Communicate show an 11 point Tory lead
Communicate Research’s monthly poll for the Independent has topline voting intentions, with changes from last month, of CON 40%(+6), LAB 29%(nc), LDEM 17%(-4). The poll was conducted between the 23rd and 25th February.
On a uniform swing the eleven point lead would be enough to deliver a workable majority to the Conservatives – only the second poll to do so since the election (the 13 point ICM lead which received so much attention last week was a hypothetical question on how people would vote with Brown as leader, not a straight voting intention question).
Before Conservative supporters get too excited, it is worth pointing out that Communicate Research’s polls are not politically weighted by past vote. This has two effects – often it means that the polls are slightly more favourable to Labour than politically weighted polls like ICM’s and Populus’s (though clearly not in this particular case), secondly political weighting dampens down volatility since every sample will have of the same proportion of people who say they voted Labour in 2005, the same proportion of people who say they voted Conservative in 2005 and so on.
When averaged over time non-politically weighted voting intention polls like Communicate’s or MORI’s tend to produce figures that are broadly in line with the figures that are produced by companies that do use political weighting – however, individual polls tend to be far more erratic – you only have to look at the 6 point leap in the Conservative vote in this month’s poll. Until we start seeing more polls showing an eleven point Tory lead it’s probably best to put this one down as an outlier, rather than a Conservative surge.
Filed under: Communicate, Voting Intention
When we “did” graphs at school 45 years ago we were instructed to “smooth”. That holds good, doesn’t it? But it also, as 45 years ago, shows trends.
‘Until we start seeing more polls showing an eleven point Tory lead it’s probably best to put this one down as an outlier, rather than a Conservative surge. ‘
The figures are not far off the Guardian/ICM poll of 20 Feb.
Anthony – thanks for an excellent analysis as usual.
I was just wondering whether you think it’s possible that some people’s past vote recall changes with their present voting intention? i.e. when voting for party X becomes more fashionable and someone decides they’ll vote for them at the next election, mightn’t they be more likely to say – through false recall / shy Tory / bashful Blairite or whatever – that they voted for that party at the last election as well?
I suppose what I’m asking is, do ICM / Populus run the risk of artificially depressing the Conservative vote share because the past vote weighting changes along the same trend as the their present voting intention?
Anthony,
One possible effect of these “shock polls’ is that they actually influence how people see things. is there any evidence (or indeed way of telling), if subsequent polls tend towards the results of a “Rogue poll”.
How would we know if a poll like this was heralding a change or influencing it? Has anyone done any research on this?
Peter.
Interesting. The WMA is now 38:31:18 so the C lead is back to 7 which it was in September. However it’s true that Communicate’s polls have tended to understate the C lead (by 2.6+/-4, sample size only 3 though) and it may be that we are getting people who are switching to C and have “forgotten” that the voted Labour in the last election, which would tend to cause the other polls to understate C support. The next couple of polls will be *very* interesting.
The message from the polls seems to be -as others before me have pointed out-that little has changed since last May. Public perception of the way the results have gone for the parties at the various elections due this May could thus be crucial. If the Tories do not sustain a sense of overall momentum then with internal sniping at Cameron likely to come to the fore they may find themselves facing the threat of a drift in their support both to the smaller right wing parties and the Lib Dems . As there is absolutely no sign of any Tory revival in Scotland then they will need to do exceptionally well in England and Wales in May – which begs the question-when is somebody ie the ‘ Western Mail’ going to conduct a Welsh poll?
Peter Cairns. I am sure you are right about “dynamic drag”. On a practical basis I have found that electoral movement associated with poll movements or media publicity can have a very potent effect. “Me too” is with us in politics as in all other walks of life. 1997 was much worse for some Tories on the ground than it seemed to us it was going to be. I am sure that was “drag” from the media. I am not saying they set out to do this, it was what they had to report.
Anthony
As a long time passive observer of this most exellent website, it strikes me that events an polls wills clearly and obviously being related to one another, are not necessarily instantaneously so.
Polls taken in the immediate aftermath of an event do not always seem to reflect the event (Even where the field work takes place immediately after the event)
There seems to be a lag of 14-21 days for the polls to “absorb” events.
I wonder whether (for example) the sight of paedophiles on the six o’clock news admtting they were lucky to not be in jail is only now reflecting itself in some low Labour ratings in the most recent polls?
We seem to be getting into a “Kinnock in 1990″ situation with the polls now, which is interesting but doesn’t necessarily indicate a Tory victory.
In order to prove the validity of these polls, the Tories will need to poll at least 40% in the local elections in May, I would think.
Yet another poll showing UKIP only getting 1%.
I asked on another thread if Yougov were reviewing the makeup of their responses as they get UKIP results in the 5% range which is 4 or 5 times the level other polling firms are achieving.
So far no one from You Gov seems to have responded to my question.
I also posted evidence of UKIP sites encouraging their people to join Yougov or is it Youkipgov?
What is the supposed margin of error in this poll? I thought normal MoE’s were about 3% so a 6% jump being “just” an outlier is surely possible but unlikely? Possibly a mix of both an increase and an outlier.
I find it intriguing that recently the two ‘left-wing’ papers (Guardian and Indy) are getting polls supporting a big Tory lead, while the more ‘right-wing’ papers (Times and Telegraph) are showing a more even contest with a more moderate Tory lead.
M of E for a 1,000 sample is 3% but only if the full 1,000 responses are taken into account . The detailed figures on the website show that in fact the sample size is really only 647 which puts the M of E up to nearly 4% . This applies to ICM polls also with a notional sample size of 1,000 .
Catherine – yes, that probably is one of the factors and this is exactly the reason why MORI (and presumably Communicate) don’t think past vote is suitable for weighting.
The reasons behind false recall and many and not particularly well understood – I think the main reasons are people being embarrassed to admit not voting and instead saying they voted for who they would have voted for, and tactical voters saying they voted for the party they really supported, not the one they actually voted for. There is also the case of people adjusting their past voting intentions to how they would vote now though. What we do know is that in practice, while recalled vote does change slowly over time – and ICM and Populus gradually move their targets to account for this – it also seems to be pretty stable…at least so far.
Because ICM and Populus have dynamic targets for their weighted vote share, if recalled Conservative vote was gradually creeping up then their weights would also slowly creep up – so yes, that could potentially be a problem, but I unless it was a very rapid change ICM and Populus’s systems could cope with it.
Peter – not so blatantly. That said, when polls get large scale media attention they do themselves have an impact on public opinion and, therefore, other polls. As you know, we do daily tracking of politicians and parties on BrandIndex and there have been occassions when there has been an impact on people and parties scores after coverage of a poll (though I would define it as being an improvement in their image after lots of press coverage of them being successful and doing well, but that came on the back of a poll. It is an indirect effect via the media).
Simon – Thanks, and thanks for your maiden comment! Lynton Crosby has a theory that it take two to three weeks for an event to sink into the public mind before it shows up in a poll. I only agree to an extent – things like new policies probably do take a while to sink in, but a lot of things have a very immediate effect. I suspect that who has a lot of publicity on that day’s news has an immediate and short term effect…but that events also have a deeper, longer term effect that could take a while to settle down.
Harry – the difference in UKIP and BNP scores (and Green scores for that matter) is probably largely a mode effect from unwillingness of people to admit to supporting fringe parties, particularly those perceived as extreme, to human interviews and comparatively willingness to admit it to an anonymous computer screen.
It isn’t possible to pack the YouGov panel in that way – largely because the majority of people on the panel are not recruited through the open website and the answers given by people pro-actively recruited and recruited through the website are checked for discrepancies – if people who joined through the website were 5 times more likely than other people to vote UKIP we’d notice.
There are also various other anti-fraud measures in place. The bottom line though it that it wouldn’t be feasible mathematically – the panel is of 140,000 people, there are 2,000 in a poll, and the polls are rounded to 1 decimal place. In other words – if you wanted to move a poll by a single percentage point you’d need 20 people in that poll, and to get that then on average you’d need to have 14,000 stooges on the overall panel. If a couple of hundred people joined the panel and behaved in an unusual fashion our panel manager would notice it. If 14,000 suddenly joined…it wouldn’t take a rocket scientist to spot it.
Anthony,
Apart from the obvious issue of Cost, (£70,000 for 140,000 at 50p a time) has YouGov ever attempted a really large sample poll, just to see how it would compare.
Equally I would assume in terms of proportionality that in Scotland membership is probably in the region of 12,000. I am sure that the larger parties could put together 1,000 people if they tried.
Any chance of a 10,000 people poll for Holyrood on April 1st….
Peter.
Anthony: Do the various pollsters have different response ratio’s and how do they handle that?
I recall in an episode of the West Wing they said something like that when polling they only normally got 25% to answer the phone and then answer the questions, so they had to call 4x as many people to get the desired number of responses. Is that the way it works? So if ICM, Populous etc get people who say “go away I’m busy” etc, do they call another person to make up for it?
How about Yougov? You say that you poll 2,000 people – is that 2,000 people who actually answer the questions, or 2,000 people you invite to answer the questions by email? I guess many on the panel don’t necessarily answer every poll they get sent, but unlike a phone call when you know someone hasn’t responded how do you know an emailled person won’t and handle that?
The last Yougov poll sample size was 2,292 but actual responses excluding don’t knows etc seems to have been around 1,900 .
Andy, not wishing to pile even greater burdens on the Tories in May (!), but the Conservative share in local elections tends to exceed their national opinion poll rating – conversely Labour’s tends to lag; both by around to 5%. So if the Tories are to convince us that their national share is at 40% (39%, after all, being what they managed last year), they need to put in something like a 45% share – at least in the English elections.
I personally am not sure that this difference between actual local votes and national polls is about polls being inaccurate – because an element of local voting is genuinely local (and people’s perceptions of which parties run councils better differ from their perceptions of who is best for government), whereas polls measure the latter.
I think (and this is entirely speculative, but based on observation) that Labour has a systemic problem in council elections in that:
a) I think people by and large are less convinced that Labour runs town halls as well as other parties (possibly a throwback to the terrible press loony left councils got in the 1980s) and
b) A greater proportion of Labour voters tend to regard themselves having done their duty to vote by turning out in general elections, and it’s much harder to mobilise them for council elections – Labour voters are also far less likely to appreciate that when they vote at General Elections they’re not electing councillors as well – the number who think one vote elects all their representatives is extraordinary.
Finally, can I just say: yet again, ridiculous Lib Dem ricocheting!
Anthony, I don’t have a problem with the idea of “shy” UKIP and BNP voters (or Respect voters, come to that), but “shy” Greens is a new one on me. Apart from Cornerstone and (some) points right, I don’t think they’re regarded as “extreme”, or that the perform better in real elections than their opinion poll performance, where measured, would indicate.
I shall of course defer to your more expert opinion if I have to
with conservatives now showing consistent large leads,it is clearly going to be a while before the trendy lefties get the hang of this!
with the country in such a mess and the police preoccupied catching people on mobile phones in cars,whilst real crime soars,these polls should not be a shock.
I am attempting to draw comparisons with when Mrs.T had been PM as long as Tony Blair- early February 1989.
At this point,the Tories still had a bare opinion poll lead (1-3%),having won by an 11.8 % marging of the popular vote in 1987.
Blair won by 3.0% in 2005;an average of the last three polls (using comparable Brown v Cameron parameters),these three polls yield am 11.0 % lead.
Therefore,the mean 2 party swing is somewhat greater for Tony Blair at this stage than for Mrs.T in 1989.
Extropolating the raw poll data into constitencies again shows an advantage for Cameron.The 3 polls averaged in the last week,using the lists of marginals on this website,would give 348 Conservatives,240 Labour,34 Lib Dems,28 others -an overall majority of 48.
In February 1989,having won a 100 majority in 1987,Mrs.T ’s Conservatives would still (just) have been the largest party in a hung parliament scenario.
Finally,looking back to 1989,Mrs .Thatcher’s poll rating deteriorated rapidly as 1989 progressed -rapidly rising mortgage rates,the approach of the poll tax.
Whilst,as a home owner,I trust this years rate movements will be far more modest,I do see things ‘getting worse before they get better,for the present Labour administration
Innocent – can’t say I’ve ever taken the trouble to compare their results with real results! I was thinking more that people can be embarrassed to say they are voting for any minority fringe party – some people like to run with the crowd! (Though obviously even more so if their party of choice is both a fringe party and an extremist party).
Philip – response rates. How pollsters handle it comes down to whether they are based on quota sampling or probability sampling.
Quota sampling is based on the theory that a sample is a little model of society, if you get the right number of men and women, rich and poor, young and old and so on, then your sample will reflect reality. MORI use quota sampling for their face-to-face polls – if their interviewer has a quota of 4 working class women to fill, they will keep on knocking on doors till they find 4 working class women – if people are out they’ll go to the next door. No idea what the response rate is for face-to-face polls.
Probability sampling is different. Here non-contact bias does matter. Probability sampling is based on the theory that, if you select your sample in a truly random way it will accurately reflect the population it is drawn from, and will do so in all sorts of ways that a quota sampling company couldn’t possibly draw up quotas for. In this case, non-contact or non-response bias buggers things up, because suddenly the sample is no longer a truly random sample (obviously in practice it isn’t anyway, becase not everyone has a phone, hence weighting). For that reason when Populus or ICM phone someone and don’t get an answer, they go to great lengths to go back and contact them later. I think over two days they will try ringing a number about 8 times in an attempt to get an answer. Response rates for phone polls are something like 1 in 6.
YouGov’s method is more like quota sampling than random sampling – though obviously there aren’t interviewers searching for particular demographics to fill an actual quota – the orginal invitation list is computer generated to include the correct proportions of men, women, young, old, etc, etc. The 2,000 people is the target number for a poll – more invitations than that will go out, and the number invited compared to the number required is based on past experience of how fast different groups respond. Because YouGov’s emails are going out to people who’ve willingly signed up for a polling panel, not unsuspecting members of the public, the response rate is far higher. The number of invites tends to err towards generosity, better to have too many responses than too few – as Mark says above, the last YouGov poll for the Telegraph had 2,292 responses, so 292 more than needed. If a poll does fall short, then obviously an email reminder can go to those who haven’t replied yet, or if necessary a top up can be drawn.
Andy (and Adam at 6:54),
There is a fundamental difficulty in comparing local elections with opinion polls, not only because people may genuinely vote differently in local elections, but also because (a) elections don’t occur across the whole country and (b) not all local elections are contested.
This year, for example, in England there will be local coucil elections in all Metropolitan districts, some Unitaries, and most, but by no means all, shire districts. No elections in London, or large parts of shire counties. The inclusion of metropolitan & unitary districts gives a pro-labour bias. Exclusion of London may have been neutral / pro-Tory up to 2004, but London has since shifted bluer, and most polls (certainly on yougov) show Tories peforming better in London than elsewhere (though I know the yougov London region does include suburban fringes). At comparable elections in 2003, Tories won 34.6% to Labour’s 27%, so we could expect to see 40% as a reasonable target.
In Scotland, most polls and comments have focussed on the Scottish Partliament, while the local councils have been a cinderella. But, since the Holyrood election will be fought on ol boundaries which no longer translate to Wstminster, these will not give as good an indicator of a general election result as they may otherwise have done. The local councils conversely may give a better view, in particular since the change in voting system means that people may be more likely to vote with their hearts.
NBenchmarks to watch are that in 2003 respective shares were:
C: 15%, Lab 33%. LD 14.5%, SNP 24%, others 14%.
2005 results (Westminster) were: C: 16%; Lab 39%; LD 23% SNP 18%, others 5%. Although Tories only held one seat at Westminster in 2005, I would be surprised if they did not improve on their 3 FPTP seats at Holyrood in May. Labour took 46/73 FPTP seats in 2003, but could easily lose 10 or more this year.
Wales has no local council elections, but does have the Assembly. This will be interesting since the Assembly and Westminster boundaries are roughly the same and so results can be compared not just with 2003 but also with 2005. That would indicate that Labour is likely to lose several FPTP assembly seats (in all directions)
On the basis of all votes cast in May, if Tories get 40% that would be a very good result due to drag from Scotland and (to a lesser extent) Wales, but if they hit 45% in England, then Cameron is on track to win.
I’m wondering what could have caused this apparent 5% swing from Lib Dems to Conservatives, neither are in government, so it’s not as if they could do anything spectacularly good or bad that really means anything right now. Noise, surely?
Not strictly a polling question, but anyone know how national vote share is extrapolated from council by-election results?
I can work out how it’s done if the ‘base’ elections are the same – i.e. if all the seats the by-elections are caused in were originally vacated in, say, 2004, then you know what the overall national shares were in 2004, how that compares with the 2004 vote share in these seats, and then can compare them with the by-election results and from that work out the projected share.
But if some by-election seats were first elected in 2003, some in 2004 and some in 2006, how can you calculate a base figure?
Keith: “I’m wondering what could have caused this apparent 5% swing from Lib Dems to Conservatives, neither are in government, so it’s not as if they could do anything spectacularly good or bad that really means anything right now. Noise, surely?”
My favourite explanation is that apparent LD -> Tory swing can actually be caused by a combination of Labour -> Tory swing and LD – Labour swing (disaffected voters annoyed with Blair returning to the fold). Though the problem with that is I can’t see Labour at present gaining as many votes as they lose.
Adam: I agree regarding Council election results. The Tories should really win big in May, and it could be an awkward night. Will be hard to top the expectations that are building up. Tory voters are disproportionately more likely to regard voting in Council elections as important, even a civic duty.
I think the notional national share of the vote is done by the results in certain key wards which are pre-selected to be representative of the nation. NOT the whole votes cast.
I think think this latest poll showing the 11% Tory lead is very accurate – and will be proved so at the next local elections – as it was in 2006 . Labour are polling similar figures to their bleak days under Mchael Foot . You will be hard put to find anyone openly admitting to even voting for Labour at the last election ! The interesting thing is – the Tories normally poll slightly more at the elections than they do at the polls prior and the reverse is true of Labour – if this was so – that would make the Tory lead more than 11% in true voting intentions on the day of an election .
the country is in a much bigger mess than the situation michael foot inherited.these poll now seem to becoming consistent.
the lefties are just in denial.
I believe Philip is right – the notional national shares of the vote in local elections are calculated using a basket of wards that supposedly represent the country as a whole. The BBC do one on election night, and then Rallings and Thrasher do a larger one a bit later. Neither of them are the “actual” share of the vote, since that would be heavily skewed towards Labour when the Metropolitian Boroughs are up for election and heavily skewed toward the Conservatives when the shires are up.
Anthony , yes the published notional vote shares are based on a selection of wards but then also adjusted by some unknown or unpublished formula to reduce the % of Others . The actual shares of the votes are available on the Plymouth Uni website ( and I also work them out myself ). I agree you cannot compare this years to last years for the reasons you give but you can compare this year’s results with 2003 for example . and in 3 years out of 4 the Met district results are available and a direct pointer to changes in party support at least in the cities .
Thanks, Anthony – I didn’t suppose you had! In any case, the Greens don’t have the resources to fight anything like all Council seats – they’ve never contested the ward I live in, for example.
Philip , there is no real evidence for any 5 % shift in support from LibDem to Conservative . If you use the change in support figures in the ICM and CR polls in Jan and Feb to justify this claim , then you also have to accept and explain why there was a similar 5 % shift in support from Con to LibDem from December to January in the ICM and CR polls .
I still maintain that all we keep seeing are changes in each month’s figures caused by sampling and M of E variations . What is the true Conservative lead ? certainly less than 11% . What is the actual level of LibDem support probably around 20% . What evidence can I give for that . Look at the council byelections so far this year , the Conservatives have polled very poorly in the Northern ones and only a bit better in the Southern ones . If the Conservatives were 11% ahead you would expect them to easily hold this week’s byelection in Brentwood , instead it is likely to be a LibDem gain from them . Similarly the Conservatives should have easily gained the byelection in Huntingdon last week but it was held by the LibDems on a massive swing to them .
“Extropolating the raw poll data into constitencies again shows an advantage for Cameron.The 3 polls averaged in the last week,using the lists of marginals on this website,would give 348 Conservatives,240 Labour,34 Lib Dems,28 others -an overall majority of 48.”
Hence the problem with taking opinion polls and applying them uniformly to constituencies.I don’t know anyone who thinks the Lib Dems are going to lose 30+ seats and, as has been outlined previously, the Tories are largely flatlined in areas like the North West a,North East and Yorkshire, Thatcher made inroads into these areas, we’ve yet to see Cameron’s Tories doing so.
“the country is in a much bigger mess than the situation michael foot inherited.these poll now seem to becoming consistent.
the lefties are just in denial.”
Err no, we have the bitter experience of ‘winning’ elections based on opinion polls, then losing elections based on actual votes.And, although I’m loathe to clutter Anthony’s excellent site with party politics, I don’t think the most partisan political animal would regard Britain in 2007 as being in a bigger mess than 1981.I seem to recall Michael Foot being way ahead in the polls during the early part of his tenure as Labour leader.
The Scotsman has the latest ICM poll on Scottish Holyrood voting intentions . It was vastly overhyped last night by Iain Dale on his site as being a disaster for Labour and good for SNP but in fact the changes on the previous month are well within normal sampling/MofE variations and not all Labour to SNP anyway . He also falsely said LibDem support was down but that again was wishful thinking and in fact false .
Constituency
Lab 29 -2
SNP 34 +1
LibD 16 +1
Con 16 +3
List
Lab 28 +1
SNP 32 -1
LibD 17 n/c
Con 15 + 1
“the country is in a much bigger mess than the situation michael foot inherited.” – I don’t know a single person, Tory or Labour, who’d claim that today is worse.
Mark: “Philip , there is no real evidence for any 5 % shift in support from LibDem to Conservative . If you use the change in support figures in the ICM and CR polls in Jan and Feb to justify this claim , then you also have to accept and explain why there was a similar 5 % shift in support from Con to LibDem from December to January in the ICM and CR polls .
I still maintain that all we keep seeing are changes in each month’s figures caused by sampling and M of E variations . What is the true Conservative lead ? certainly less than 11% . What is the actual level of LibDem support probably around 20% .”
Look at Anthony’s historic voting data (on the right). This isn’t just the sixth poll in a row to show the Lib Dems below 20, but 16 of the last 19 polls have all showed the Lib Dems below 20. This poll result gives the Lib Dems a pretty average result for them.
If you want to explain a couple of polls as outliers then it’d probably be the CR and ICM ones before last which are the only 2/16 polls since the start of December to give the Lib Dems 20+
Ok Philip , so the polls you don’t like are the outliers . You just do not seem able to grasp the fact that with polls sampling 1,000 people and ignoring don’t knows and won’t votes the sample is much less then you are going to get large fluctuations from month to month . As to the true level of LibDem support is at 17% as in the last 2 polls , 20% as I claim or much higher as all the local byelections would indicate . We will get another clue tomorrow with the byelection in Brentwood a Southern seat which if the Conservatives were 11% ahead they should romp home .
We really need as I have said before polls with larger samples as the Germans do – and weekly . The 2 polls out today for example have samples of 2.500 and 4.500 and the results for all parties vary by no more than 1% . Having said that of course , all the German pollsters were way out at their last GE .
Thanks for the explanation, Anthony, that makes sense
Even in 1979 (Thatcher) and further back in 1970 (Heath) the eve of election opinion polls suggested that the Tories were flatlining in the north and romping away down south but it did not quite work out like that. Why? I seem to recall-but my memory is not what it used to be- it was because the marginals behaved like marginals wherever they were located in the country….except of course in Scotland! I suspect turnout in the marginals will hold the key once again and upset a few poll predictions.
Mark,
in last month’s Scotsman/ICM poll, the Lib Dems registered 17% on both votes; this month’s constituency rating is 16%. So, even if it’s not worth getting excited about either way, Ian Dale was, in fact, correct.
Not your fault though – the Scotsman claimed the Lib Dems were 1% up, but they also halved Labour’s drop on the front page. Maybe no-one at The Scotsman has ten fingers to count on?
While the changes are all within the MofE, the results confirm the picture we’ve seen since last autumn – the SNP between 2 and 6 points ahead of Labour, and the Tories and Lib Dems somewhere in the mid teens. If the results are similar on polling day, it will be a disaster for Labour.
Mark: “Ok Philip , so the polls you don’t like are the outliers . ” – That is not what I said. The only two (out of 16) in the last three months since the start of December are probably the outliers because they are the rare extremes in that period, not because I don’t like them.
In the same time frame there has been the exact same number (2) of polls that show the Lib-Dems on 15 or less. One at 15 and one at 14 – I would just as equally say those two are outliers, again because they are the exceptional extremes and not because I don’t like their results.
Ignoring the two outliers at either extreme, the Lib Dems are remarkably consistent. 12 out of 16 have all been 17,18 or 19. 75% of the polls say the Lib Dems are on 18 plus or minus 1, and this poll fits in with that perfectly.
Reading the Scotsman (under protestas it had a poll in it) it touches on the 6% issue.
In it’s simplest form, the 56 List seats are split among eight regions, seven each. But with FPTP seats each region actually has around 16.
The rule of thumb for threshold is 100/( Seats +1) which means you need 6% in each region to get a list seat. 5% means none, 6% means seven.
In theory 12% should give you 14, and 18% around 21, which is more or less what happens with the Tories because they hardly win any FPTP seats.
In contrast thye LibDems because they do very well in a few rural areas, (Highlands and Islands, and Southern Scotland) perform above what you would expect from a Party getting 16%.
On these polls partly because of geography even though the Tories will get over 90% of the vote, compared to the LibDems, they will only get about 75% of the seats.
Just goes to show that no system is perfect. With an average of 33% of the vote the SNP are predicted to get 44 seats, where as 33% is 43 seats , which at 97% is pretty fair.
Labour on an average of 28.5% should get 38 but are predicted to get 41, which is still 93%, so for the larger Parties it is fairly good, with the distortion having greater effect the smaller the vote.
It’s hardto see how STV could make it better as although it would end the two classes of MSp’s it would mean the same regions or an even worse deal for the smaller parties.
Peter.
Frankly I’m just glad that the undemocratic PR nonsense is in Scotland etc only and I hope it never gets introduced into the UK Parliament.
Thanks for the replies about calculating national share, but sorry, you’ve misunderstood what I’m asking – it’s not how the national share is calculated on council election nights, but how national share is projected from any given week’s council by-election results.
Most weeks, the Press Association reports on the week’s council by-election results (see http://politics.guardian.co.uk/byelections/story/0,,1999462,00.html, for example) – in which they end by stating something like: “Analysis of four comparable results in January suggests a projected 18.5% Tory lead over Labour”
To reach this figure, they must logically have a base figure – they can’t just add up the results in those by-elections because if the by-elections took place in Conservative strongholds, for example, this set of results would be utterly unrepresentative. Taking February’s by-elections, for example, Labour polled the most votes with the Tories in third because most of this month’s by-elections occurred in seats where the Tories were very weak. Unrepresentative wards producing a meaningless overall vote share.
To me, the logical way to work out a projected national share from by-elections would be to add up the results from this week (possibly excluding those seats a major party did not contest), compare them with results when those wards were contested in the main elections – when we know what the national vote shares were, courtesy of Rallings and Thrasher – and then project the national share based on the % change in each party’s standing.
However, that only works if you’ve got the same “base” year to project from – whereas in most weeks, by-elections will take place in seats that were originally contested in four different years, from 2003 to 2006. In 2006, the Tories polled 39% nationally, in 2003 they polled 31%.
Hence, the base shares of the vote are different and incomparable. So, if anyone can explain how this isn’t voodoo psephology I’d be much obliged.
Oh, and just one other thing: “the situation is much worse than that Michael Foot inherited”? Michael Foot didn’t inherit anything: he lost.
Just to add re. the national projection on local election night – yes, all of you who’ve commented on it are right about how it’s done – using “representative key wards”, but subsequently the entire national vote (and that of Mets, districts, unitaries, London boroughs etc) is calculated by Rallings & Thrasher and published in their results compendiums.
Philip,
Whats undemocratic about PR, the make up of the national parliament broadly represents the voting intentions of the nation. seems democratic to me.
In 2005 Labour got 35.2% of the vote, but the Tories got slightly more, while Labour got 356 seats to the Tories 198.
So on virtually the same number of votes Labour got 55% of the seats and the Tories only 30%, and you think that is democratic.
Peter.
Philip – OK Look at the last 16 polls eliminate the ouliers and you get the Conservatives at 37% with a 5% lead nowhere near 40% and 11% lead and the supposed 5% shift between LibDem and Con is a myth . As I repeat again , there is no evidence that public opinion has changed to any significant extent at all for 8-9 months now . You can debate whether the average figures are correct or not but other evidence from council byelections is pretty consistent that LibDems are performing much better than the poll figures would suggest .
Peter, this really isn’t the place for a debate about PR, but the overwhelming reasons why PR is undemocratic are:
1) that it removes power from voters and gives it to politicians to make back-room, behind the scenes, grubby and cynical deals trading policies they have no mandate to discard or accept.
2) It makes representatives more distant from their electors by imposing systems that either require vast, unmanageable electoral areas and/or (even worse) creates lists that give representatives seats for life with no possibility of defeat (and before you come back with “but under STV lists are open” that only works if parties field full slates of candidates, which we know from experience in Northern Ireland they do not).
Hence, PR isn’t just undemocratic – it’s corrupt – and only supported by pathetic nerds with nothing better to do than invent byzantine systems that obscure the will of the people, or those who (surprise, surprise) will gain more representation from it. Hardly a strong reason to endorse it in either case.
“In 2005 Labour got 35.2% of the vote, but the Tories got slightly more, while Labour got 356 seats to the Tories 198.
So on virtually the same number of votes Labour got 55% of the seats and the Tories only 30%, and you think that is democratic.”
As a Tory, I say: Yes, Absolutely! What can be more democratic than for our MPs to have the knowledge, the fear, the terror, that they’ll lose their own personal seat. What can be more democratic than in 1997 seeing the Secretary of State for Defence, Michael Portillo, lose his seat? In First Past the Post, if a party polls badly they can find half their MPs are now unemployed. Cabinet secretaries can go unemployed.
That just won’t happen in PR. The likes of Portillo will never lose a seat in PR, they’re too near the top of their party’s list.
You’re talking about Labour having a disastrous election, are we going to see half their MPs go? No, you’re talking about them going from 50 down to 41. They keep 82% of their MPs and that’s bad? That’s not keeping them to account. That’s barely more significant than natural churn.
—————
Mark: “Philip – OK Look at the last 16 polls eliminate the ouliers and you get the Conservatives at 37% with a 5% lead nowhere near 40% and 11% lead and the supposed 5% shift between LibDem and Con is a myth . As I repeat again , there is no evidence that public opinion has changed to any significant extent at all for 8-9 months now . You can debate whether the average figures are correct or not but other evidence from council byelections is pretty consistent that LibDems are performing much better than the poll figures would suggest .”
I don’t think council byelections are relevant whatsoever, so I’m not disputing them. I have honestly never even looked into their results, so again that’s not dismissing their results based on a dislike of their outcomes, I’ve never considered them relevant. Council elections have barely half the turnout of a general election, what’s the turnout to a council byelection?
You’re right, over the last quarter of a year there is no reason to suggest that there has been just recently a 5% swing from Lib Dem to Tory, because there’s no reason to believe that last month the Lib Dems were on 22%. That was an outlier (just as much as the 14% one was) and its reverted back to being 18% +- 1% now
Looking long-term over 8-9 months and you’re right, there has been little change. Since the Lib-Dems have never been averaging 20 or above. They’ve gone from an average of 19 down to an average of 18.
Philip , Of course council elections and byelections have SOME relevance to how the parties are standing nationally . Turnout can be variable but in some cases as in the 2 Burnley byelections last month quite high over 50% in 1 of the 2 wards fought despite pretty horrendous weather on the day almost on a par with GE turnout . Analysing the results can give clues as to the health of the parties standing in a particular area . For example Calder Valley and Halifax are 2 marginal seats the Conservatives would need for an overall majority . The Conservative results in last year’s May elections were poor in Calder Valley and abysmal in Halifax . Clearly there is something wrong there with the local Conservative party . Similarly LibDem performance in Bolton eas particularly poor . All these factors mayl impact on GE performance in due course .
Adam,
All you say about PR actually happens now, only it happens within parties.
Take a look at a One Party run council compared to a No Overall Majority. In the NOM council there is real debate abut the main issues, in a one party one the main council meetting can be over in an hour, but the internal OP meeting the day before can last all day as the deals are carved out in private.
It’s all very well for supporters of FPTP to talk about the like with their MP, but more than 90% of the electorate never contact their MP, and in some cases more than 50% don’t even know who it is. Under FPTP the real power lies with the party whips and machines and the individuals are just lobby fodder.
The average MP voteds with the party line 90% of the time, hell look at Iraq, do you honestly believe that all those Labour MP’s who supported it, would have done so if the Tories had been in power.
Close to their constituents and representing them…. Dream on.
The parties are stilll really powerful under PR but at least there are a range of them. Sure they make deals like Labourand the LibDems, but people can see what policies have been dumped or ammended, and judge them on that.
In Scotland the LibDems campaigned on No Tuition Fees and settled for No (Up front) Tuition Fees, a bit like animal farm , when no one was looking they climbed up the ladder and paintedin some new words. But people knew what they had did, because their original policy was on record.
It’s the same with Free Personal Care, the executive opposed it tooth and nail until 35 minutes before the final vote ( a back bench rebellian by a few LibDems and Labour MSP’s ment they were going to lose and it was never an executive bill anyway), and then tried to take credit for it because they had signed it in to law.
It’s a bit like Germany and Japan taking the credit for ending WW2 because, they signed the surrender.
My point is under FPTP a deal like that would have been done behind closed doors and we wouldn’t have know because all the people involved would be from the same party. PR isn’t perfect but it works better in practice that FPTP.
For all the executive has had lack lustre policies and peformance, and the parliament building and expenses have been a fiasc, in terms of open debate and democracy, holyrood leaves westminster for dead.
Oh and of coursewith a PR parliament you don’t need a second chamber to keep it in check.
The only people I have ever heard defending or promoting the “link with constituents”, MP’s, Councillors and Party activists, because their jobs depend on it.
I was a Councillor for four years and the number of my constituents I actually met and helped was a tiny fraction of those i represented,far more of them gottheir help or advices from the CAB.
In contrast, in a one hour meeting with officals i could make budget decisions that could have a real impact on hundreds of lives.
My hope is that with PR for local government in Scotland come May, we will have a system where councillors shift to spending far more time making councils deliver services better for everyone, that running around trying to help individuals when they don’t.
Peter.
Peter, your argument is that when FPTP produces no overall control then the consequences are similar to under PR – absolutely, they are – but FPTP greatly reduces the risk of no overall control which is its overwhelming strength. All you do by making that case so eloquently is to buttress my assertion that single party government is far more democratic than coalition government.
Your dismissal of the “representative” element of an electoral system rebounds upon you, because your very first claim when you opened this discussion was that PR is more representative of the people’s will – but given you appear to regard the representational aspect of our democracy as largely symbolic don’t then bemoan the unrepresentative nature of FPTP.
Elections aren’t about ensuring proportionality – they’re about electing governments and representatives. But if you disagree, presumably you advocate one electoral constituency covering the entire UK, electing 650 MPs – the ONLY way to guarantee exact proportionality?
The problem you have, of course, if you don’t believe in one constituency with the national vote allocated strictly proportionally is that any alternative is less proportional, and therefore you cannot claim proportionality to be a defining principle – you either believe proportionality is paramount or you don’t. FPTP isn’t entirely disproportionate: it’s just one of many less proportional systems that places other priorities ahead of that principle.
Your comments about the amount of constituents you came across during your time as a councillor are not untypical but I have to say criticise yourself for that – not your constituents or the system – I was a councillor for eight years and went out of my way to make sure my constituents engaged with me – it was the only possible way I was going to hold onto my seat in an area that was far from natural territory for my then party; but in any event, the issue is surely that constituents have the opportunity for straightforward access to their representative, rather than that they use it?
If that is the issue then it is incumbent upon the system to ensure that the link is clear and that their representative is accessible and accountable, not a pick-and-mix jumble who aren’t accountable because the system means that in most instances at least one from each party gets returned regardless.
Finally, sorry to point this out, but yet again we have a supporter of a party that would benefit from PR arguing for PR. Forgive me for being cynical, but even those in Labour who argue for PR only do so because they misguidedly believe it would forge a left-of-centre majority against the Tories in perpetuity (curiously, they then never have an explanation as to how, if that’s true, right-wing parties are in power under PR systems…).
Sorry, but the next time I come across a PR supporter who supports it despite its detrimental impact on their political cause will be the first.
Adam – I do not have a political cause, I am not a hard-and-fast supporter of any political party. I have voted for three different parties during my 12 year voting career. However, I am very much in favour of PR, as I simply cannot regard the idea of one party having TOTAL, unshackled control of government when it has received support from barely a third of those who have voted. I would ask you to address two points:
1) It is entirely possible that Labour could get fewer votes than the Tories at the next election and still be returned to power with a small overall majority. i.e. more crosses against Conservatives candidates nationwide than against Labour ones, but Labour in absolute power. Can you explain how this is democratic?
2) If PR is such a crazy idea, why does every other country in Europe have some form of it in place? Why is it perfectly good for them, but out of the question here?
1: Even as a Conservative voter I can still recognise that as democratic. A large part of the “Labour bias” is simply the fact that safe Labour constituencies have a far lower turnout than equivalent safe Tory constituencies. Constituencies rightly are done per capita and not per voter, so Labour’s support is under-represented in the final vote due to their turnout but this is naturally corrected through our constituency system.
This provides a healthy correction (even if it penalises my party) that PR lacks.
2: Why compare against Europe? That is meaningless, why not just as equally compare against the rest of the Anglophone world: American, Canada and Australia etc – none of those three use PR. One could just as easily ask: Why is using constituency based Parliaments good enough for Canada and Australia but not good enough for us?
As for comparisons with Europe and how PR works in practice – I would rather our stable democracy, along with the likes of Australia and Canada than the unstable travesties generated through PR in nations like Italy and Israel.
Incidentally, Australia does use a form of PR for their Senate – though their two chambers have similar powers as ours do. The lower house is elected through constituencies, but on a form of transferable vote and not first past the post. That model works very well for them.
Toby, as you don’t appear to have a political cause, my challenge doesn’t apply to you. However:
What you seem not to appreciate is that one party governing alone is more democratic – under any system – than two separate parties governing together UNLESS there has been a clear statement of intent that the two will govern together and a joint manifesto explaining what this coalition’s policies will be prior to the election.
Of course parties won’t do that – and you’ve seen why, in part, with the furore over the botched spin over Menzies Campbell’s supposed conditions of supporting a minority Labour Government. But without such a pre-election public commitment, there is no democratic mandate whatsoever for any form of coalition.
It is simplistic and disingenuous to say “ooh, Labour’s got 35% and the Lib Dems have got 17% so 52% have voted for a Labour/Lib Dem coalition”. In fact 35% voted Labour because they wanted a Labour majority, and 17% voted Lib Dem because they wanted a Lib Dem majority.
Personally speaking, under no circumstances would I want the Lib Dems to hold any power anywhere at any time – and in fact if I believed there was even a chance that Labour would sell out to the Lib Dems in the event of them losing their majority, I would vote Conservative to help ensure that doesn’t happen.
The reason why FPTP is democratic is because this isn’t a national election: it is 659 elections happening simultaneously, the majority of which result in the representative chosen winning over half the vote (and the vast majority of others winning more than 40%). That, under your apparent definition of “democratic” must be acceptable.
Now, if you want a national election that’s fine and if your principal criterion is proportionality the only consistent position for a PR advocate is a single constituency. Is that your position? Again, if it isn’t, let’s not hear so much about how a government with less than 51% support is undemocratic.
In respect of your second point I could note that both Italy and Israel have been rowing back from their terrible PR systems, or that France’s two-round system isn’t proportional, but as Philip has explained, compare like with like, not Britain with Europe.
Toby, instead of just complaining about the system you have – that’s easy to do; tell us what system you want and explain why it’s better than all the others. I always enjoy this part – because the moment a PR supporter comes out with their preferred system, all the other PR supporters who favour other systems turn on them and a fabulous fist-fight ensures – it’s one of the best spectator sports around!