Hi,

I am analysing data with R and recently came across an issue in my setup. I have limsurvey server on a laptop and people connect to it via WiFi provided by an access point that I carry around too. That access point in dhcp points to itself as DNS server and therefore can point to my laptop's site lime.lan Works very well, but I came across something interesting:

For some responses submitdate (which is mostly identical to datestamp) seems to be an hour off. I need to understand if it is user problem (they submitted responses early) or is there any chance server might have recorded responses in a different timezone (in which case it probably should be for all not a few responses?).

Surveys are non-anonymous from limesurvey perspective (i.e. I record all variables) but practically anonymous, because no responder-identifying information is collected. One could link response timings and CCTV, but that would be paranoia and besides there is no CCTV

So a couple of questions:

1) What is the difference between submitdate and datestamp fields?

2) Do they record time strictly from server (i.e. my laptop's clock in this case) or is there any chance the client could provide the time for those records?


Background: This is used to collect feedback on teaching quality and timings presently serve to distinguish which talk the feedback relates to. There should be time difference between response groups to be able to distinguish which teaching piece the given feedback refers to. This linkage can be improved and I am working on it, but I would really appreciate clarification regarding those 2 questions.

Thanks,

Robert

As far as I know, the date writen to the database is based on the system time on the server. In your case, this should be your laptop. I don't think the client provides the date/time. This could be pretty problematic when running online surveys in different time zones, I think.

But maybe one of the developers can shed a different light.

What OS is running on the notebook?
The dates/times are provided by the OS which are accessed by PHP.
Exactly one hour off seems to be a timezone related.Daylight-saving would be an hour off too.
The clients are not providing any time/dates.


Is your survey all in one? Or group/questionwise?

The difference between submitdate and datestamp field is easy.
The submitdate is only shown for completed surveys. The datestamp is the last time the survey received answers by the respondent (which depends if you have all in one or a group/question wise display mode).

holch wrote: As far as I know, the date writen to the database is based on the system time on the server.

Confirmed.

And if you're server didn't have good time (for you) there are a settings : timeadjust

Thanks very much Jelo and Denis,

Laptop is running arch linux. Survey is all-in-one - that is there is the one mandatory group - it has 4 questions in it (3 sliders and one text box), but I display neither group name nor description. My experience is - if I want audience engagement I want to provide something that is simple and easy for them to do.

So I then assume the submitdate is the laptop system time the moment web server receives request after user clicks the survey-submit button. And datestamp is whenever last activity detected, even though user may have not submitted (probably only if save-survey?).

Also - did not know we had timeadjust. This is very good to know.


I think in my case the explanation then is that the savvier users (young doctors) submitted their feedback not after but during the lectures. I am careful that my time on server is ntp-synchronised.