Our business requires us to do show/hide of questions in questionnaires. We do this through DDE's since question filtering is not appropriate. However, once we go down this road, we can no longer rely on Archer's built-in questionnaire scoring.
Anyone who's worked with Archer questionnaires and attempted to do show/hide of questions via ACL DDE's and needs to do scoring knows what I'm talking about. We've typically handled scoring in such cases via calculated fields but the number and complexity quickly get out of hand. Does anyone have another approach to make this easier?
- Community Thread
- Forum Thread
- questionnaire scoring
- RSA Archer
- RSA Archer Suite
- Solutions and Use Cases
Is what we have done. In some cases it isn't too bad such as when questions are stable. On the other hand when want to add or remove questions it quickly becomes a big deal.
You either do calculated field or you can use some custom object to get score and put to the filed upon user saving. However, mind that custom object would not work in case of scheduled recalculation or datafeed.
I do not think, there are any example, but it is not build with good JS skills either.
I've used the DDE hide/show in the past as well. To handle the scoring issue, I add a default N/A value with a score of 0.00 that is correct to any question that is hidden. That allows your hidden questions to still be scored and also added into the progress completed calculations. Also I only use non-questionnaire text fields as otherwise the text fields really mess up scoring/progress.
Thank you very much for replying to my post about questionnaire scoring. We used to go the default N/A value as well but moved away from it because we had to use elaborate scoring calcs to determine if all questions were answered with a non-N/A response. For a while now we’ve eliminated a default N/A (in our case we used “**Not Answered**”, but no difference in essence, just a way to bring the user’s attention to unanswered questions) and made all questions displayed on the layout as required via DDE’s once Submission Status changed to “Submitted”. This allowed the tool’s built-in response to unanswered required questions to kick in and let the user know specifically which questions lacked answers. When you have a few hundred questions, as we typically do, this avoids the users trying to hunt for unanswered questions like a needle in a haystack. So in your case how do you determine if the questionnaire has been fully answered?
Thank you very much for your comment. I’m not understanding it, however. We do use hide/show via DDE’s. Is that what you’re referring to or are you referring to filtering of questions? If so, we’re wondering about that as well. And, if so, does Archer’s scoring take into account only those questions on the layout that are there via filtering and not the others?
Our typical scenario where we use ACL DDE’s to display/hide questions is where one answer to a question that always displays triggers (via DDE) several sub-questions to display below it. For example, answering “Yes” to question 10 causes questions 10.a – 10.f to display underneath it. We’ve had instances where our users answer “Yes” to question 10, go ahead and answer some or all of questions 10.a – 10.f, and then go back and change the answer to question 10 to “No”, causing the sub-questions to no longer display. However, Archer still records (and scores) the answers to them as if they were still displaying. Our business does not want them to be scored if they’re not displaying. This is the crux of what we’re dealing with and what I’d like feedback on if you wouldn’t mind.
DDE's and Hide/Show Rules in the questionnaires are 2 different things.
However, I now understand your issue with the extra detail you've provided and indeed n a situation where you have dependency questions, this is unrelated tot he hide/show rules.
I've done the same as David Merrick mentioned above.
Another idea -
Make it hard for them to change their mind on an answer.
So, using DDE's don't allow the dependent questions to be hidden if their value is something other than no selection.
Basically, what we want to do here is is make it apparent to the end user that they didn't deselect the answer.
Also, thinking about filtering out the parent question's answer if any of the children are populated.
This way the user can't change the parent answer unless they "deselect" all the children.
Or make the parent question read only when children are not no selection or n/a.
I'll keep thinking on other things we may have one.