Poll that requires you to choose more than 1 response doesn't actually require you to do so

[silverflight8]

Title:
Poll that requires you to choose more than 1 response doesn't actually require you to do so

Area:
Entries

Summary:
Either the polls requiring more than 1 response should be unskippable, or there shouldn't be such an option, and more than 2 responses should be the minimum.

Description:
You can post polls that restrict the number of votes (ticky-boxes). This works as intended for anything requiring 2 or more votes. However, you can also choose to skip questions. The ability to skip questions makes the option "at least one answer required" useless.

It should either be:

1. an unskippable question; or
2. the minimum setting should be "at least two answers required".

I'm not sure which would be better, and which would be easier to code (especially considering the fact that a poll can be constructed using custom poll HTML, not just the poll wizard), so I'm throwing this to the comm at large.


(with thanks to azurelunatic for bringing this up and discussing it with me in the first place).

Poll #6505 Poll that requires you to choose more than 1 response doesn't actually require you to do so

Open to: Registered Users, detailed results viewable to: All, participants: 44

View Respondents

This suggestion:

View Answers

Should be implemented as-is.
11 (25.0%)

Should be implemented with changes. (please comment)
4 (9.1%)

Shouldn't be implemented.
7 (15.9%)

(I have no opinion)
22 (50.0%)

(Other: please comment)
0 (0.0%)

Comments

[rho]

"Answering" a poll without actually answering any of the questions is currently used as a way to say "I'm not interested in filling out this poll, but I want to see the results". The addition of unskippable questions would break this behaviour or, at the very least, require it to be special-cased in code.

[deborah]

it's kind of a weird workaround, though, isn't it? And not everyone knows about it, which is one of the reasons I usually put in a "I'm not going to answer any other questions in this poll" radio button.

Making the "display results" link switch to displaying directly in the page as if you had answered the poll instead of taking you to another page might serve the purpose better.

[rho]

Oh, it's definitely weird and not remotely ideal. I'd be entirely in favour of adding a less opaque way of achieving the same thing, but until and unless we get that, I'd be sad to lose what we do have. I wasn't intending my comment as "we shouldn't do this because..." so much as "if we do this, then we also need to consider..."

(As an aside, I don't like adding "I'm not answering" options to polls I create, because it messes up the answer percentages.)

[matgb]

This are because you're a stats obsessive ;-)

So I'm ticking 'with changes' so that if we implement this, we ought to make sure a better 'not answering' option is present at the same time, something we ought to do anyway but it'd be a low priority for me, unless there's someone that knows the poll code well enough to just do it quickly.

[cesy]

+1

[arethinn]

Making the "display results" link switch to displaying directly in the page as if you had answered the poll instead of taking you to another page might serve the purpose better.

+1

I was going to say that I was confused by what rho was saying - sure you can "not answer" a poll to see results, but it seems like the wrong/unusual way to do it, because isn't that what the view results link is for? It'd be great if results displayed inline like when you actually fill out a poll.

[ciaan]

Sometimes polls have questions that are not applicable, or that are boring, or that have no correct answer shown, and I would hate to be forced to answer them in order to fill out the rest of the poll.

[matgb]

Conversely, sometimes the creator of the poll needs an answer to one question to put other answers into a context, and not answering that question can mess up the data acquired from the rest of the poll.

I think it should be poll creator choice, and if a user doesn't want to complete the poll due to that choice, that's entirely up to them.

[ciaan]

Just so long as it is a specific setting the poll creator can do, and doesn't automatically apply to everything.

[matgb]

That is the suggestion, yes.

[ariestess]

+1

[feathertail]

This reads more like a bug report than a suggestion, even though it contains a suggestion for how to resolve it.

[arethinn]

I agree, and I'm not sure I understand just what the supposed problem is.

[azurelunatic]

It's a bug, but there are two possible ways of fixing it. It's which way to fix it that's in question.

The bug is: you can select a "must pick at least one of these options" thing, but someone can enter zero options and still fill out the poll. The label does not match the behavior.

[arcanetrivia]

Ah, ok.

[silverflight8]

Since there is debate on which way to implement it, and I can see both sides have good points, so I'd rather solicit suggestions first.

[solitarywalker]

I don't think that there should be "unskippable" polls. If someone doesn't want to answer your poll, just accept that and go on with your life. It seems to me that the best solution to the problem presented would be a rewording of the cited text: "multiple answers permitted", "please choose at least one", etc.

[cesy]

If you don't want to answer the poll at all, that's easy. But if you're answering the poll, the required questions within the poll should be required.