clickthroughrate — 2011-10-15T20:43:24-04:00 — #1
A tip to survey creators:
Please display all the questions on one page: formulate the questions accordingly.
It is very tiresome to answer 5 questions, click to proceed, 5 questions, click, 5, click...
It does not make you want to finish the survey.
Thoughts on this?
ralphm — 2011-10-15T22:34:11-04:00 — #2
You don't necessarily want too many questions on one page. I just like to know, in advance, how many questions to expect. I don't like knowing how many there will be. (Either state at the beginning that there are 25 questions, 5 per page, or on each page indicate how many to go.
clickthroughrate — 2011-10-16T15:13:18-04:00 — #3
So you are ok with repeatedly clicking for next page? What is your limit?
I do agree that in this case, they should state how many questions there are. Nothing worse than not knowing you are answering 10 pages of questions.
ralphm — 2011-10-16T17:56:37-04:00 — #4
That doesn't bother me, to be honest.
What is your limit?
I guess it depends on the time I have. I appreciate it when I'm told "this survey will take about 15 minutes to complete", or "this survey contains 15 questions". Then I feel I know where I stand, and I can wait until I've got some time available. (It's an excuse to get a cup of tea!)
What really annoys me is when new pages keep appearing and I have no idea how many more there will be, or how long the survey will take. Sometimes I just abort in these circumstances.
system — 2011-10-18T10:27:00-04:00 — #5
I've got to go with the "tell me up front how many" -- and to be frank do NOT try to tell me in minutes, tell me how many questions... Also, I really HATE the one question per page idiocy just as much as I hate all the news websites out there which have so many ads you can't find the article and then break the article up into 1k slices over 20 pages. "Read one paragraph, oop, time to load the next page".
... just be sure you don't make a page too long -- there should be a nice happy middle-ground like say... five questions per page.
clickthroughrate — 2011-10-19T20:01:44-04:00 — #6
Cant argue with that. Usually cancel the whole survey after three pages of one questions. Only five questions per page?
technobear — 2011-10-20T09:35:05-04:00 — #7
I'm in full agreement with ds60 on this - especially
do NOT try to tell me in minutes, tell me how many questions...
I'd say number of questions per page depends a bit on how wordy or complex they are. More than five is fine for very simple questions.
eruna — 2011-11-10T09:51:20-05:00 — #8
One benefit of having a multi-page survey is that you can capture partial data. If it was all on one page, then you would lose data if the survey was abandoned part way through. Though, AJAX can also be employed to capture fields after they have been completed.
It seems like five-ten questions per page is a good amount. But I agree, there should always be an indicator of how far along you are in the process. It is common with progress bars for the graphic to be disproportionately slower in the beginning and faster towards the end. This would probably help with conversions.
stevie_d — 2011-11-10T13:35:14-05:00 — #9
I would really not recommend submitting data before the survey is complete. Not only is it a technical and logistical nightmare (eg, what happens if someone clicks 'back' and changes one of their answers?) but it is potentially in breach of data protection rules, and having partial data may not be helpful anyway.
eruna — 2011-11-10T21:09:35-05:00 — #10
Its not really all that complicated. There are several methods of doing this. One method is to store a temporary id in a session and use that to track the responses. There are many different types of surveys. There are strict scientific surveys, but there are many others that are more informal where any information is helpful. I'm speaking here from my own experience of gathering member responses.