Propose new features
Designing Decidim together
The problem
Currently, debates allow participants to post comments/arguments in favour or against a proposal, and they can also express if they agree or disagree with the arguments issued by others. However, this information is not aggregated, and thus, it can be difficult to asses the debate results (that is, if the proposal should be accepted or rejected).
The solution
It would be really useful to aggregate all the opinions issued in a debate so that the result becomes apparent. The attached images illustrate this idea (thumbs up/down are just to illustrate the idea, proper design would be required).
This aggregated opinion would help to increase the quality of the debate (an thus, of the democratic process it represents), and may incentivise participants to pose their opinions.
Considered alternatives
The Barcelona City Council has funded several research projects for applying Artificial Intelligence techniques to solve this problem. Along these projects, several aggregation methods have been proposed but not yet added as functionalities to Decidim. These methods go beyond the simple idea of considering the proportion of comments in favour or comments against the proposal because, as shown in the attached images, some arguments may be stronger than others (that is, more people agree with them).
Additional context
These projects have provided the code for the aggregation computation as open source, but it needs to be translated to Rubi for it to be incorporated in Decidim.
We are researchers from the University of Barcelona and the Artificial Intelligence Research Institute (IIIA-CSIC), and are in touch with Platoniq to see how this new functionality could be incorporated into Decidim.
Does this issue could impact on users private data?
No, participants’ opinions are publicly shared.
Funded by
It could be partially funded through Fundació Solidaritat UB.
Related images
Report inappropriate content
Is this content inappropriate?
Comment details
You are seeing a single comment
View all comments
Conversation with Ivan Vergés
It will apply to every thread of comments @carol . It wouldn't be difficult to set some configuration options for this at the component level also
Yep, hence my question. If everything we do to improve deliberation applies to comments, then perhaps the Debates component doesn't make sense? (my opinion so far is that it does make sense for it to exist, and some comment functionality should be specific to that component).
Good point, the way I see it Debates is like a proposal without all the steps before publication. Maybe in this component deliberation options should be enabled by default while disabled in others (but still available if needed).
Maybe this is also an opportunity to try to add some features to the component debate so it have more sense (this could be just a design change, more oriented to discussion maybe, or different nesting level).
Very much agree with you :)
I think this kind of "stacked bar chart" showing the weighted result of positive, negative, neutral comments is cool _for comments_. Available everywhere (in each component).
_Maybe with some extra:_
-- Admin could deactivate this "deliberation options" for each component.
-- Maybe it is interesting to add "question" option?
For the _Debates component_, I think it might be very interesting to review the approach that https://consider.it/ is using.
Video example: https://d2rtgkroh5y135.cloudfront.net/images/product_page/trimmed-overall.mp4
_PROS:_
- Better granularity between agree/disagree or Against / Agree.
- Focused on the arguments. The system promotes defining PRO/CON arguments for each person
- Interesting way to visually filter and explore arguments (you can explore arguments by highlighting positioning segments and display weights per argument with more supports and other options).
Could even be better if:
- We should be able to filter or represent the weights in some way, to see which arguments have been viewed or rated by how many people. Because one of the biggest problems is that people don't read enough, and maybe they rate only the last ones or in any case a few contributions.
- In line with that could be interesting to give people an easy way to explore contributions in order to evaluate more quickly and with more insights what others are saying & been able to contribute with more context in mind. (e.g. show how many arguments are less evaluated or seen...). This could be achieved with to axis graph like scatterplot or layers of density plot or some kind of heatmap.
Thanks a lot @aram_l_roca for pointing consider.it. I agree it offers a useful visual display that shows in a very intuitive way the general trend of the deliberation in terms of the distribution of participants that agree and disagree with a proposal (that is, it becomes very apparent if most people agree, or if participants are polarised, etc.).
However, I would like to elaborate further on a couple of the issues you mentioned:
- Granularity: In consider.it participants can express how much they agree or disagree with a proposal as a percentage (for instance, someone can agree 40%) and provide pros and cons for it (this is similar to the comments/arguments). So, indeed, as you mention, they offer better granularity, but only for the proposal. Pros and cons are included by writing a new one or by doing “drag and drop” of the ones previously posted by other participants, so there is no granularity here, and in fact participants cannot express if they do not agree with a comment (thus, drag and drop is equivalent to the green ^ in Decidim’s comments shown in the second attached image, but there is no equivalent to click into the red v). As for the mathematical method we have developed, we can use any positive or negative number to express the agreement or disagreement with comments, so it is a matter of interface (as it is now we can map v to -1 and ^ to +1, but we could use any other granularity, as for example, 5 stars or a sliding bar).
-Focus on PRO/CON arguments: This could be translated into Decidim if, when issuing a comment, the participant had to always specify if the comment is “In favour”, “Against” or “Neutral”. Currently, the default is Neutral (which may also be associated to question, if I understand your comment) and only for the first level comments. In fact, our mathematical computation works with labelled (“in favour” or “against”) comments, so this is a requirement for us. Additionally, to give visual hints about pros and cons, comments in favour and against could be displayed in two columns, which is highly recommended by usability experts and implemented, not only in consider.it, but also in other deliberative platforms such as kialo or your priorities.
Having said so, and acknowledging again that consider.it provides a very intuitive idea of the overall status of the discussion, let me remind you that this proposal functionality is about providing an aggregated value that helps to assess the debate results (that is, if the proposal should be accepted or rejected). Min 0:26 of the third tour video in https://consider.it/tour shows that proposals have several associated scores: the total score, which corresponds to the sum of opinions (where opinions are on [-1,1]) but does not consider pros/cons; the activity score, which is the raw number of opinions and recognized pros/cons; the average score, were each proposal is scored by the average opinion score; etc. However, these scores are used to sort proposals rather than to assess the debate results. Our claim in the functionality proposal is that we should have a metric that goes beyond averages, since some arguments may be stronger than others and this should be taken into account.
Maybe a differential point for the debates (as opposed to other commentate components) could be to force the user to use the in favor/against/neutral button in order to comment?
That makes a lot of sense @microstudi ! Comments in a debate become arguments in favour and against (or neutral) the topic under discussion, and then it has to be crystal clear (thus, forced).
Loading comments ...