bring "narrative" front and center

[jywarren]

Wireframes have complexity without narrativizing their data--narratives could make it "make it personally valuable for people to contribute."

And focused on relationships rather than questions or data:

Answers are means of relating generally rather than specific nodes that people can cluster around "I'd much prefer to have some sort of mode of engagement that is relationship building (e.g. showing me how my responses relate to other peoples'), and have other ways for free discussion, but not directly attached to my response.

[shapironick]

My response just now to #3 was also a response to this issue.

[jywarren]

@shapironick's questions on #3 -- first, how often and where do we invite people to look at others' stories? After each time they answer a question?

Nick: Looks good--I do wonder if we should be encouraging exploring after each question or if maybe after each category there would be space to explore. Because 1) people probably like doing about 5-9 things of the same kind of activity before switching to another kind and also 2) we will not want to actively distract them from entering information. 3) we can decenter relating from single Q&As and move them towards a larger space of dialog where interactions could go between and beyond questions.

Also important - can we ask for open comment (paragraph) after each question, to open up to the stories people want to share? This makes me think we should not make people feel like "finishing" the questionnaire is as important as telling their story and reading others' -- and communicating with each other. The questionnaire is important to the research, but the primary use of the site (for its users and for me) will be its use as a platform for communication and exchanging useful information.

[shapironick]

I think after every section of questions we could have a space for more narratives as they can then be cued by several questions within a section to tell their story related to a specific part of their trailer life. Like charlie said I think the connectivity and the narrative aspects should be less around each question and more around general aspects which could relate to the different phases of the survey--like a story about how they got swindled when they bought it at the beginning... a story about their general wellbeing in the symptoms section... a larger life story in the demographic section. I think having narrative after each individual question will fragment to stories have have them loose power, not to mention have them be harder for users to navigate. I was really surprised in my research how much people wanted to be surveyed. they wanted their harm to be captured to prevent future harm, and then after that they would fill in the rest of the story. Which surprised me as an anthropologist who deals in stories. So that repeated trope might be usefully harnessed by having paragraph comments at the end of each section. [get ready for some serious doses of caps lock. most people seem to lean on it when sending me their stories]

Also we should maybe also keep in mind that "suggest a question" function. so they can tell us not only what their story is but what they thing is significant for inclusion in the survey.

[jywarren]

We should have a status box (can defer to Phase II if too much work) with how many questions you've answered, tips or suggestions for next steps.

[jywarren]

OK, so reading Nick's last comment, this implies a different structure for the survey - where more than 1 question is shown at a time. We could either show only one at a time, and at the end of each section, prompt with "Describe your experiences related to the last few questions to others" or something a bit less dry, but which makes it clear that you are telling a story which other people will read.

Whether we show questions and ask for a story on a single page (which would make for a more complex page layout, but fewer clicks, perhaps) or show things one at a time, we'd now be talking about "Sections" of the questionnaire, which would contain both quantitative form-questions and narrative-free-responses. This adds one extra table to the database. I wonder if we could simply order all the questions 1,2,3,4 etc and have a "question type" which is "multiple choice" or "text" or "long text" like Google Forms does, though. Would we then be able to skip the extra table and just have a numerical "section_id" numbered 1,2,3 and/or a "section name" like "Around your home", "Your health", etc etc.?

[jywarren]

Another issue with displaying multiple questions is that in the current design, we show people graphs of others' responses immediately after they answer a question, then lead them to the next question. If we batch them into "sections", do we show several graphs once they finish the whole section? Or does a graph appear under each question as it's answered? My instinct is to display each question on its own page.

[shapironick]

I think we could leave the questions as is--one on each page-- as that is also part of our public/private filtering which could get complicated.

So it would just be the graphs and the narrative section that would be in one page. It would be nice to have the graphs together on one page for each section so people could see how they stack up in the multiple ways across a section. That would increase the likelihood that they would see trends, see a narrative emerging from their answers (hmm I'm assuming here that the narrative part would come after the graphs but I guess that's not necessarily the case), and contribute a connection or observations that we couldn't have anticipated.

So yeah. each question on its own page. and then perhaps a section graph/summary page and a narrative prompt.

I've divided the questions up on the long survey--are we good to go on that one?-- and will work on putting the question type together in a list.

[mmnoo]

You can find a working draft of the narratives interface here. (it will prompt you to log in)

There are a few obvious design issues with it. The most striking being that all of the data is displayed separately. The way it is now, might integrate well though with the questionnaire, so that while a user is answering questions, they can click a button submit and view other's answers in addition to having a submit and skip button. By doing this, they can see other peoples answers as they go through the questionnaire. (might help encourage them to complete the questionnaire if they get curiosity rewards along the way)

Questions

• 'long-form narratives' = comments?
  • If we have commenting, maybe showing questions one at a time or in sets might be better than showing a page with all of the graphs (we can do both of course). Commenting might get lost on a page with so many graphs.
• For comments, does it make sense to have the ability to comment on the narratives page (probably need a lay-person-friendly name for that page) or integrate it into the questionnaire somehow (or both)?
• it is possible to prevent a user from seeing a graph/narrative to a question they have not answered. Is that desired? This could be useful and serve to prod them to go back and answer skipped questions. *The narratives page isnt well defined in the work plan for phase I, and there werent many hours allocated for it. Is it safe to assume that this iteration of the narratives page suffices for phase I? (There is about an hour left in the phase I time budget for tweaks)

[jywarren]

Repeating from #4,

Im thinking the long responses/comments after each section could be an opportunity to break up the questionnaire and offer a different kind of activity in between... Like the puppies but more like

Q, Q, Q, Q, discuss and story telling with fellow contributors, Q, Q, Q...

Therefore we could batch the graphs by question group. On Jul 31, 2014 3:50 PM, "Melissa" notifications@github.com wrote:

You can find a working draft of the narratives interface here http://dev.wherewebreathe.org:3000/narratives. (it will prompt you to log in)

There are a few obvious design issues with it. The most striking being that all of the data is displayed separately. The way it is now, might integrate well though with the questionnaire, so that while a user is answering questions, they can click a button submit and view other's answers in addition to having a submit and skip button. By doing this, they can see other peoples answers as they go through the questionnaire. (might help encourage them to complete the questionnaire if they get curiosity rewards along the way)

Questions

• 'long-form narratives' = comments?
  • If we have commenting, maybe showing questions one at a time or in sets might be better than showing a page with all of the graphs (we can do both of course). Commenting might get lost on a page with so many graphs.
  • For comments, does it make sense to have the ability to comment on the narratives page (probably need a lay-person-friendly name for that page) or integrate it into the questionnaire somehow (or both)?
• it is possible to prevent a user from seeing a graph/narrative to a question they have not answered. Is that desired? This could be useful and serve to prod them to go back and answer skipped questions. *The narratives page isnt well defined in the work plan for phase I, and there werent many hours allocated for it. Is it safe to assume that this iteration of the narratives page suffices for phase I? (There is about an hour left in the phase I time budget for tweaks)

— Reply to this email directly or view it on GitHub https://github.com/publiclab/wherewebreathe/issues/1#issuecomment-50816702 .

[jywarren]

Sort of like this.

[shapironick]

Looks good! Maybe also ask steve wing his thoughts on this while y'all are meeting today? This gets back into the questions @novakn posed in #4. Perhaps the narrative sections could be inbetween the categories of symptoms (fatigue/cognitive, respiratory, skin, other) and then to avoid contaminated answers about other categories of symptoms we could have a warning: "In this section please only discuss respiratory (for example) symptoms. Other kinds of symptoms will be discussed elsewhere, and its very important for our research methods that we keep them separate." We could eventually auto-detect for words that relate to other symptoms and send additional warnings. Or could we just have the symptoms section be a marathon section, as its particularly sensitive, and have breaks between sections like housing type, demographics, mitigation, other indicators of IAQ etc? The latter is probably the safer bet.

would that work, nicole? I know you were worried about cueing.

[jywarren]

The latter makes a lot of sense and would be much easier to implement and maintain -- we don't want to be policing or censoring submissions, so let's structure it so that it's OK to talk about whatever you like.

[mmnoo]

I like where this is all going.

[mmnoo]

In response to would displaying narratives (comments/discussion) at the end of each question group be a problem? from #4:

There probably isnt time for it in Phase I, since it would require a rethink and an addition of some complexity to the questionnaire and narratives logic, but it sounds doable for phase II. Its mostly a matter of budget and expense for you guys. I'm happy to do it.

[jywarren]

Melissa - forgive me for not following completely, but could you summarize what you are currently doing? I want to understand the scope of the changes we are proposing compared to the way you are currently implementing it.

Sorry to make you regurgitate things you originally got from us, but the discussion's been scattered across github a bit and I just want to hear it from you so I know we're synced. If you've already laid it out thoroughly, please just link me to it! Thanks a lot, and this'll help us decide how to structure any proposed changes.

[mmnoo]

Im currently working on exporting the data.

Next up is a modal welcome box, then I am planning on wrapping up some loose ends such as making nightly scripts to clear out unverified newusers and password reset tokens (you can see some loose ends I am tracking here. That doc isnt up to speed with GitHub, but is more of just some personal tracking on my side of things. )

Then I am planning on swinging around and working on the GitHub issues that have come up, and hopefully get a head start on some phase II stuff with some (hopefully) remaining hours.

Does that answer your question?

[jywarren]

Oh, sorry, i meant what you're currently doing with the sequence of questions and the narrative -- i.e. when are you directed to the narrative page for the questions you've answered, such that my suggestion of doing that after each question group is completed would disrupt the logic and/or create more work for you? I just want to be considered in our planning. Thank you!

[mmnoo]

Oh, I see.

Right now it is just temporarily a menu item. Its not integrated into the questionnaire at all. I dont think it will be crazy technically to implement most of what has been suggested above, I just have no time left allocated specifically to that task. Although, until now, I havent been considering all of the extra time we actually will probably have remaining in phase I from the margin of error that we incorporated in to the time budget.

If there is extra time in phase I, which I think there will be now that I am getting close to having checked off all of the phase I asks, and am considering the margin of error we have to work with, would this be the first thing you want me to work on?

Also Monday is a holiday here, so I probably wont be super responsive to emails/messages

[jywarren]

OK, yes, I think figuring out how the questions lead you to the forums would be the first thing to tackle with any extra time. But we do want to reserve time at the end for design polishing and other finalizing issues. Does that sound reasonable?

[mmnoo]

Yes, totally.

[jywarren]

OK, I'm going to close this, and reopen a new one for the "Forum" based on question group.