I've been looking through the local election results, and have again been struck by how difficult it is to get an accurate overall picture of how people voted, either within a particular local authority or across Wales as a whole.
The main reason for this is that some wards return more than one councillor, and each voter in these wards has as many votes as there are seats to be won. This makes it almost impossible to meaningfully analyse how people voted, because different voters have different numbers of votes.
I have wondered if the best way to get a clearer picture would be to say that instead of having more than one vote, voters in a four-seat ward have four "quarter-votes", that voters in a three-seat ward have three "third-votes", and that voters in a two-seat ward have two "half-votes". This wouldn't affect the individual election results in any way, but would mean that each voter had only one vote in total.
This is just a question, but has anyone ever seen an analysis of local election results done on this basis?
21 comments:
Why not just divide the total votes cast in a 3 seat ward by 3, 2 in a 2 etc. Rough and ready but it will give you the picture.
Err, thanks Glyn, but that's what I thought I'd suggested.
Do you know of anyone who's done it this way? Or does anyone want to try it for their local authority and see what it shows?
Divide the number of votes a candidate gets by the number of ballot papers (i.e. people who voted). That why you can see that candidate X has the support of 45% of the voters.
This figure isn't always readily available.
I've been looking for these as well, I think it's hopeless. Apparently the Lib Dem vote across the UK went up form 15% (in 2008) to 16%, which is a 'Disaster'. The Labour vote went up from 37% to 38%, in other words a 'Triumph'.
Grangetown, Cardiff
Ballot Papers: 4165 (i.e. turnout)
AWAN PC 1467 35.2%
BAILEY Con 308 7.4%
BARKER Green 205 4.9%
BATE Christian 84 2.0%
BEST Green 215 5.2%
GOVIER Lab 1812 43.5%
HARDING LD 489 11.7%
KAEHNE Con 266 6.4%
LOMAX Lab 1709 41.0%
MONTEMAGGI LD 505 12.1%
MORGAN LD 501 12.0%
NEWNHAM Communist 77 1.8%
NICHOLAS PC 1391 33.4%
ORAM Con 251 6.0%
SATTAR PC 1511 36.3%
THORNE Lab 1704 40.9%
Expressed as % of people voting.
"I've been looking through the local election results, and have again been struck by how difficult it is to get an accurate overall picture of how people voted, either within a particular local authority or across Wales as a whole."......
Across Wales? lots for Labour. Within local authorities...not many for Plaid.
Does that help?
I'm bemused by how little we hear about the shambolic FPTP voting system. No-one provides us with information on whether the distribution of seats reflects with accuracy the voting of electors. We don't even appear to know the overall shares of votes in Wales for each party.
I did a back of envelope estimate for one county based on total votes (so not corrected for wards with multiple votes). Labour received some 52% of the votes for which they were rewarded with a luxurious 69% of the seats.
Here again, we lag behind Scotland with their STV system. In Glasgow in 2003, Labour took 71out of 79 seats under FPTP. This year they won just 44 under STV.
Never mind who won or lost, let us at least have a fair and democratic voting system that gives us what we voted for.
Why does Leanne Wood duck out of press conferences?
http://www.bbc.co.uk/blogs/thereporters/vaughanroderick/2012/05/dydw_i_ddim_yn_un.html
Anon 17:19- who cares?
Anon 17:56,
journalists, constituents, prospective Plaid voters, anyone with an interest in the opinions of our elected representatives.
Anon 17:56
Did you not follow the 2008 US vice-presidential contest?
The problem with counting 1/2, 1/3 or 1/4 votes (as appropriate) is voters who plump for 1, 2 or 3 candidates where they could vote for more. Do we then say that they abstain or even spoil their ballots for the other 1/2 or 1/3 or 1/4. Or, do we rather say that this a consequence of of our system where voting for more than one candidate can actually count against the person we really want to see elected?
It is probably possible to work out some sort of algorithm based on the total number of ballots in the ward (B), the total number of preferences expressed (V) and the number of votes by party (P) expressed by the number of candidates the party has fielded (C). A fifth factor would be the number of seats (S).
So, who is the psephologist who is willing to posit a useful relationship between variables B V P C and S? I'll buy you a Mars bar if you can solve MH's quandry!
Anon 20:12. I don't really appreciate the sentiment behind making anonymous criticisms of Leanne Wood on poor grounds.
I care about Vaughan Roderick's blog. As do you. Most people don't. Most people will never hear of it.
Vaughan's blog is important in terms of shaping opinion among journalists, but only as important as, say, Syniadau or John Dixon.
You're not supposed to 'appreciate the sentiment', you're supposed recognise a problem. You would sooner brush it under the carpet.
There is only one sensible voting system for multi-member wards, and that is STV.
I voted for 3 Plaid and one green, I would have liked more scope in a wards where there were at least 20 candidates for 4 seats.
Back to the question in hand...
Isn't it simply:
Total votes available (overall number of ballot papers [turnout] x seats available)
Divided by total votes obtained by each party (the sum of their candidates' votes) or independents (the sum of independent votes, aligned or non-aligned for example)
Equals... share of the vote.
I accept that this is just the mathmatical reverse of your proposal Syniadau, but is perhaps easier to explain (?).
Either way, I think it is a mathmatically credible measure and of course more informative of underlying party support than the FPTP system allows (as we all know from Westminster elections).
Phil Davies
I'd add that I think it immaterial if a voter
a) splits his/her vote (in a multi-seat ward) - this will be reflected in the aggregated totals for the parties
b) exercises his/her right to vote for only 1 or 2 candidates in a multi-seat ward - again, this will be reflected in the aggregated party totals
Putting voter motive aside for the above two options, the end result is still a good (and only empirical) measure of overall (and underlying) support for the parties/independents
Phil Davies
Using the example data provided above for Grangetown. The result would be:
Total available votes (4165 x 3 seats) = 12495
Lab 5225 42%
PC 4369 35%
LD 1495 12%
Con 825 7%
Green 420 3%
Christian 84 1%
Communist 77 1%
Clearly if a party does not field all three candidates (as was the case here for the bottom 4 parties) their overall potential vote is reduced as an elector "might" have voted for them. But we do not know. And this is not an opinion poll. It is the result of a real election.
This method can easily be aggregated up for whole council and national picture.
Or have I missed something?
Phil Davies
So, in the Grangetown example...
Labour won all three seats in the ward on FPTP (that is 100% of the seats) on 42% of the vote...
Which leads neatly into the discussion raised by others on this trail about the fairness, or not, of this system.
PD
Not strictly related, but a policy paper was submitted to Caernarfon Plaid branch last year urging them to offer prospective candidates to all wards in Gwynedd that couldn't muster a Plaid candidate. The vote at branch refused to go with the policy motion (I have a pdf copy if you let me know where to send/post it).
The implication? If the policy had been adopted, Plaid would have taken Bryncrug unopposed and by now would be in sole control of Gwynedd.
Yes, please send it to me at this address.
I know that Ceredigion had problems like this, too. In the end Plaid didn't have a candidate in 7 out of 40 wards. And in Sir Gâr we didn't put up a candidate in 6 of the 58 wards. These are meant to be our heartlands, and I think we should put up a candidate in every seat unless there is a very good reason not to (for example not to split the vote of a co-operative independent). It might not yield fruit now, but it will help establish our profile for future elections.
Post a Comment