Hi everybody

I received a link to fill a survey.
There are two parameters with the link : token and lang

Does it mean that this survey is NOT anonymous ?

[In the manual, it is written in the "Notification & data management" part : "Enter a Token code (only possible if your survey is not anonymous)"]

Please, forgive my poor english...

"lang" is only for presenting the survey in the right language to you. So this will not have an impact on anonymity, as long as there is more than one person with this language, of course.

Now the token is a different aspect. The token is usually unique to one respondent. However, here is the passage about tokens and anonymity:

www.limesurvey.org/manual/Tokens#Can_a_s...onymous_responses.3F

So the fact that there is a token, doesn't mean the survey can not be anonymous.

thank you holch.

Well, is there any possibility, for a user, to check if a survey is anonymous, or not ?

I've been told that the survey I have to fill is anonymous, but I doubt.
And I think that my answers may differ if it is anonymous or not.

As far as I know, there is no way to see as a user if the survey is really anonymous. There might be one option: Limesurvey has a privacy message that it shows usually when the survey is anonymous. But, even this could be probably "faked".

However, if you don't trust the creator of the survey that guarantees your privacy, then rather assume it is not.

Often in market research the anonymity is guaranteed rather through the ethics code than technically. Often we need to know if someone has respondet (e.g. to be able to remind them or to give them the promised incentive), but we still guarantee to treat responses anonymously. We could for example attribute the responses to a certain person, but we won't as for our work ethics.

So it really depends if you trust the sender or not.

well it's clear now, thank you.
As the anonymity is conditional (it depends on the ethics of the administrator), I think that this kind of survey should not be said "anonymous" : we do not know what is the ethics of the administrator, it's not fair to suggest that the administrator is ethics by default because he/she may not be ethics, and, in such a case, the user is tricked.
We should invent a new term to precisely name this property. I suggest (sorry, english is not my mother tongue) : :woohoo: anonymous-if, or quasi-anonymous.
It is important to respect the user and to give him the proper and accurate information because his/her security is at stake : the user may not ask exactly the same way if a survey is anonymous or quasi- anonymous.

Do you know how to enter in discussion with the limesurvey community about this ethical problem?

What you are asking for is a sweet dream... If you are asked to participate to a survey and think that it may be sent by a liar and dishonest person or organisation, just don't answer it. Also remember that all core texts, alerts, messages etc. are editable either server or client side (nothing written in stone)

I agree with Ben, I see this as a relationship issue between the respondent and surveyor. I don't see how it relates to LimeSurvey in particular.

I also agree with Ben and Tpartner. While Limesurvey can provide an anonymous mode, saying if the survey is anonymous or not and really keep it anonymous is something only the administrator of the survey can do.

And even if with the completely anonymous mode, there are always ways for the administrator to implement certain things that might still pass on personal data to the survey.

When it comes to anonymity on the internet, you either trust the other side or you don't. It is almost impossible to guarantee 100% technical anonymity. If you have your doubts about the creator of the survey, rather don't answer or answer as if it was a non anonymous survey.

Ben, tpartner and holch, thank you very much for your responses and reactions.

Explaining the context seems now useful.
I work in an organization where there are pressions, tensions, between people, eventually between chief and employees, etc. as in a lot of work places.
My organizarion chooses to use limesurvey to ask for the opinions of the people and decides that it is compulsory (I've received a reminder). So I can't choose to not answer (it's real life, not a sweet dream).
Now, I've deeply understood the limit of limesurvey, so I will cheat my responses (what else could I do ?).

This is a real problem to pretend that a survey is anonymous if it may not be (It is illegal in countries where personnal data are protected by laws).

I'm wondering why nobody answers about my suggestion of a new word to qualify limesurvey-anonymity. This new word must convey honestly that there is no resl anonymity, but only a conditional anonymity. This word will help the people, administators and users, to take care about this limit.
No idea?

latnahc wrote: ....
My organizarion chooses to use limesurvey to ask for the opinions of the people and decides that it is compulsory (I've received a reminder). So I can't choose to not answer (it's real life, not a sweet dream).
Now, I've deeply understood the limit of limesurvey, so I will cheat my responses (what else could I do ?).

Hi,

The only good solution it's to give hosting and survey management to an external Society. I already ork with "Délégué du personnel" (staff representative in France). I host the survey : give access to staff representative and direction to the survey data except all user information. This was contractualized with staff representative .

This is a real problem to pretend that a survey is anonymous if it may not be (It is illegal in countries where personnal data are protected by laws).

latnahc wrote: .I'm wondering why nobody answers about my suggestion of a new word to qualify limesurvey-anonymity. This new word must convey honestly that there is no resl anonymity, but only a conditional anonymity. This word will help the people, administators and users, to take care about this limit.
No idea?

Because we can do anything we want in LimeSurvey : any dev can hack it .... (same for all system : only a good pact can do this).

Denis

I very much understand your concerns. We normally work with an external consultant that receives all respondent data, anonymises the results and only reports information about groups when there are more than 10 respondents, so results are never traceable to individuals.

Even then respondents sometimes ask the same questions as you do. And rightfull, because the way the external consultant works is based on contractual agreements that the respondents have not seen.

In the end it all comes down on trust.

I agree. A company/organization that has so severe trust problems within the different stakeholders should consider to contract an independent company that guarantees the annonymity of respondents.

As Dennis says, this is not a Limesurvey problem, this is a general problem. There is no 100% guarantee that a tool is annonymous, especially when there are things like tokens involved (to be able to remind).

Hello.
All the participants of this discussion agreed that a survey could appear as anonymous even if it is not.

The problem is that this information, obvious for all of us, is not expressed at all in the manual.

It is written, for instance :
"Anonymized responses?: This allows you to determine whether responses to your survey are matched up with information from your surveys tokens table, or kept 'anonymous'. The default is 'No'. If you choose 'Yes' then your survey is set to radically anonymize responses - there is really no way to connect answers and participants. Even the admin can't link response data and participant/token data."

In addition, in real life, it is socially quasi impossible to express doubt about the administrator, especially at work.

There is no warning about the fact that the user have to trust about anonymity without any proof and that he/she can be tricked. :woohoo:

This important limit should appear in the manual and should be known by the participants.

This question involves the ethics of limesurvey.
The reality is that it is, until now, impossible to have an anonymous survey (really anonymous, as I can do with papers, not maybeanonymousiftheadministratorishonest) AND to control that each user fill it only once. It is a scientific result.
And it's lying to pretend the contrary.

Errare humanum est :S , perseverare diabolicum ! :evil:

Paper surveys are not anonymous. Really. It is just as risky as with Limesurvey in Anonymous mode. If they want to trick you, they trick you.

The anonymous mode means that it is almost impossible for them to connect personal data with survey data. However, if they use a token, you can not see if it is in anonymous mode or not.

I can understand that this is a tricky part, but this is basically the same problem for any survey tool (e.g. you can register IP addresses via Javascript, etc.). With paper there are similar tricks that you can do if you want to. Just so you are aware.

lternative solution :
- Create the token (witjout information)
- Print little papers with the link and each token
- Put it in a hat
- Distribute

Denis