Hi,
I have a situation where multiple recipients happen to share the same e-mail address. For example, a husband and wife have a joint e-mail account but they are to participate independently in a survey. In my tests it would appear that LimeSurvey (v2.5) automatically discards all but one recipient for each e-mail address even though they have different tokens.
Is there a way to send an invitation to every recipient irrespective of whether their e-mail address is unique?
Thanks,
Colin

Hi,
I've just checked it and I could send several invitations to the same email addresses. LimeSurvey automatically discards duplicate addresses when importing them and the option "Filter duplicate records:" is on.

Alternatively, you can send one email to one address and allow for mulitple responses for one token - use the option "Uses left:".

Regards

Thank you Elisa.
I deliberately bypassed the duplicates check on user import and can see that I have the multiple instances of the same address in the participants table. I can't take advantage of the multiple responses facility in this case since each invitation mail / survey includes per token personalised data drawn from custom attributes. However, if multiple mails worked for you then they should work for me!

I wonder what mail delivery mechanism you are using? I'm currently using the PHP mailer but will reconfigure and test with SMTP once my current survey is complete.

Regards,
Colin

Generally, I'm still using LS2.06 with SMTP, but for this test I've used LS2.5 with PHP mailer.

Regards

Once again, thank you Elisa.
How peculiar that I am seeing two differences under such similar circumstances.
I'll have to hunt for more clues.
Best wishes,
Colin

Dear Elisa,

In case it is of interest to you, I've carried out further testing and have established that multiple participants with the same e-mail address do indeed work provided that they have different names. In my data set my participants shared e-mail addresses and names as they were differentiated by token and various custom attributes. This is arguably correct since both participants would have registered with the same e-mail address which has just one first and last name.

My evidence appears to indicate that LimeSurvey does do an unexpected (to me at least) de-duplication of the participants table at the time of invitation generation and appears to use the same default critera as offered during participant import whether or not it is appropriate for the data-set. Since there is no mention of this functionality in the Wiki or on the application pages I suspect that it is unintended behaviour. After all, if this was intentional then surely it would have been appropriate to either mimic the criteria actually used during participant import or offer a fresh selection box to gather de-duplication criteria to use at invitation generation time. The complete absence of these features during mail generation suggests that the system works on the premise that recipient de-duplication (according to criteria specific to the recipient data-set) has already been carried out and, therefore, since recipients have already survived de-duplication and now each have a unique token they should be considered unique irrespective of any field values they may have in common with other recipients.

I'm going to re-check the bug tracker and add a new bug report if necessary.

Best wishes,
cjfzim

Hi,
I haven't tested the sending functionality so carefully. In fact I didn't need this specific behaviour so far. But you're right and it would be good to have the system working consistently. Thank you for addressing this issue.
Regards