surveys and their feedback will be analysed back at the shop.

The idea is to make the logging process - painful as it might be - as painless as possible.

Sunday, April 22, 2007
Why not run them through some automated process?  There are a lot of commercial packages and development libraries to build on.

You could even write one yourself from scratch.  I wrote a simplified 3 of 9 barcode recognizer in under a day last year, so looking for filled in bubbles should be a walk in the park.
JSmith Send private email
Sunday, April 22, 2007
Do you mean scanning completed questionnaires and "parsing" the scan?

Would appreciate any links on this. I was thinking scanning was not always reliable and sometime hand written comments could get lost.

Monday, April 23, 2007
I wouild say display the entire questionaire.  When doing a page, most data entry personnel hate to have to stop and wait for a page refresh or anything.  If it is a long questionnaire, see if you can auto scroll by tabbing, so the data entry person does not have to take his hands off of the keyboard any use the mouse.
Wayne M.
Monday, April 23, 2007
If I were just doing mark recognition, like fill-in-the-bubble answer sheets, I would write it from scratch.  For handwriting recognition, I would use AnyDoc ( ) as a packaged product or LeadTools ( ) libraries to build it myself.

AnyDoc has all the nice GUI features already in it to allow the user to verify the results, LeadTools allows better integration into your solution.
JSmith Send private email
Monday, April 23, 2007
Thanks for the links.

The LeadTools link appeals the most but the runtime licensing will be a real pain to administer it will also require revising the price point of the software or else making it an optional add-in. These are chunky decisions to make.

Thanks for your help.
Tuesday, April 24, 2007

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
Powered by FogBugz