Hi Alex,
Good plan. I believe that mySociety is intending something similar with
a questionnaire to be sent to each candidate. This ties in with the
Democracy Club election plan and MyNextMP. My impression of the
situation (not knowing the full details) is,
- Democracy Club will source local questions (or rather statements) from
volunteers - we've started doing this
- mySociety will source national questions from all interested parties
- This will then be combined into a questionnaire personalised for every
constituency's candidates (sourced from MyNextMP)
- Candidates will then reply with a ranking of agreeness, with free-text
explanations to their responses, and a rating of how important it is to them
- These answers will then be used in a find-your-candidate type exercise
as you described
- People can then publish their choice of candidate (or even to state
that they don't intend to vote, or that they're exercising the right to
a secret ballot) to Facebook/Twitter/whatever and advertise the
questionnaire to more people.
- Follow up elected MPs with their answers to see if they keep them.
This seems to be quite close to what you're proposing, clearly it's
desirable to avoid duplicated work. Maybe further discussions can go on
about how it's all going to happen and the details - like whether the
questionnaire should be posted to candidates in one go, or as more
questions are suggested they should be sent as a second questionnaire -
the latter might be a bit annoying to candidates and reduce the chance
of a response!
Tim
On 30/01/10 15:27, Alexander Hilton wrote:
Ben, Timothy
Some nationalist fundamentalists have about 1700 candidates listed on
their site, which would be a good first step for scraping
http://albionalliance.org.uk/
Plus they're xenophobes so i don't mind stealing the data they took so
much effot to compile.
I have a project I want to build using this data to try to break down
people's personal ties to political parties where they are based on
preconceptions rather than actual knowledge of the candidates.
1. there is a problem with the http://www.politicalcompass.org/
applications in that they average out qualitative opinions in a linear
(quantitative) manner. for example, I am anti nuclear weapons but pro
nuclear power. These left and right wing positions are averaged out to
define me as a centrist, but you could argue that I am centrist on
neither issue.
2. Therefore, assessments of candidates should be made, issue by
issue, and the result scored according to the importance to the voter.
3. So, the core of the project is questions, all framed as..."Would
you vote for ..........?" and the candidates get to answer yes, no,
not sure, and have a free text field.
4. Secondly, the users need to answer the same questions (without the
free text field) but with, in addition, a 0-5 ranking of how important
the issue is to them.
5. The questions are user generated but registered users have to
register their postcodes, therefore placing them in their relevant
constituencies.
6. Users can see which questions have been asked locally by other
users and which ones have been answered by local candidates. By
answering the questions themselves, they instantly return a ranking of
candidates, scoring the candidates according to how similar their
answers are to their own.
7. The candidates' score is calculated by adding the user's
"importance" rankings if they agree with the candidate or subtracting
them if they disagree.
8. Some questions of importance to the user may not have been asked of
the local candidates, or they may have been asked but not answered.
The user would need an opportunity to search a national list of
questions and apply them to their local candidates or to add to the
national list of questions - maybe along the lines of google moderator
http://moderator.appspot.com/
9. There is a micropage for each candidate listing their questions
answered, unanswered etc
PROBLEM - HOW DO YOU GET THE CANDIDATES TO ENGAGE?
There has to be a transaction of value to the candidates.
1. The users, regardless of who they are "most like" in opinion, get
to indicate on the site which candidate they are backing.
2. When questions are sent to candidates, they are not sent using the
users email, but are instead sent using an email address under the
system's domain. Candidates respond by clicking the yes link, no link
or not sure link - all unique links - and then get bounced to the page
where they can enter free text.
3. If a candidate signs up, they get added functions;
a. They get to answer questions in the national system before they are
put to them by local voters - in effect encouraging voters to ask the
questions they want to answer.
b. They get to add questions nationally - again helping to influence
the agenda.
c. They get to bulk-email (though without direct access to names and
email addresses) the voters who have indicated they will back them
locally.
d. They get to edit their micropage with pictures, contact details,
select from a choice of simple themes etc
OK - that's clearly a bitch to build. BUT - I'f I could acquire a few
thousand in sponsorship money, who would be up for it?
Would require...
1. heavy SQL knowledge
2. some great php skills (or something similar)
3. A sprinkling of magic ajax dust
Alex Hilton
07794 771 113
_______________________________________________
Mailing list [email protected]
Archive, settings, or unsubscribe:
https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public
