Hi,

2.51.0+160901 with a non-anonymous survey with single-use tokens. The participant database shows a valid 'date completed' for one record but results show two entries:

I'm trying to understand how the system works -- is this expected behaviour?

Thanks,
Martin

lemonbrain wrote: I'm trying to understand how the system works -- is this expected behaviour?

That is a normal behaviour. One time is counted when the survey is completed. If you have no persistence via cookies activated, you might see many entries with the same token as long as the survey is not completed.

Thanks! So, the Summary counts - 'full' and 'incomplete' responses - are not mutually exclusive. (I mistakenly assumed respondents, not responses). I'll have to be a lot more careful than I expected when exporting (all-)responses.

With token you have a settings to link one response to one token (or not). Seems it's false by default. Maybe it can be great if it yes by default.

Denis

Hi Denis,

This survey is running with:

Allow multiple responses or update responses with one token: Off (Without the hint text it would be impossible to parse 'Allow A OR B' On/Off)

... and I guess you were referring to ...

Enable token-based response persistence: Off
A few have responded with separate sessions (e.g., different hosts) and I guess locking might be an issue if someone switches from a mobile to desktop mid-session. In this case, de-duplication did result in minor data loss -- one participant offered varying responses to a few questions.

So the first one in setting "OFF" will prevent people from completing more than one survey with the same token. However, if they don't finish the first time, the next time they come they might start a whole new survey, as the first one was incomplete.

You can prevent this only by setting token based response persistence, which means that if you stop responding and come back later to the same token based survey Limesurvey should jump where you left of and you continue this one survey.

Is it possible for there to be an incomplete survey with some questions answered, and a complete survey for the same single use survey?

I have an anonymous survey I just administered and for the first time I have an "incomplete" response that had questions answered. I'd like to count the partial data if I can be sure that the person wasn't already counted among the "complete" responses, but can I be sure of that? (Other than trying to match up answers to a completed survey, which I don't want to rely on).

nbond wrote: I'd like to count the partial data if I can be sure that the person wasn't already counted among the "complete" responses, but can I be sure of that?

You cannot be sure. It's an anonymous survey. Even with a restricted access via token, a complete and incomplete survey can be done by the same person. Just stick to the complete surveys.