The question has been raised whether it's really required of the survey system to guarantee anonymity of the users taking these surveys. Even if the IP addresses of users are saved or any of the access control values are saved, there is no way to tie them back to the actual answers that person gave in the systems current design (and planned design). You'll be able to know that a certain person did take the survey and when they did, but not their actual answers (unless you actually ask for identifying information within the survey).

I'm very inclined to leave it this way, but interested in hearing your, the users, thoughts. When this system was originally designed for administering command climate assessments for the U.S. Army, anonymity was a must. For some of the surveys you're using, though, I can see how it's not as much of a requirement.

What I really don't want to see happening, though, is people taking the survey lulled into a false sense of anonymity when there really isn't any. I can't guarentee that the people designing he surveys will adequately give a notice to those people taking the survey whether the results are anonymous or not.

Those are my rambling thoughts... what are yours?

---John Holmes...

I must admit that I might have been a bit to fast when asking for the feature.

It was newer my intention to 'go behind the back' of the users. So why let the last question be, something like this; "so far you have been 100% anonymous. However, we would like to ask you the same questions that you have just answered, again in a years time. If you'll let us send you another questionnaire at that time, please provide your email below. - Note that this way it will be possible for the administrator to see what you have answered. However, we guarantee that this information will remain strictly confidential."

- I guess what i'm saying is; I'm amazed to find out how flexible your system really is - all I got to do is to learn how to think before I ask questions.

Heh... I wasn't referring to you going behind user's backs...just that it was possible or could be possible.

You can do what you describe now by simply gathering the email address in an S-answer type (small text box). Run the next survey and just make sure they use the same email address. Then you can export to CSV file and "join" the two surveys on that email question.

There's not much of an issue with that because the users should realize that they're providing an email address along with the answers so they can be related. I'd still give a good notice saying that it's possible just so there are no questions, though, kind of like you mention.

Anyone else have any opinions on this issue?

---John Holmes...

I'm using UCCASS on intranet. Users should pass authentification before taking the survey. As they pass auth, I already know who he's taking the survey, there is no anonymity and it's better because each user reports budget !

For each validated survey, I record the user and the sequence in a specific table.
(it means less code than record the user for each answer...)

For each quitted survey before validation, I temporary record the $_SESSION in another table with the userid, so it's possible for a user to quit the survey and come back later at the same page he precedently quit.

I was wondering; could one way around the dilemma simply be to publish two different versions of UCCASS? One of them could in the copyright footer contain some text a la "anonymity guaranteed" - maybe together with a "read more" link to this page explaining that it off cause is a possibility for a php programmer to alter the code etc.

I know that as long as the open license is in play it will be difficult to actually guarantee anything at all, but so it is today as well.

sincerely yours
Andreas

I'd rather not have to keep up with two versions, although I suppose it would not be that difficult. I'll take that into consideration.

---John Holmes...

Kick Butt program BTW!!! I love it.

I would like to be able to compare results of users. Here's the theory:

Jane Doe takes a survey

Jane Doe's company/group/whatever does things to change Jane Doe's answers on the survey.

Jane Doe takes survey again

Now I can go in and say 31% of the respondants say things have gotten worse, 4% remain the same, 60% say it is better 5% did not retake survey.

Perhaps collecting Jane's name and doing and MD5 encryption on it (I think the feature is built in to PHP) would allow her name to be stored with the answers she provided, but the admin would still not be able to say "This is Jane's responce"

Of course, my PHP knowledge is very limmited, but that might be a middle of the road feature.

James
PS Pardon my spelling

I think this should be an option for the survey author, and the responders must be aware if the survey was anonymous or not. Maybe the welcome/intro to the survey should indicate that.

So when the author clicks "Create Survey", a radio button should appear with yes/no fields under the title "Anonymous Survey?". This can be placed under "Survey name". Once the survey is created, this option should not be changable. Then on the take_survey template, there should be a highlighted fieldd stating "This surver IS (or IS NOT) anonymous, so the responders are fully aware of their privacy options.

That's what I'll do. You'll be able to switch from "Anonymous" to "Non Anonymous" and the program will wipe the relation data, but you can't go the other way once the survey is made.

---John Holmes...

I'd also like to modify my vote/recommendation

I think it should be both, a configuration option in survey.php.ini to allow/disallow the option, and, if allowed in survey.ini.php, the survey author with survey creation priviliges can select the anonymous yes/no option when creating the survey.

The reason I think it may be important to make it a site-wide allow/disallow option first, is that some sites (mostly corporate and governemnt) have extensive privacy policies, and may want to allow members to create surveys, but do not want to allow them to select the non anonymous option, as it may presnt them with privacy policy concerns... Other sites, like chit-chat forums may not really care, and the survey contents may be trivial.

That's a good idea. I can make it similar to the "survey_text_mode" and "user_text_mode" options... where you set a "maximum" value in the survey.ini.php file, and then set the specific value on a per-survey basis.

I think I'll make an extra template that'll be included in "not anonymous" surveys that'll serve as a warning/notice to those taking it. A programmer could remove the notice, of course, but at least I'll be making the effort to notify the users.

---John Holmes...

I am glad this came up again.

I have just completed a survey with ca. 300 respondents. After completion of the survey I realise that I need to ask one more simple question. However because of annonymity it is impossible to sendt out an invitation for a one-question survey, and then add the answers to my existing data.

This functionality would be good to think in to the solution you make.

(if I am wrong and this IS possible - pls tell me how)

No, it's not possible right now. Once the anonymity choice and save feature are added, though, it would be simple to do so. I'll need a way to "continue last survey" even if it was already completed, though. Then you could just add your extra question and have them retake the survey. The answers already provided would be there...

The only issue is, if they've take the survey more than once, then what one do they continue? Do I give them a date choice? Do I just allow them to continue/update the last completed survey? Do I only allow this feature if they can only answer the survey once? What is most useful to you guys?

EDIT: Just so you know, I'd lean towards the second option of only continueing/updating the last survey...

---John Holmes...

So an admin option would be to

For current survey being edited:
( ) Do not allow users to continue/edit surveys (no save feature)
( ) Allow users to continue/edit last saved/completed survey within [___] [minutes/hours/days] only
( ) Allow users to continue/edit any saved/completed survey (choose which one according to date/time stamp) within [___] [minutes/hours/days]

The "within" feature would let you set a time limit on when the user can come back to edit/complete the survey. So you basically say they have "10 days" after saving or submitting the survey to come back and continue or edit (for example).

And if you're only allowing them to take the survey once, then there's really no difference between the second and third option.

I think that should cover all the bases, no? He who wants "maximum control"...

---John Holmes...

Love this script!

I'd like to see the option to resume a poll as described.