> On Nov 21, 2019, at 9:39 PM, Craig Russell <[email protected]> wrote: > >> On Nov 21, 2019, at 7:33 PM, Georg Link <[email protected]> wrote: >> >>> 3. We should disallow multiple responses from the same individual to the >>> survey. This might involve assigning an immutable ID to each person who >>> receives or requests a survey. >> >> Tricky. If we share one link on social media, then everyone using that same >> link expects to take the survey. We cannot prevent anyone from taking the >> survey again. I do not like to add a separate step of requesting to >> participate and then us manually sending out personalized invites. > > I thought that this would be the trickiest part. > > I thought that it might be possible to send an individual invitation to each > person that we either know of (email address from our committer base) or who > requests a survey for their specific email address (non-committer > contributor) and we could then send an immutable survey ID. >
Yes, we can. But that would not allow us to have a public survey. The survey would be behind a “token-wall” (I just made up that word). Every participant would need to either enter their unique token or use a link that already contains their token. This would make it difficult to advertise the survey on social media, because we would not have a link to share for directly taking the survey. I recommend to keep the survey open and risk that someone provides multiple responses. I believe the benefit of having an open survey and inviting everyone outweighs the risk (but I don’t know everyone in the ASF community to accurately make a prediction). > Is there any other way to discourage if not eliminate multiple responses? Technically, not if we want to encourage everyone to participate, even though social media. Socially, yes. We can draw attention to the fact that multiple answers are possible and then tell people not to use that possibility. Seems counterproductive though. Do you have someone in mind that wants their specific demographic information to be counted several times? Or someone who wants to dilute the responses? It would require answering 20+ questions multiple times — quite a bit of effort. -- Georg Link (he/him)
