Loomio

Experiment with some new polling types!

JK
James Kiesel Public Seen by 88

Hello product development community!

We've been hard at over the summer, developing up some brand new ways of making decisions with Loomio, which we're ready to release for beta testing. The current additions:

  • Polls Ask users to select one or multiple choices from a user-defined list of options
  • Counts Ask users to respond to a yes or no question

And coming up on the roadmap:
- Dot voting Users can distribute a set number of 'points' amongst a set of user-defined options
- Event planning Find a time that works for everyone (Doodle style)
- Brainstorm Allow users to submit ideas en masse without the need to judge them

I've enabled the new polling features in this group, which you'll see in a card marked 'Decision tools' on the right hand side. We're hoping the functionality will explain itself; if not, let us know!

Imgur

We're very excited about this new step in the Loomio product, and look forward to your feedback. Enjoy!

RG

Rob Guthrie started a check Mon 27 Feb 2017

I'l tried the new polling types Closed Fri 3 Mar 2017

5 - Yes
1 - No
RG

Rob Guthrie
Yes
Mon 27 Feb 2017

here we go!

JK

James Kiesel
Yes
Mon 27 Feb 2017

BH

Bob Haugen
Yes
Mon 27 Feb 2017

JK

James Kiesel
No
Tue 28 Feb 2017

MB

Matthew Bartlett
Yes
Tue 28 Feb 2017

What is count?

JK

James Kiesel
Yes
Tue 28 Feb 2017

IR

Ignacio Rojas
Yes
Tue 28 Feb 2017

S

Simon
No
Wed 1 Mar 2017

IR

Ignacio Rojas
No
Wed 1 Mar 2017

IR

Ignacio Rojas
Yes
Wed 1 Mar 2017

BH

Bob Haugen started a poll Mon 27 Feb 2017

Do you like polls? Closed Thu 2 Mar 2017

Outcome
by Bob Haugen Thu 2 Mar 2017

Looks like most people in this poll like polls!

13 - Yes
3 - No
2 - Please! Can we talk about something important?
BH

Bob Haugen Mon 27 Feb 2017

Please! Can we talk about something important?

Only because that was a stupid poll...

JK

James Kiesel Tue 28 Feb 2017

Yes
DS

Damian Sligo-Green Tue 28 Feb 2017

Please! Can we talk about something important?

I don't like binaries

SW

Sir Wumpus Tue 28 Feb 2017

No

Polls are too rigid and pre-frame the discussion in the terms used by the pollster. Simple discussions are better than polls for gathering initial opinions and coming up with ideas, while regular loomio proposals are better for finalizing.

JW

James Wilson Tue 28 Feb 2017

Yes
B

Bady Tue 28 Feb 2017

Yes
MB

Matthew Bartlett Tue 28 Feb 2017

Yes

Especially these ones: https://twitter.com/dailyvotingpoll

LP

Louis Parisot Tue 28 Feb 2017

Yes
I

Immi Tue 28 Feb 2017

Yes
ST

Stacco Troncoso Tue 28 Feb 2017

Yes
IR

Ignacio Rojas Tue 28 Feb 2017

Yes
RG

Rob Guthrie Tue 28 Feb 2017

Yes

We're long overdue for supporting more kinds of decision processes. Looking forward to the creativity that this unlocks.

S

Simon Wed 1 Mar 2017

No
S

Simon Wed 1 Mar 2017

No

Testing, testing

JD

Josef Davies-Coates Wed 1 Mar 2017

Yes
CS

Christian Seppa Thu 2 Mar 2017

Yes
SP

Sam Parkin Thu 2 Mar 2017

Yes

I like new shinies

JG

john gieryn Thu 2 Mar 2017

Yes

clicking into the poll from my email worked great

DS

Danyl Strype Thu 2 Mar 2017

No
F

fauno Tue 28 Feb 2017

great! but how do you enable them on custom installs? i've set the
group to enable experimental features already, but i don't see the card.

RG

Rob Guthrie Tue 28 Feb 2017

You can enable polls from the /admin section of loomio. There is an enable polls button.

GC

Greg Cassel Tue 28 Feb 2017

It's possible to directly access @bobhaugen 's poll without seeing this discussion topic which it's attached to. That was confusing for me as I scanned my new Loomio content quickly this morning.

I think that it'd be fine for polls to be associated with a single comments section for a discussion topic which they're attached to-- if you make the relationship as clear as it is for Loomio proposals. (I can see a discussion topic in a window to the left of each Loomio proposal which I consider.)

With that in mind, please accept this version of the comment which I originally made (in a confused state) in another thread:

It's possible I should give this input on Loomio Roadmap or github, but in case it feels like a discussion "subtopic" to anyone: I think that it's very important for polls to have comments sections.

For instance, @bobhaugen created a poll last night in the Product Development subgroup. Please correct me if I'm wrong, Bob, but I believe that poll is supposed to be intentionally silly in a way which criticizes the perils of poll design.

I don't seem to have any option to comment upon that poll without registering one of the predetermined positions. That's a problem for me, because it would be inaccurate (and misleading) for me to register any of those positions.

I care lots about the potential of good poll design, so I'd like to directly and efficiently give my feedback on that specific poll.

I hope that my comments here seem sufficient to indicate the importance of providing an option to comment on a poll without choosing one or more of its options. Does this make sense @gdpelican and others?

BTW, I love having previous poll results hidden by default, but available to people who haven't voted. :)

BH

Bob Haugen Tue 28 Feb 2017

Please correct me if I'm wrong, Bob, but I believe that poll is supposed to be intentionally silly in a way which criticizes the perils of poll design.

Yes, intentionally silly, but not meant to criticize the perils of poll design. As I wrote in that other thread, I was just trying to test the feature quickly, not design a good poll. I apologize for the silliness. Please design a better poll!

DS

Danyl Strype Sun 9 Apr 2017

Chronological comment threads are a hammer we're so used to online that there's a danger every UX problem starts to look like a nail. I think if the devs have got the range of new facilitation tools right, and the design of each of them right, then they each offer a solution to the many problems we've all had in Loomio group comment threads (eg the infamous Editing Comments' discussion). If this is the case, do you think there's a risk that tying polls tightly to a meta-discussion via a comment thread somewhat defeats the purpose of them?
EDIT: fixed link

RG

Rob Guthrie Sun 9 Apr 2017

I think the question is moot.
The new facilitation tools are not tightly tied to discussion threads. Or groups. They've been built so they can be integrated into any other platform you like really. Check /p/new and start a poll without any discussion thread at all. Much more granular.

GC

Greg Cassel Sun 9 Apr 2017

I definitely do not think that chronological non-threaded comments are the right solution to our online discussion problems. I personally design systems for structured conversation. Nonetheless, I think that having a comments option is generally much better than not having one.

I don't think that providing a meta discussion thread inevitably means that a poll is tied "tightly" to the meta-discussion. The strictness of association depends deeply on social context.

With that in mind, I reiterate part of my earlier comment here:

I don't seem to have any option to comment upon that poll without registering one of the predetermined positions. That's a problem for me, because it would be inaccurate (and misleading) for me to register any of those positions.

I care lots about the potential of good poll design, so I'd like to directly and efficiently give my feedback on that specific poll.

With that additionally in mind, I should point out that IMO, poll creators would ideally have the freedom to enable or disable commentary, just as they'd be able to enable or disable others' ability to add poll options/ categories. However, in a world with such technical freedom and power, people could & should be held socially accountable for any questionable decisions, such as disabling commentary on polls.

FYI this all ties directly with the two key questions I ask of technological design options:

  1. Does this technology create ability for people to communicate or organize?

  2. If this technology restricts ability to communicate or organize, should that be a technical standard or a community standard?

GC

Greg Cassel Sun 9 Apr 2017

@robertguthrie commented while I was typing, such that my last comment is mostly irrelevant now. The granularity/ modularity he refers to is IMO far superior to tying user choices to a monolithic platform.

DS

Danyl Strype Sun 9 Apr 2017

Ae, I think we're iterating towards the same position from opposite starting points. Loomio's original interface traditionally worked by "tying polls tightly to a meta-discussion via a comment thread", and a major design decision has been made to move away from that. We all agree (I think) that being able to associate a poll with A discussion area is good, but also that it doesn't have to be a comment thread on Loomio. With the polls becoming more "much more granular", as @robertguthrie puts it, it opens up possibilities for integrating them into other platforms that folks like @bobhaugen have been asking for #UXWin !
EDIT: formatting

QG

Quentin Grimaud started a poll Mon 13 Mar 2017

Which voting methods are acceptable according to you? Closed Mon 1 May 2017

I create a new poll because I when I found this thread the poll was already closed and I wanted to test it :) See https://plato.stanford.edu/entries/voting-methods/ for explanation of several voting methods

2 - Plurality Rule
2 - Borda Count
2 - Plurality with Runoff
3 - The Hare Rule
2 - Coombs Rule
2 - Negative Voting
4 - Approval Voting
1 - Cumulative Voting
1 - Dodgson's Method
7 - Condorcet Rule
QG

Quentin Grimaud Mon 13 Mar 2017

Condorcet Rule
Approval Voting

great feature! :) :)

DK

Dmytri Kleiner Thu 23 Mar 2017

Condorcet Rule
JK

James Kiesel Wed 5 Apr 2017

Coombs Rule
The Hare Rule
LP

Louis Parisot Wed 5 Apr 2017

Condorcet Rule
Plurality with Runoff
IR

Ignacio Rojas Fri 7 Apr 2017

Condorcet Rule
Approval Voting
Negative Voting
The Hare Rule
DU

[deactivated account] Mon 17 Apr 2017

Condorcet Rule
Approval Voting

I dream of a system that successfully combines Condorcet ranking/preference and approval methods into one.

SP

Steve Phillips / @elimisteve Wed 26 Apr 2017

Condorcet Rule
Borda Count
Plurality Rule
ZB

Zeljko Blace Fri 28 Apr 2017

Condorcet Rule
Dodgson's Method
Cumulative Voting
Approval Voting
Negative Voting
Coombs Rule
The Hare Rule
Plurality with Runoff
Borda Count
Plurality Rule

why make choices if all can be implemented as applets/plugins?

JB

Jo Booth Sat 29 Apr 2017

DW

Doug Webb Tue 9 May 2017

Hello! I missed the boat with this one... Approval voting has been mentioned and scored fairly high in the poll. Approval voting is a sub-type of score voting (a.k.a. rating.) In approval voting you rate using two 'levels', in score voting you rate with 3 or more levels.

Rating methods such as approval/score (cardinal systems) have a lot of theoretical and practical benefits over ranking methods such as Condorect/Coombs/Hare/Borda/etc (ordinal systems) in terms of actually identifying what people want.

Very interested to hear if score voting is in the pipeline and if not, happy to advocate further! :sweat_smile: For further info on this subject do check out this amazing email list for ongoing discourse about Social Choice theory and this informed but opinionated website about Score Voting.

QG

Quentin Grimaud Tue 9 May 2017

On a side note, during the last few months in France, a recent voting method is being very positively discussed: it's called Majority Judgment. I am curious what you guys think about it :)

DU

[deactivated account] Wed 10 May 2017

Wow thanks for that email list link Doug! - I just had a look and just found out that PAV had already been invented/theorised about a decade before I even began to come up with it myself - it is really awesome to know that PAV academia already came into existence (about 10 years) before I invented/coined the exact same theory of PAV in my own informal intellectual ramblings seen here: https://www.loomio.org/d/zgYdEaFp/internet-party-democracy

I understand that PAV, which combines both Approval and Preferential/Condorcet voting, is the most optimally inclusive method to achieve a voter convergence outcome.
When openly electing from the group member base it also produces new organic data about where social edges/bonds exist within a group, and which members already naturally tend to serve as central points of group cohesion for a given type of member function (i.e. the responsibilities of the role that is up for election).

DW

Doug Webb Sat 13 May 2017

Hi Quentin. Majority Judgement looks pretty good: it's a score voting system with some additional features. I recall reading somewhere that the authors found good reasons to use verbal indicators instead of numeric indicators...

DW

Doug Webb Sat 13 May 2017

Glad you find it useful! Don't get dragged in though... some of those conversations are a bit mind-warping. I would be very interested to test out a proportional approval or score voting app. For me I'm still mostly concerned with single-winner methods.

DU

[deactivated account] Sat 13 May 2017

True.

When it comes to proportional ​voting I understand that there is no single best way to do it - because it is generally premised on assuming that groups can be partitioned into distinct groups of unique identity - a good example of attempted proportionalism is the party vote.

However identity is multidimensional and no individual belongs solely to one kind of group - but can and does belong to many categories.

This is why I think designing voting systems based on identity politics (and proportionalism) is inherently flawed.

But rather single winner and multi winner methods are best if they are premised on finding the nodes where edges of popularity tend to centralise.

And if in multi winner methods the outcomes also just happen to look like​ a proportional selection of various 'identity' categories, then this fortunate coincidence is just how the cards stacked up when everyone wrote their ballot.

DW

Doug Webb Sun 14 May 2017

Ah! Yep, for electing representatives to positions I think multi-winner methods could drastically improve current governance systems. However, I believe that 'single-winner' methods are ultimately required for direct decisions (like setting up the systems for having representatives in the first place...)

In a community setting for example, you might need a single-winner method for selecting a policy for sharing the projector. However, that policy might specify a multi-winner method (like PAV) for selecting which films are watched during the week.