Voter Power update

I thought I should post an update on the ongoing issues with http://www.voterpower.org.uk/ I stumbled across on  Friday my original plan for that post was to lightly touch on electoral reform by showing how a system already in use in some elections in this country could be used for Westminster elections. This is a hugely complex set of topics which I am unlikely to be able to cover in sufficient depth on this blog but I thought the concepts were worth mentioning.

Voter Power was launched with quite a bit of publicity and seemed a good way of highlighting alternative voting systems online, especially in comparison with the original report document which understandably lacks mass market appeal. It was therefore a huge disappointment that it had obvious glaring errors. It had just been something subtle like rounding errors in the numbers or a fault with the visualisations then that wouldn’t be so much of a problem, but the site is currently peddling serious misinformation. The fact that the fault in the data shows on the front page in the two prominent lists of top and bottom ten show that the quality control standards here don’t seem to include having a quick glance over the results let alone a full test suite.

Why is this an issue? This is the so called Internet Election, where social media and networking, blogs and mashups are as important to getting information out as door-stepping and leaflets. The power is there for everyone to be able to put forward their view of the country and how it should be run and provide visualisations on how this compares with reality and the viewpoints of candidates. This sort of increase in information and comparison tools is, in my opinion, a hugely valuable way of increasing engagement and hopefully turn out, so high profile errors like this set the whole process back as they undermine trust,

Martin has been in touch but is away for the weekend so can’t look at the issues that are marring the site until he gets back.

Curioser and curioser

Looks like there is a one seat difference in the bottom 10 list on voterpower.org.uk, also involving Sheffield.

Top 10
Constituency	VOTE POWER INDEX
ARFON	1.490
CEREDIGION	1.390
CLWYD WEST	1.387
BELFAST SOUTH	1.228
WIRRAL WEST	1.200
PRESELI PEMBROKESHIRE	1.180
SITTINGBOURNE & SHEPPEY	1.176
NA H-EILEANAN AN IAR	1.174
ABERCONWY	1.160
EDINBURGH SOUTH	1.154
Bottom 10
Constituency	VOTE POWER INDEX
KNOWSLEY	0.001
EASINGTON	0.001
BARNSLEY EAST	0.001
SHEFFIELD BRIGHTSIDE & HILLSBOROUGH	0.001
BOOTLE	0.002
LIVERPOOL WALTON	0.002
WASHINGTON & SUNDERLAND WEST	0.002
BELFAST WEST	0.003
COATBRIDGE, CHRYSTON & BELLSHILL	0.003
NORMANTON, PONTEFRACT & CASTLEFORD	0.003

Voter Power and Sheffield South East

The more I look at the Sheffield South East inconsistency I spotted when doing a post about Voter Power this morning the more it looks like a software or data glitch. Looking at the spreadsheet published by the Guardian, which if I am reading right should be the same back end data, then Clive Betts has a safe seat. So something seems to have gone wrong in the data import or the software.

Other sources have Sheffield South East as safe and Magnus from the ERS said in reply to my question by email:

Our determination is an educated guess based on previous results, local election trends, incumbency and demographic trends. Sheffield South East and its predecessor seats have been held by Labour since the 1930s, and the notional 2005 results taking into account boundary changes gives the Labour incumbent Clive Betts a majority of 42%. In my view this is a pretty safe seat.

I am still awaiting a reply from voterpower.org.uk

Voter Power

Today’s political chatter is about a report from the new economics foundation called spoiled ballot, "why less than three per cent have a fair share of power in Britain." Actually very little of it is about the report, most of it is about the Voter Power Index which is a very web 2.0 friendly interface to the data including the compulsory video explaining what the site is about for those who don’t like to read.

In this methodology Sheffield South East turns out to be one of the most powerful in the country in marked contrast to the Electoral Reform Society data I reported Wednesday which makes me worry about the methodology and I’m thinking of emailing the authors to ask them about this. Their reports for the rest of our constituencies are Penistone & Stocksbridge, Sheffield Central Sheffield Brightside & Hillsborough, Sheffield Hallam and Sheffield Heeley.

Following on from this I decided to have a play with the data from the last general election to see what would happen if Sheffield was a super constituency elected using the method used by the UK in the European Parliament elections called the d'Hondt method. You’ll note that the nef report doesn’t have this as an option they suggest using the same mega constituencies that we do currently use for the EU elections for party list elections. They do have a suggestion that could see Sheffield elect as one, but it is based on Single transferable vote which would be impossible to do the kind of back of a napkin maths for the table below involves as it would mean guessing at voter’s second, third, etc choice intentions.

In my fictional system instead of 5 labour MPs and 1 LibDem we would have 3 Labour, 2 LibDem and a Tory (the figures in red represent an elected person). Of course there are flaws with even this extrapolation such as Green supporters only being able to express their preference in half the old constituencies and voting method influences choice but I think the result is interesting none the less.

	Sheffield Central	Sheffield Attercliffe	Sheffield Brightside	Sheffield Heeley	Sheffield Hillsborough	Sheffield Hallam	Total	D'Hondt /2	D'Hondt /3	D'Hondt /4
Labour	14,950	22,250	16,876	18,405	23,477	5,110	101,068	50,534.0	33,689.33	25,267.00
Liberal Democrat	7,895	6,283	3,232	7,035	12,234	20,710	57,389	28,694.5	19,129.67	14,347.25
Conservative	3,094	5,329	2,205	4,987	6,890	12,028	34,533	17,266.5	11,511.00	8,633.25
Green	1,808	-	-	1,312	-	1,331	4,451	2,225.5	1,483.67	1,112.75
Respect	1,284		-	-	-	-	1,284	642.0	428.00	321.00
BNP	539	1,477	1,537	1,314	2,010	469	7,346	3,673.0	2,448.67	1,836.50
UKIP	415	1,680	779	775	1,273	438	5,360	2,680.0	1,786.67	1,340.00
Socialist Alternative		-	-	265	-	-	265	132.5	88.33	66.25
Christian Peoples	-	-	-	-	-	441	441	220.5	147.00	110.25