Wednesday, 7 November 2012

US Elections 2012: Some initial thoughts on the results and polls

The big political story is no change.  Obama is still the president, there looks like there will be very little change in the party composition of the Republican held House and Democrat led Senate.  Given the polarisation of US parties at the moment, this seems to be a recipe for continued legislative gridlock.  The last Congress was the least productive in over 120 years.

The House results, though still not all in, look like the most striking. In 2010 the Republicans won 56% of the seats on 52% of the votes and a 7 point lead over the Democrats.  In 2012 they look like they will get still get 54% of the seats on less than a percentage point lead, if any at all. So it looks like the electoral system is seriously unresponsive and biased to the Republicans. This is most likely because of a mixture of incumbency advantage and most strikingly disproportionate Republican power in the Congressional redistricting process

The results of the presidential election were as expected given the polls.  I'm glad to see that predictions from polls did well on average, but there were clearly big systematic differences between polling companies (e.g. Gallup had Romney ahead for most of October and at the end while, Reuters/Ipsos polls tended to be more favourable to Obama). While the predictions run by Simon Jackman, Nate Silver and others are impressive and did well, I would still have been more cautious about the probability of an Obama win given the risks of systematic bias in the average polls (industry bias). While US polls do have a good record, with increasing problems with survey research (declining response rates, mobile-only rates increasing, vagaries of turnout etc) it is getting increasingly precarious trying to estimate election shares of the vote from polls. Attempts to predict the outcome of the 2010 British general election suffered from an average error in the Lib Dem vote intention polls of around 4 points.  That kind of difference in the US would have completely changed the story (to either a comfortable Romney victory or Obama improving on his 2008 result). So in truth I was still thinking it was too close to call, for similar reasons given by Andrew Gelman.  

Peter Kellner has an interesting discussion of the socio-demographic breakdown of the vote in YouGov polls here. They differ substantially from those in the exit poll here which show much smaller gaps than YouGov do. Regardless of why, Kellner's main point that, from a British perspective, the gaps between men and women, old and young, between ethnic groups and some other bases are startling. The exit poll figures suggest the gaps widened slightly between 2008 and 2012 except for the age gap where the Obama vote among 18-29 year olds looks to have dropped by 5 points, but it is still at 60%. This still looks like a nation divided and polarized. 

I'm looking forward to the analysis of what happened (both with the votes and the polls) to come.

Sunday, 6 May 2012

English Local Elections 2012: Who won votes from whom?

This is quite hard to tell the pattern of vote switching from one party to another from the aggregate ward-level data alone because of the well known ecological inference problem, but some insights can be gleaned from the pattern of correlations in the changes in party shares.

If one party is gaining at the expense of another we would expect to see a strong negative correlation between the in the changes in the shares of the vote, as one tends to do relatively well where the other is doing relatively badly.  Below are the correlation matrices for the ward level changes in the shares of the vote and changes in turnout from the BBC Key Wards data for both change since 2008 (when the incumbents were last elected) and change since last year.  

Labour's rise seems to have depended on squeezing minor parties and independents at least as much as the Liberal Democrats. Stronger UKIP performances clearly did damage the Conservatives, both relative to 2008 and last year.  The strong negative correlation between the BNP and UKIP changes (-0.96 for change since 2008) suggest that UKIP was picking up most of the former BNP voters.  

The final row in each matrix shows the correlations between party performances and the change in turnout.  It has been suggested that part of the explanation for the poor Conservative and Liberal Democrat performances is that the supporters of these parties disproportionately stayed at home.  If there was a strong tendency of this kind then we should expect to see the coalition parties performing worse where turnout fell most heavily, i.e. positive correlations between the share of the vote changes for these parties and turnout change. Instead the correlations are negative or not statistically significant.  So these results suggest that Conservative and Liberal Democrat voters were not disproportionately staying at home.

            |  con0812  lab0812   ld0812 aoth0812 ukip0812  grn0812  bnp0812
     con0812 |   1.0000 
     lab0812 |  -0.0631   1.0000 
      ld0812 |  -0.3094* -0.3865*  1.0000 
    aoth0812 |  -0.3370* -0.4412* -0.3582*  1.0000 
    ukip0812 |  -0.2946* -0.2975*  0.1721   0.3583*  1.0000 
     grn0812 |   0.0373  -0.1896* -0.1047   0.2482*  0.0468   1.0000 
     bnp0812 |  -0.1222  -0.1171  -0.2698   0.5735* -0.9616* -0.1965   1.0000 
    turn0812 |  -0.1626*  0.0537  -0.1114*  0.0546  -0.3388* -0.1778*  0.2861 

             |  con1112  lab1112   ld1112 aoth1112 ukip1112  grn1112  bnp1112
     con1112 |   1.0000 
     lab1112 |  -0.0787*  1.0000 
      ld1112 |  -0.1639* -0.2890*  1.0000 
    aoth1112 |  -0.4163* -0.4644* -0.3263*  1.0000 
    ukip1112 |  -0.3205*  0.0546  -0.1212   0.4517*  1.0000 
     grn1112 |   0.1147*  0.0554  -0.0820  -0.0386  -0.0165   1.0000 
     bnp1112 |  -0.2853  -0.0569   0.1040   0.3543* -0.7879   0.1525   1.0000 
    turn1112 |  -0.1173* -0.2631*  0.0289   0.0769  -0.1449* -0.2106* -0.0071 

Note that these are pairwise correlations so the N's will vary, and for the ones between minor parties they can be very small. aoth refers to all others collectively. The changes are only calculated for when a party stood both times, so these matrices say nothing about the effects of candidate entry or withdrawl. Correlations that are statistically significantly different from zero at the 5% level are starred.


Thanks to the BBC, John Curtice, Rob Ford and Jon Mellon. This who provided data and analysis.  The views and errors are all mine.

Friday, 4 May 2012

Scottish Local Elections 2012

by Stephen Fisher, 4th May 2012

In yesterday's Scottish local elections, both Labour and the SNP were up and the Tories and Lib Dems were down relative to when they were last run in 2007. 

The BBC collected the results of the first preference votes for the 93 wards in 6 councils. (Local elections in Scotland use STV.) The shares of the vote were as follows (with 'aoth' referring to all but the four main parties aggregated, and turnout in the final column).

    council_name   con12sh   lab12sh   ld12sh   snp12sh   grn12sh   aoth12sh   turn12  
        Aberdeen       9.7      29.7     15.1      31.3       2.5       14.2     33.0  
          Dundee      11.3      30.1      9.0      43.4       0.6        6.3     35.9  
       Edinburgh      21.2      28.1      9.3      26.9      11.4       14.5     41.9  
            Fife       7.8      38.5     13.1      31.1       1.0        9.5     37.2  
         Glasgow       5.9      46.9      2.9      32.6       5.5       11.6     31.3  
    Renfrewshire       9.1      47.6      4.4      35.3       0.0        3.7     41.6  

The corresponding changes in the share of the first preference shares of the vote were as follows.

    council_name   con0712   lab0712   ld0712   snp0712   grn0712   aoth0712   turn0712  
        Aberdeen      -4.4       5.1    -11.8       1.8       0.9        9.2      -17.2  
          Dundee      -1.3       0.8     -2.3       3.4      -0.0       -0.6      -14.0  
       Edinburgh      -0.9       5.2    -12.7       6.5       3.2        1.8      -15.7  
            Fife      -2.8       9.8     -9.3       3.3      -0.4       -1.0      -12.9  
         Glasgow      -1.6       3.7     -5.2       8.1      -0.9       -5.0      -12.1  
    Renfrewshire      -2.4      10.8     -5.4      -0.1      -0.2       -2.8      -14.0  

The average changes since 2007 across all the 93 wards were Con -2.2, Lab +6.3, LD - 8.2, SNP +4.3. 

Turnout was down heavily mainly because the 2007 elections were fought on the same day as the Scottish Parliament elections in 2007.  The confusion and the number of spoilt ballots that arrangement caused led to the local elections being delayed by a year this time round.

Given that the SNP replaced Labour as the largest party in Scotland in 2007 it is perhaps unsurprising that Labour's share of the vote has gone up more than that of the SNP's.  But while the SNP improved on their 2007 result, it has come nowhere near being able to match their performance in last year's Scottish Parliament elections. Labour however have recovered a bit on their poor performance then.  Comparing the first-past-the-post share of the vote from 2011 with the share of the first preference votes this year in the BBC Key Wards the overall changes were Con -0.7, Lab +1.6, LD -1.0, and SNP -12.2.

Overall the SNP won 424 seats and Labour 394. But it is not yet clear which party will be ahead in terms of the share of the first preference votes. 


Thanks to the BBC, John Curtice, Rob Ford and Jon Mellon. This who provided help with data and analysis.  The views and errors are all mine.

Wednesday, 2 May 2012

About this blog

This blog will be about elections and other things, such as public opinion and social attitudes.  I hope to post a few comments and analyses on election results, polls and aspects of research I'm engaged in.