Ranked Choice-enquête gebruikt min/max-instellingen niet correct

Functie
1

Wanneer ik een meerkeuzestemming maak, kan ik het minimum- en maximumaantal keuzes instellen voordat er gestemd kan worden, maar dit lijkt niet te werken voor Rangschikkingskeuze stemmen.

Als ik een stemming maak met vijf opties:
poll type=ranked_choice results=always min=1 max=3 chartType=bar

De parameter max=3 lijkt ineffectief; een gebruiker kan nog steeds vier of alle opties rangschikken.
De parameter min=1 werkt alleen als het 1 is. Als ik deze op 2 of meer zet, wordt de stemming ongedaan gemaakt en is het slechts een lijst met opsommingstekens.

Misschien is dit opzettelijk of een probleem dat specifiek is voor de manier waarop rangschikkingskeuze werkt?
We hebben een use case waarbij we willen dat gebruikers een minimum aantal van de poll-opties rangschikken, en rangschikkingskeuze heeft de voorkeur boven meerkeuze.

2 likes
2

It feels to me related to ranked_choice implementation, maybe @merefield has some ideas here?

I am sure we can sort something out here.

1 like
4

Yeah, this isn’t currently supported, but I would definitely not call that a bug.

Perhaps the markdown interface is too permissive and that’s misleading and you might well argue that’s a minor bug. But it shouldn’t be a surprise to people who know instant run-off that this isn’t part of the algorithm.

It’s not really a standard feature of Ranked Choice, Instant Run-off Voting, where the voter has the discretion to select the number of candidates they wish to rank, and can abstain on the others.

Imagine restricting someone on a ballot paper. I guess someone could call your paper spoiled if you rank 4 out of 5 candidates and somewhere on the instructions you were told you could have only 3, but suspect that’s not usual as it would cause confusion and a lot of spoiled ballots which might upset the electorate.

The image is a ranking prompt asking to arrange four names, with "John Citizen" marked as the top choice and "Mary Hill" as the second. (Captioned by AI)
The image is a ranking prompt asking to arrange four names, with "John Citizen" marked as the top choice and "Mary Hill" as the second. (Captioned by AI)372×480 12.9 KB

But of course Discourse doesn’t run on paper :slight_smile:

And sure you could definitely add that feature! - it could make sense to limit things to 2 or 3 choices out of 5.

See:

1 like
5

To follow up I had a re-look at the build dialogue to refresh my memory.

The image shows a user interface for creating a ranked choice poll with options to customize voting settings and views. (Captioned by AI)
The image shows a user interface for creating a ranked choice poll with options to customize voting settings and views. (Captioned by AI)577×662 8.41 KB

As you can see there is no misleading setting for Ranked Choice on the form. So the bug is not here. I respect it remains visible on the markdown shell for Polls and that could lead you to believe it might have an effect.

I think this is fairer as a Feature request. Moving it there.

I’d be happy to work on it if hired to do so (I’m not a member of the Discourse team and as a freelancer not paid a salary, time is money … ), consider posting in Marketplace or PMing me direct for a quote if this is important to you. I’d be very happy to PR if acceptable.

3 likes