This is a very interesting product observation, I am glad I was able to with the customer on this.





While comparing information in two seprate objects (ao participants and flattened survey) we found a data discrepancy.





A survey participant had been sent the same survey in two separate months, the second being a follow up because the participant did not respond to the first program.





The initial send was the 18th of December and the follow up on the 3rd of January.





The data showed that the respondant took the survey on the 10th of Jan while you would assume they took the 3rd of January survey as it was active and the 18th of December send showed "knocked off".





After investigation we found the link from the 18th send was actually still active and the particpant took the original send.





We then found that this is actually expected. If the Participant Dropped/Knocked off then we are not storing survey responded details in AO objects but updating in Survey objects.





To explain a bit more if a participant is manually Knocked off then the system won't deactivate the Survey link immediately, that participant can respond to survey until it met the survey expiration date mentioned in either the Survey or Program Conditional wait.





My ask is to still update the AO participant object because this is not documented anywhere and can easily cause confusion.