We recently completed a bicycle and pedestrian trail user survey using Survey123. One of the questions was "What is your primary residence?" and a list of locations was provided as well as an other option. When looking at the "analyze" feature, the graphic shows that 84% of users checked Janesville. However, when the data is exported to excel Janesville is not present in the data column. There are "yes" and "no", which were not an original option, as well as the names of other cities included on the list.
Any idea on what can be done for this data to show properly in the exported data?
The issue was that the question was originally a yes, no and then the author later changed the yes and no to other words and added additional options. Still exported the raw data as yes/no +. Apparently, that's a bug.
That doesn't sound like a bug to me. That sounds like the person changed the label for the answers *after* the survey was published instead of removing the original options and adding new ones. It's my understanding that, once the survey is published and the feature layer is created, you can still change the labels that are displayed publicly for the survey but not the value. That's intentional. In order to have a different label from what it was originally, you need to remove the response options and add news ones and ensure that the values are correct when you republish.
Thanks @Davidsalmon for reporting. We are looking at potential solutions to this for our next update as it can lead to data inconsistencies.
The fundamental issue here is that we cannot guarantee that the choice label and choice code match once you make an update. For this reason, we may need to disable updates on existing choices once a survey has been published. We will let you add and remove choices, but not to update them (once a survey has been published).