Sunday, December 21, 2008

Chávez poll

Hinterlaces shows 61% of Venezuelans opposed to the constitutional reform allowing indefinite re-election, 31% in favor, and 8% undecided. The percentage in favor is very close to a poll Boz linked to a week ago.

There is, however, still a long time to go--about two months. Chávez has proposed that the vote coincide with the 20th anniversary of the Caracazo (February 27), which was the beginning of the end of the puntofijista era.

15 comments:

boz 1:41 PM  

During today's Alo Presidente, he proposed 15 February, which is apparently the anniversary of a major Bolivar speech.

Anonymous,  8:52 PM  

there is no such a speech. He is just rushing it because he knows his changes are reduced by the oil barrel price

Justin Delacour 11:27 PM  

Hinterlaces shows 61% of Venezuelans opposed to the constitutional reform allowing indefinite re-election, 31% in favor, and 8% undecided.

I have a very difficult time understanding how the PSUV could gather 4.5 million signatures favoring a new constitutional referendum if Chavez were losing by the margin that Hinterlaces suggests.

boz 9:15 AM  

He is just rushing it because he knows his changes are reduced by the oil barrel price

I doubt two weeks (time, not this blog) will make much of a difference in terms of the economy.

If there is a political motivation to the date, it's more likely that Chavez believes that holding the referendum more quickly will give the opposition less time to organize.

Greg Weeks 9:22 AM  

I do think, however, that he is doing it quickly (i.e. early 2009) rather than later for economic reasons.

Justin Delacour 12:42 PM  

But if Greg thinks that Chavez could still win, one wonders why he gives credence to an extremely dubious poll with a 61-31 spread in favor of the opposition. Given (1) the PSUV's recent 11-point victory in the nationwide tally of gubernatorial elections and (2) the fact that Chavez himself has long been more popular than his party, the notion that Chavez could be trailing by 30 points simply defies logic.

Greg Weeks 1:49 PM  

If I see other polls, I will cite them--I haven't seen anything else. I think the key question is whether Chávez's personal popularity translates into support for indefinite election. There is not an automatic correlation.

boz 4:47 PM  

It's worth noting that this poll doesn't attempt to predict the election results. It's discussing all voters rather than likely voters. There are reasons to believe Chavez supporters will turnout in greater numbers.

Justin Delacour 11:54 PM  

It's discussing all voters rather than likely voters.

Perhaps if Boz didn't place such blind faith in Latin American polling, he would stop to consider the possibility that some pollsters' survey samples are not truly representative of the populations they purport to reflect. Any cursory glance at the history of Venezuelan polling would demonstrate that many pollsters are way off in their projections. Most such pollsters are aligned to the opposition, but there are also some government-aligned pollsters that have been off in their projections.

Before one places blind faith in Hinterlaces, it would be wise to look into the firm's history.

boz 10:54 AM  

I was simply making a point about "all voter" vs. "likely voter" models. It's a significant point for those who understand polling.

Hinterlaces was about four points off in the 2007 constitutional referendum. That's fairly accurate and reflects an improving model over time.

Justin Delacour 1:06 AM  

I was simply making a point about "all voter" vs. "likely voter" models. It's a significant point for those who understand polling.

Oh, I understand polling quite well, Boz.

I suspect there are some things about polling that you choose not to understand, though.

Simple question for you, Boz. Why do you think it is that Venezuelan pollsters routinely present two sets of polling data --the so-called "all voter" and "likely voter" data-- in a way that is uncharacteristic of how polling data is usually presented in other countries?

boz 7:56 AM  

Nothing uncharacteristic about it at all. Most polling firms across the world use various models. If you look at the last election cycle in the US, for example, at least five firms were publishing both "all voter" and "likely voter" models on a daily basis.

Justin Delacour 3:09 PM  

Nothing uncharacteristic about it at all.

No, you're way off on that, Boz, and I think you know it. Some U.S. pollsters presented expanded voter models for this election because there was some uncertainty about what the level of turnout would be, but these models are quite distinct from what Venezuelan pollsters do.

The bottom line is that Venezuelan pollsters present the so-called "all voter" and "likely voter" data because that's what their funders want them to do. The partisan funders of Venezuelan polls believe in the bandwagon effect (whereby a favorable poll can help build momentum for their side), so they have their pollsters produce two different sets of polls so they can parade about with whichever one of the two sets is more favorable to their side.

It is a completely unethical set of practices (for which Boz is an obvious apologist). The role of any ethical pollster is to make projections, not to operate as a partisan actor that issues skewed data for the purpose of building political momentum for one side or another.

boz 5:52 PM  

Nearly every polling firm worldwide does an all voter and likely voter model when they are polling near an election. It's standard practice. Likely voter models get more accurate as the election gets closer. As this election gets closer, I expect Hinterlaces and others to release likely voter results as they did prior to the 2007 referendum.

Justin Delacour 7:03 PM  

Nearly every polling firm worldwide does an all voter and likely voter model when they are polling near an election.

First off, Boz, the "all voter" label doesn't even make sense. "All voter" models? How could they all be voters if significant portions of them don't vote?

Regardless of what kind of numbers-crunching a pollster may do internally, the question is about what kinds of polls are typically understood to be worthy of being presented as legitimate indicators of the voting public's intentions. As will soon become clear, there is a lot of poll-cooking in the period leading up to a vote in Venezuela because those who commission the polls want polls that they think will build political momentum for their own side. It will soon become clear that the poll showing a 61-31 spread in favor of the opposition was cooked for the purpose of building opposition momentum.

  © Blogger templates The Professional Template by Ourblogtemplates.com 2008

Back to TOP