Full candidate gender breakdown by party

May 13th, 2010

We’ve crowdsourced the genders of the candidates in our database.  Here’s a chart showing the candidate gender balance by party (just for parties with representation in the Commons).  The average across all candidates was 22%:

Below is the data for all parties with 10 or more candidates. The obvious pattern is a left-right/female-male correlation. Is this because of positive discrimination tactics in left-wing parties, a gender pattern reflected in actual party membership, or something else?

Party Name female male Percent female
Scottish Green Party 12 8 60%
Alliance – Alliance Party of Northern Ireland 6 12 33%
Trade Unionist and Socialist Coalition 11 23 32%
Labour Party 198 433 31%
Green Party 98 217 31%
Scottish National Party 17 42 29%
Christian Party – Proclaiming Christ’s Lordship 20 51 28%
SDLP (Social Democratic & Labour Party) 5 13 28%
Respect – The Unity Coalition 3 8 27%
Socialist Labour Party 6 17 26%
Conservative Party 155 476 25%
Ulster Conservatives and Unionists 4 13 24%
Liberal Democrats 137 494 22%
Plaid Cymru – Party of Wales 8 32 20%
Scottish Socialist Party 2 8 20%
Christian Peoples Alliance 3 14 18%
Sinn Féin 3 14 18%
British National Party 58 279 17%
UK Independence Party – UKIP 87 470 16%
Independent 35 306 10%
Traditional Unionist Voice – TUV 1 9 10%
English Democrats Party 10 97 9%
Official Monster Raving Loony Party 27 0%
National Front 17 0%
Democratic Unionist Party – D.U.P. 16 0%
Total Result 899 3251 22%

Posted in Uncategorized | No Comments »

What can you do with our data?

May 8th, 2010

Lots of people have  told us that our candidate survey tool really helped them decide who to vote for. We’re truly delighted about this.  The next step is  to learn what we can from the data, partly for fun, and partly to see what we can learn, in order to an even better job next time.

All our raw data is now online. Based on this, we’ve done a preliminary analysis of the survey responses, showing average answers broken down by party. Some patterns immediately emerge; one that jumped out at us straight away was the extraordinary distance that Labour party candidates are from everyone else when it comes to the statement “there are too many CCTV cameras in Britain”:

party agreement %
Labour Party 35
Independent 67
Conservative Party 68
Liberal Democrats 75
Green Party 78
UK Independence Party – UKIP 78
British National Party 82

We’ve got a basic grasp of stats, but we’re not statisticians.  So we’re really hoping other people can do interesting things with our data.  @tomatosquid has already made a multidimensional map of the average figures linked to above, which seems to show the closest pairings of parties are Ind/Con, Lib/Lab, Green/Lib, and UKIP/BNP.  (Remember this is based only on our questions; if we’d included a question about “voluntary repatriation” then no doubt UKIP and BNP would have been further apart, and if we’d included a question about ID cards, then Lib Dems might well have been closer to the Conservatives than Labour).

It’s also possible to cross reference our data with the biographical data available at YourNextMP; in particular, I’d like to crowdsource a complete gender dataset soon, so we can do interesting things with that.  I’d also like to add 2005 poll figures, so we can see if there’s a correlation between seat marginality and likelihood of a candidate answering our survey (about 50% of candidates we contacted by email answered in the end, but the distribution of answering rates varied enormously by party; if we can understand what correlates with survey answer rate, we can focus our attempts to get that figure higher next time).

What else can be done with our data?  How about comparing how gender or age compares with party membership as a predictor for survey answers; running a spell checker across the free-text part of the answers and finding out the best spellers; identifying which questions divide candidates the best, or which parties have the widest range of answers.  It’s also been suggested that a Principal Components Analysis might show us unexpected cross-party groupings. Please show us what can be done, and drop a comment here!

Posted in statistics | 5 Comments »

Lies, damn lies, and gender balance

April 30th, 2010

As more and more candidates fill out our survey, we also get more demographic detail about them. And as a commenter has already pointed out, we’re amassing a great deal of interesting data about the people who want your vote.

As a very simple taster of what’s to come, it seems1 that around 80% of candidates in this election are male. Here’s the breakdown of female candidates by party (just the ones we have enough data from to make some basic generalisations2):

party female candidates
Labour Party 32%
Green Party 30%
Liberal Democrats 18%
Conservative Party 13%
English Democrats Party 12%
UK Independence Party – UKIP 10%
Independent 9%
1We don’t know yet if the sample of demographic data we have is representative of the candidates as a whole.  This is one of the things we’ll address after the election.
2Actually, to be boringly honest, all we can currently say at a 95% level of confidence is that Labour and the Greens have significantly more female candidates than the Conservatives, English Democrats, and UKIP. This may be true, but it’s more boring than the misleading table above, which is why I put it in a footnote.

Tags: , ,
Posted in statistics, survey | 3 Comments »

Candidate survey: FAQ

April 29th, 2010

Our survey of candidates is finally live!  We’re getting thousands of people flooding in and reviewing the responses we’ve got so far.  It’s really exciting to have got this far, as part of a team of thousands of volunteers.  Here’s an FAQ about the survey.

Where are the answers?  I can’t see them!

For each statement, you first need to select what you think in order to reveal what the candidates had to say.

What does it mean if someone hasn’t answered?

In some cases, the candidate might not know about the survey, but in the majority of cases, they will have had an opportunity to reply.  We are doing everything we can to contact as many as possible:

We have email addresses for 80% of candidates.  Some of them will inevitably be incorrect; part of the project has been to try to find valid contact details.  That means we’ve maybe successfully emailed about 75% of them.  Candidates are very busy, so might not have noticed our emails; but we’re trying to make sure voters have reminded them in each constituency (you can see how many reminders they’ve been sent on our constituency listing page)

For the remainder, we’re sending old-fashioned letters to everyone we have postal addresses for.

So, for an individual constituency, you can’t be sure that a candidate definitely saw the survey request.  However, overall, the statistics showing responses by party are definitely of interest.

Why haven’t you got responses from the BNP?  Are you biased?

We are completely neutral about this project, and want as many candidates as possible to answer the survey.  Sadly, in the case of the BNP, it’s very hard to get their contact details — we’ve tried.  It’s also hard to get contact details for many Independents.  If you find any, or you’re from any party which we have poor email coverage for and can supply us with email addresses, please let our partner site YourNextMP know.

How did you come up with the questions?  Why haven’t you got one about the stuff that’s important to me?

Our volunteers helped come up with the local questions, and a balanced panel came up with the national questions.  Getting the balance of questions right was hard; we think we did a good job, but suggestions for how to do better next time are welcome!  There’s a full blog post about the process on the TheyWorkForYou blog.

I’m a candidate, and I think this is all a Labour/Tory/LibDem push-polling project

We can’t please everyone with our selection of questions, but we’ve tried to be balanced and open.  Some of the local issues sometimes come out unbalanced because we had to rely on volunteers, and in some constituencies only one volunteer came forward; so the local questions are sometimes only representative of a single person’s concerns.  We still think the answers can be interesting, though!

I’m a candidate, and I object to being forced into expressing my opinions in this format

To be sure, a multiple choice plus very short text box is a limiting format, and we understand it can be difficult to answer some of the questions as fully as you’d like.  However, we wanted to give voters an way of comparing what their candidates have to say which was easy to understand.  Voters are tired of spin and media management of messages, and this is a refreshing and new way to engage with them, openly and transparently, on a fair and equal footing with other candidates.

Also, we encourage you to change your mind about any of these questions in the future! One of the problems with our current political/media system is that it’s seen as a bad thing to change your mind. We think this is wrong, but equally, you should be prepared to defend your change of mind should it happen; which is why we think it’s good for you to go on the record. We might, for example, run the same survey again in a year to give you a chance to show how your opinions have changed.

Why are you asking questions about issues over which my MP will have no power?

We know that many of the questions are the responsibility of local government, or of devolved parliaments and assemblies.  However, we think that a candidate’s views on these issues give some insight into the person, rather than the party they represent — something we think voters want and deserve.

Posted in survey | 14 Comments »

The Scottish Green Party are ahead of the Tories by 56 points!

April 26th, 2010

…at least, in terms of answering our candidate survey.

We’ve now sent requests out to over 3,000 candidates for them to fill it out, and we’ve got responses from over a third of candidates already. Here’s a graph of the response rate so far of the top 6 parties (measured by numbers of survey invites sent, as of 26 April):

survey response rate bar chart, 26 april 2010

Click through to see a full list.

It’s really exciting to see the responses start rolling in.  Our survey is a mix of:

  • statements based on local issues gathered by our volunteers; and
  • statements from across the political spectrum, chosen by our balanced panel of experts (about whom we’ll write a little more when we get a chance)

Our goal is for voters to be able to type in a postcode, and compare what each of their candidates has to say about this selection of issues (we’ll be releasing that soon).

It’s great to see that most candidates are responding really positively, despite the huge work load many of them are under.  However, some are reluctant to commit their responses, many on the grounds that the survey “might not properly represent the nuances of specific party policies.”

We completely understand the point that a survey is necessarily and arbitrarily narrow, but we believe voters should get a chance to learn a little more about the people they’re voting for, rather than the parties they’re voting for, and this is the best way for them to do so.

It’s no surprise the smaller groups like the Greens, UKIP and Independents are responding more quickly: they have a greater incentive to fill it in, and a lower volume of emails to reply to.  However, what really stands out to me is that out of the big parties, one in three Lib Dem candidates and one in five Labour candidates have answered the survey so far, but only one in twenty Tories.  Come on, Conservatives, join in the fun!

If you want to help encourage your candidates to answer the survey, please sign up and log into Democracy Club!

Posted in candidates, statistics, survey | 7 Comments »

Play our new game, win a duck house

March 30th, 2010

Currently Homeless (photo by N.R)

I know, you probably don’t need a duck house. But that’s no reason not to have one, right?

If you join DemocracyClub right now, you’ll have the chance to play our new candidate researching game. As you play it, you’ll win duck points.

The person with the most duck points from this (and possibly other) election based games by midnight on election night will win this duck house*, donated by mySociety.

OK, we confess. The real reason to play this game is not actually to give your feathered friends some shelter (although we aren’t pretending – we are giving away a duck house ).

The real reason to play the game is that it is a playful way to help us gather contact information of candidates standing for election. Once we have that information we can send a survey to those candidates, and the resulting data can be added to TheyWorkForYou , our own election quiz, and any other sites that choose to use it.

So go on, have some fun, bring some transparency to the election, and help some ducks, all at the same time. How often do you get a three-in-one offer that good?

* In the strange eventuality that the winner doesn’t want the duck house, we’ll work out a good cause for it to go to.

Posted in candidates | 3 Comments »

Tesco and potholes

March 24th, 2010

What kinds of things our volunteers care about in their local areas?  We’ve been asked this a few times by journalists. I’ve noticed a few unusual ones like setting up re-enactments of the Boer War, but until now I’ve not had a proper look at the overall picture.

It’s all quite rough and ready, and I’d love to do some proper work on sorting and categorising the issues we’ve got. Until then, I’ve done a very quick exercise to get a feel for what people are reporting.  First, I made a word cloud:

Then, I came up with some categories and keywords and counted issues against these keywords.

The most obvious thing to note is the preponderance of transport-related issues, and particularly car-related ones: potholes, parking, bypasses, road tax evasion, congestion.  At the tail end of the distribution (not shown), there are things around broad band provision, immigration, the BBC, and foxhunting.  The most interesting minor issue that I’ve included here was supermarkets, with Tesco alone being singled out for 20 issues.

I was surprised that health comes quite low in the list, compared with (say) education.  Green issues came suprisingly high, as did housing and regeneration.

I’ve no particular analysis to offer at this stage; any comments welcome!

Posted in local issues | 8 Comments »