Do not show poll results to logged out users without user action

[matgb]

Title:
Do not show poll results to logged out users without user action

Area:
Polls

Summary:
If a user is not logged in, they should see the 'Fill Out Poll' display, with the 'submit' box marked unavailable, with a prompt to login in order to be able to vote.

Description:
Currently, if the reader of an entry is logged out, they see the results of a poll, without seeing the option to take part in any way. This leads to many users to assume they can't vote or the poll is closed; this is especially true of readers from other, non-LJ derived sites.

This is leading less off-site readers to take part in polls, and creates a bad usability experience and poor workflow, especially for users who have not logged in with OpenID before.

Instead of showing the current poll results to logged out users, they should see by default the fill out poll display, with a clickable option to see the results instead. In addition, a clear link to the login page, with a returnto keyed in, makes it much more likely an individual user will actually take part.

Ideally, of course, the logging in bit can be AJAXified alongside the other plans for polls.

I have mocked up a partial example of this using some CSS:
http://matgb.dreamwidth.org/376612.html the poll is from Miss_S_B's journal (this suggestion is at her request), and the returnto link takes you there. Unfortunately, a logged in user using style=mine will see the box intended for logged out users, but that's the best I can do without a change in the base code.

Thinking through for drawbacks of this, none spring to mind, logged out users, even DW account holders not knowing they're logged out, shouldn't see the results immediately before they've voted, logged in users don't, and it improves both interop and usability.

Poll #4089 Do not show poll results to logged out users without user action

Open to: Registered Users, detailed results viewable to: All, participants: 35

View Respondents

This suggestion:

View Answers

Should be implemented as-is.
15 (42.9%)

Should be implemented with changes. (please comment)
10 (28.6%)

Shouldn't be implemented.
7 (20.0%)

(I have no opinion)
3 (8.6%)

(Other: please comment)
0 (0.0%)

Comments

[azurelunatic]

If implemented, I would make sure the prompt to log in is at the top, below the master metadata of the poll, because I have more than once had a frustrating experience with a poll where I put some serious thought into filling stuff out, only to find out at the end that something did not go the way I planned it to, and I had wasted the time I took to fill things out. (Sometimes I did come back and fill it out again; sometimes I didn't.)

(I didn't see the mockup, as it doesn't appear to be displaying quite right in my alternate, logged-out browser, Chrome.)

[instantramen]

I for one hate websites that make it look like I can vote in a poll only to inform me I have to log in or register to participate.

[azurelunatic]

So the whole poll fill-out greyed out, and not just the submit?

[instantramen]

I think if it was changed at all, I'd prefer to continue showing the poll results, but where logged-in users who've voted see "Change Your Vote" logged-out users would see "Log in" and "OpenID" links or something to that effect.

[noracharles]

+1

[goodbyebird]

+1

[matgb]

My principle reason for cutting the results view as the default for logged out users is that it can discourage participation and it can also skew results, both are valid concerns within polling theory.

But you're right that having it display as if you can vote when you can't would be bad, that should've occured to me.

[zvi]

+1

[turlough]

+1

[kyrielle]

I don't think it should show the results, because it can skew the responses and is not ideal. I think it should show the poll title, a count of questions, and links to View Results, or Log In To Vote or the like.

[susanreads]

Something like this, yes, where you don't automatically see the questions or the results but can if you want to.

[syderia]

This

[ratcreature]

Yes. Me too.

[azurelunatic]

This is probably a separate suggestion, but supposing someone preferred to not have OpenID users answering a poll. We shall ignore the obvious "because people with accounts are better than people without them" argument, and proceed to "So what if we get slammed with a whole bunch of OpenID spammers such that it's a reasonable antispam measure"? or ...

... you know, I think that's kind of a bad idea. But I'll leave this here in case anyone else can think of a genuinely compelling use case.

[matgb]

Partially linked to my discount votes suggestion?
http://dw-suggestions.dreamwidth.org/406160.html

But yeah, I don't want to have the option of excluding external users, if we're treating them as full users without a journal, then that's what we should do. Perhaps validated OpenID users only?

[azurelunatic]

Really, as far as decision to implement goes, this seems like two separate suggestions.

Suggestion 1: when a logged-out user is encountered at a poll, insert a login prompt, including basic facts of life about OpenID, right where it should not be able to be missed.

Suggestion 2: when a logged-out user is encountered at a poll, show the poll questions roughly as they would be presented for voting, rather than showing the results first.


I think the former is totally absolutely sure to be taken up. The latter, that's probably going to have a bit of debate.

[feathertail]

How about just a "Log in to vote" button, next to a "See the results" button?

[matgb]

Yes, this makes sense. The latter, as I've said above, is partially linked to polling theory and specifically UK law, in the UK polling stations close at 10pm and exit polls aren't allowed to be published until 10pm, anyone that's seen or heard the result of the exit poll isn't allowed to vote.

Essentially, if you haven't yet voted in a poll, you shouldn't automatically see the current standings.

[jazzfish]

ISTM there's two different classes of "logged-out users": people who don't have accounts and aren't going to vote in the poll regardless, and people who just haven't logged in yet and are going to vote in the poll. The first class want to see the results, and ought to be able to. I'm sympathetic to the argument that the second class shouldn't see the results until after they vote, same as everyone else who votes, but I think that needs to be secondary to letting the folk with no account see the results.

I /do/ support putting a "Log in to vote" link / button at the top of (non-closed) polls.

[matgb]

See, theoretically, the number of people in (1) is very small. If you're browsing online and you don't have an OpenID of some sort then you're in a small minority.

The subset of people that know they've got an OpenID is, of course, tiny, but I think we need to do a lot more to promote the use of OpenIDs generally, not just on DW. An OpenID user can vote in polls, and specifically it would be good for some polls (the ones miss_s_b is currently running being a good example) to encourage external users to vote.

A clickable link to open the poll results is fine and part of the suggestion, so logged out users that want to see the results can do so (ideally, I'd like that link work through AJAX, but there's already a listed bug to make a lot of the polling system work with AJAX).

Most sites that accept OpenID logins are now moving to the point where people can use it without necessarily even knowing it's OpenID they're using; ideally I'd like us to get to that point, but the current poll display acts to exclude people that I think we should be trying to include.

[daweaver]

If you're browsing online and you don't have an OpenID of some sort then you're in a small minority.

That's a red rag, I shall adopt a suitably bullish demeanour.

I think we need to do a lot more to promote the use of OpenIDs generally, not just on [Dreamwidth].

And I think that the internet needs to be aware that OPENID is not a suitable trust mechanism. It was never designed to be repeatable, it was never designed to be accountable beyond the instant of posting. Bradley Fitzpatrick's ability to say "this is not a trust system", and implement "this is a trust system" still astounds me.

I'm yet to find a problem to which OPENID is a sensible answer.

[yvi]

I'm sympathetic to the argument that the second class shouldn't see the results until after they vote, same as everyone else who votes

You can already always see the results even if you haven't voted - every not filled out poll has a "[ Display Results ]" link.