Pedro Magalhães

Margens de Erro

Pitagórica, 8-12 Oct., N=503, Tel.

Posted October 22nd, 2012 at 11:00 am4 Comments

PS: 34,1%
PSD: 29,5%
CDU: 9,1%
CDS-PP: 8,3%
BE: 7,2%

Here.  Other findings: massive government disapproval and preference over cutting expenditure over increasing taxes, although this last question appears to have been formulated in a way - "Should the government cut expenditures instead of increasing taxes?" - that is probably not the best. Also interesting is where people thought where cuts should be made: 61% public companies, 38% local government, 0,7% education, 0,3% health. Besides the question of what the response categories actually were, one wonders how many respondents understood the real implications of their answers...

by Pedro Magalhães

Polls in Portugal, update

Posted October 19th, 2012 at 2:58 pm4 Comments

The LOWESS line uses a 25% bandwidth. The rest is self-explanatory, I trust.

by Pedro Magalhães

Rescaldo sondagens Açores

Posted October 15th, 2012 at 12:41 pm4 Comments

Pouca coisa para trabalharmos. O site da ERC regista uma sondagem de intenção de voto, realizada pela Eurosondagem, com trabalho de campo terminado a 2 de Outubro. Entre parêntesis, diferença em relação aos que vieram a ser os resultados finais:

PS: 41,9% (-7,1)
PSD: 35% (+2)
CDS-PP: 9.7% (+4)
CDU: 3,7% (+1,8)
BE: 3,3%(+1)

O erro absoluto médio é 3,2, bastante acima do que sucede normalmente em sondagens nacionais. A sondagem à boca das urnas da Católica teve resultados, como teria de ser devido à sua natureza, bastante melhores:

PS: 50% (+1)
PSD: 31% (-2)
CDS-PP: 7.5% (+1.8)
CDU: 3% (+1.1)
BE: 3%(+0.7)

Desvio absoluto médio de 1.3, igual ao que se passou em 2008. E uma curiosidade: esta é a segunda eleição seguida, depois das legislativas, em que o CDS-PP é sobrestimado pelas sondagens. Algo se deverá estar a passar, ou no eleitorado do CDS ou nas sondagens.

by Pedro Magalhães

Eurosondagem, 4-9 Oct., N=1021, Tel.

Posted October 13th, 2012 at 7:58 pm4 Comments

PS: 34.8% (+1.1)
PSD: 30% (-3)
CDS-PP: 10% (-0.3)
CDU: 9.5% (+0.2)
BE: 7.7% (+0.7)


by Pedro Magalhães

Marktest, 17-20 Sept, N=805, Tel.

Posted October 2nd, 2012 at 6:16 pm4 Comments

PS: 29.6% (+3.8)
PSD: 20.2% (-6.5)
CDU: 10.1% (+0.2)
BE: 8.6% (+1.7)
CDS-PP: 5.2% (+0.8)

Values in parentheses from previous July 2012 poll. Here.

by Pedro Magalhães

42 anos de vida sem saber que isto existia. Vou ali suicidar-me e já venho.

Posted October 2nd, 2012 at 1:42 pm4 Comments

by Pedro Magalhães

Before the debates: polls and forecasts

Posted October 1st, 2012 at 10:14 am4 Comments

Two days before the first debate, where do we stand in terms of election forecasts for the US presidential election?

1. Polls: polls are not forecasts, but they are of course informative. One obvious argument is that, the closest we get to the election, the less likely it is that any factor will emerge that may change the balance of forces. But the other - actually more plausible - argument is that, the more time elapses and the campaign progresses, the more likely it is that voters' intentions reflect the "fundamentals" of the election and the latent preferences of voters. At the moment, all poll aggregators place Obama ahead by about 4 points.

2. Models: econometric models use data from past elections, at the national or state level, and a number of contextual variables (mostly from the economy - such as change in national or per capita income, unemployment, and others - and sometimes from approval polls) to predict outcomes. The latest issue of PS collects almost all of them. The average is 50.4 for Obama, an extremely narrow advantage.  An article using ensemble Bayesian model averaging of those models estimates that "between 46.4% and 52.5% of the US voters will support the incumbent in 2012, and there is a 0.60 probability that the vote for Obama will be greater than 50%." Very close indeed.

3. Prediction markets: these are markets where investors buy and sell shares of political outcomes. Their performance has been extremely good, although there is a debate on whether they are "superior" to polls when the latter are properly treated. In any case, this is what the markets have been telling us until now:

Iowa Vote Share:

Iowa probability of winning:

Intrade Obama's probability of winning:


by Pedro Magalhães

Muito bom

Posted September 27th, 2012 at 5:25 am4 Comments

by Pedro Magalhães

O que tenho lido sobre as eleições americanas: uma selecção.

Posted September 26th, 2012 at 9:44 am4 Comments

1. O, para os últimos números das sondagens nacionais e estaduais e o que implicam para o colégio eleitoral, e os comentários de Mark Blumenthal (aqui sobre o "post-convention" bounce de Obama).

2. Também sobre as sondagens (e sobre previsões que se podem fazer na sua base), Nate Silver no FiveThirtyEight.

3. Votamatic, idem.

4. John Sides, em dois blogues: YouGov e Monkey Cage. Aqui sobre como a discussão sobre os rendimentos de Romney afectaram a sua campanha, e aqui sobre as "gaffes" e o seu (duvidoso) efeito nas sondagens.

5. Brendan Nyhan, aqui sobre a estranha persistência do mito de que Obama seria muçulmano.

6. Andrew Gelman (também no Monkey Cage), sobre quase tudo o que me interessa, incluindo e para além de eleições. Aqui resumindo por que razão Romney é um mau candidato (e dificilmente outro seria bom tendo em conta aquilo em que se tornou o Partido Republicano).

7. Pollyvote, tudo sobre diferentes formas de prever resultados eleitorais.

8. Jay Cost, ajudando a que os Republicanos não percam completamente a esperança.

9. Lynn Vavreck, que contribui para vários dos blogues acima, aqui sobre os eleitores indecisos.

Tudo disto é um bocado técnico, pouco especulativo e muito colado aos dados empíricos. Mas sucede que isso é precisamente o tipo de coisa que me interessa. Para "grandes reflexões", nem sempre impecáveis do ponto de vista empírico mas sempre impecavelmente bem escritas e inteligentes, leio o David Brooks, aqui sobre as mutações do conservadorismo americano.

by Pedro Magalhães

More poll results and caveats

Posted September 20th, 2012 at 5:59 pm4 Comments

Starting with the caveats, note that in a "poll-poor" environment such as Portugal, it's understandable that a single poll gets a lot of attention. But let's not forget that a poll is just a poll. And although this last one somehow "feels right" (unlike this one), that does not make it "right" or "wrong," whatever that means when we talk about measuring public opinion. We need more data.

Furthermore, recall that the last Catholic University poll was conducted in late May/early June. This means that the massive PSD drop cannot be assigned exclusively to the measures related to social security contributions: the "Relvas affair(s)" and the privatization of public television were two very contentious issues that preceded the last austerity measures. Finally, note that the poll was conducted not only at the height (so far) of mobilization against the government but also following a week during which PSD figures, employers, unions, and even the minor coalition partners voiced their disagreement with the government's measures. So, although it is always possible that things can get worse for the PSD, it's probably safe to say that they had never as bad as in that period when the fieldwork took place.

Having said that, a few more interesting results:

* In the previous poll by the same institute, the PM was positively rated (getting 10 or more in a 0 to 20 scale) by 48% of respondents. This figure now dropped to 34%, making him the party leader with the worst approval rating.

* 78% of respondents disagree with the notion that the reduction of companies' social security contributions will help fighting unemployment, while 81% disagree with the notion this will have an impact on consumer prices. This is, in a way, a measure of how unsuccessful the communication of this measure was, as well as of the large consensus against it on the part of the kind of sources (media, expert, partisan) people likely used to form any sort of opinion about it.

* In contrast, there is an absolutely massive support for measures that were barely discussed, such as increasing taxes over luxury items or over high value real estate: support is always above 80%.

* 48% of respondents think now likely that Portugal will face a situation similar to Greece in the short-run, against 34% just three months ago.

*60% of respondents believe that no opposition party would do better than the current government if it was in power. Considering that 77% make a negative evaluation of the work of the current government, this helps explaining why, while the PSD drops in this poll, the PS fails to capitalize.

by Pedro Magalhães