Closed ajparsons closed 5 years ago
Sorry yes, people in the future: @dracos is right and I'm talking about being able to match up surveys internally to the metadata of messages (while respecting usual rules about not examining contents of messages), not publishing individual results publicly.
Further described in: https://github.com/mysociety/research/issues/151
This adds a post-survey embed (or a link to the survey if Javascript is off) to 50% of people writing to all types of representatives.
This survey is not anonymous as it passes the message ID so responses can be reconciled with other features of message (postcode - and so deprivation by lsoa - and representative type). I think this is OK as we're not asking for personal or political information of the individual as we were with the last survey - and this is effectively the same status as the in-built questions. Open to alternate approaches, but I think even converting postcode to LSOA and passing wouldn't really be anonymous as time+lsoa is likely to be unique.