|

Venezuela’s National Assembly Elections 2010: A Poll Analysis

An analysis of poll results regarding the National Assembly elections, coming from companies with varying degrees of bias one way or another, with the aim of trying to get a sense of what the outcome will be from the Sunday elections.

What follows is a brief attempt to analyze the poll numbers regarding the National Assembly elections, coming from companies with varying degrees of bias one way or another. I’ve included much of the cited data below (and prior election figures can all be found on Wikipedia with links to their sources), including poll numbers and the result from the last vote in Venezuela, in 2009, along with the polls related to the National Assembly elections next Sunday, 9-26-10, so readers can also draw their own conclusions.

In hopes of trying to get a sense of what real public opinion might be with regards to next Sunday’s elections, I wanted to look at the accuracy of the polls for the 2009 vote (and in some cases a few other electoral processes), especially the predicted margin of victory from the final poll of each company before the vote on 2-15-09 (and sometimes the trends in their polls, if there were multiple). The final percentages of votes cast in the election were 54.86% in favor to 45.13% opposed, a margin of 9.73% (with 99.75% of ballots counted).

One very interesting and important polling company to look at is IVAD, the Venezuelan Institute for Data Analysis. Even U.S. corporate media have spoken of their credibility. Their last two polls of the 2009 vote, released in the 2 weeks prior to the vote, showed leads of 5.5% and then 8.0% for the “yes” vote. These numbers indeed turned out to be pretty accurate, but slightly underestimated the numbers for the pro-Chavez side.

We can also go back to the regional elections of 2008, which saw approx. 5.5 million total votes for pro-Chavez governors nationwide, and approx. 4 million votes for opposition candidates, coming out to about 58% to 41%. In poll results published a month before the vote, IVAD put public opinion at 43.0% for the PSUV candidates and 30.4% for the opposition, less than a 13% margin. Again, we see that this is somewhat less than the 17% differential in votes, but is still fairly accurate. If we go back to the defeated constitutional referendum of 2007, we can see that IVAD’s final poll put the level of support for the “yes” (pro-Chavez) vote at 38.8%, while the support for the “no” was at 42,7%, a difference of 3.9%. With the final margin of victory for the opposition being between 1.21% and 2.02%, depending on whether you look at part A of the referendum or part B (there were 2 separate referendums on the ballot), we can see that IVAD again slightly underestimated the amount of support for the pro-Chavez option. Now, regarding the upcoming elections, their Sept. 2010 poll reflects 54.3% support for pro-Chavez candidates, and 45.7% for those of the opposition. Assuming polls by IVAD continue to be relatively accurate, this could be a very good sign (and even more so if they continue their trend of underestimating the electoral support for Chavez and his candidates/referendums). 

One clearly government-aligned company is GIS XXI (historically run by pro-Chavez former government ministers). Their polls on the 2009 vote indicated approximately an 11% to 16% lead for the pro-Chavez camp, with their final poll predicting just under a 13% margin. Not terribly far off from the final result of 9.73%, but possibly a bit inflated, as some might have expected.  Regarding the upcoming election, their most recent polls show 52.6% for pro-Chavez candidates and 47.4% for the opposition candidates, a margin of just 5.2%. That these relatively close numbers are coming from this generally pro-Chavez company is the most concerning of all the polls in my opinion.

And now for the polling companies closely aligned with the opposition. Probably the most important of the opposition polls to examine would be Datanalisis. In 2009, they predicted landslides for the opposition side in 3 consecutive surveys, until the final one (supposedly taken in mid-January), when they predicted 51.5% for the “Yes” vote and 48.1% for the no. The results of their two polls from the previous month (one taken in Dec. 6 – 16, and another whose results were published on 1-22-09 but do not say when they were taken) indicate that in a single month, the “yes” option had a massive surge in support of between 15.3% and 21.8%, while the “no” vote lost about 5% of support during that period.  This drastic change in such a short time period does not seem to speak well to the credibility of their polling methods. Nevertheless, their final poll still underestimated the amount of Chavez’ electoral support.

Now if we take a look at the 2010 elections, in Datanalisis’ most recent poll (announced by their director, Luis Vicente León, at the Miami “Conference of the Americas” last week), pro-Chavez candidates have a lead of exactly 4.0%, 52.0% to 48.0%. He was quoted in numerous articles featuring these results as saying: “It’s clear the country is divided in two equal parts.”  Interestingly, before the 2009 constitutional amendment vote he presented results that showed extremely similar numbers (51.5% to 48.1%), and was quoted in Reuters as saying virtually the exact same thing: “This reflects a country divided into two, basically equal parts” (It turned out that it was not equal at all, with 1.1 million more voters choosing “yes” over “no”).   It all could make one wonder if publishing these numbers as their final poll before an election is possibly less the result of honest polling and more a strategy intended to influence the election, possibly by showing the opposition as more popular than they really are (but still just behind) so as to encourage turnout within their ranks. In the past, Datanalisis has refused to disclose its exact figures and methodology of some of their polls, stating that they are commissioned by private groups/individuals and only they are privy to that information. 

One of the more humorous opposition polling companies would be Keller y Asociados. Two months before the 2009 referendum vote they predicted a 37% margin of victory for the opposition. They did not do any further polling. In the current election, they are giving the opposition a 14% lead in terms of overall nationwide support, as of August. This company is clearly bogus and their relatively moderated numbers are actually a bit encouraging (one might expect they’d take advantage of recent troubles in Venezuela to really push the belief that Chavez and his movement is increasingly unpopular).

I’ll also briefly mention that the polls of opposition-aligned Hinterlaces, in the months leading up to the 2009 vote, showed a trend that saw the opposition’s lead decrease from an early lead of 30 points to just 5 points in the two weeks before the vote. With the final margin being nearly 10% the other direction, one might wonder how they stay in business. In their latest 2010 election poll, taken the last week in August, the opposition has just 4% lead in their polls. Considering how clearly biased this polling company is in favor of the opposition, these results seem somewhat encouraging. 

Based on this simplistic review of the recent accuracy of these polling companies, it would seem that the pro-Chavez candidates will likely receive a majority of the total votes cast in the elections next Sunday. The major difference in this case, of course, is that the vote is not centered around whether or not Chavez can or will remain in power, and the big wild card is how that will impact turnout (although some of these trends, as with IVAD for example, seem to hold even during the 2008 regional elections that similarly did not threaten Chavez’s presidency). Regardless, it seems clear that the electoral climate is not as negative for Chavez’ movement as the local and international media were insisting for months, despite the many unfortunate developments for the government this year and likely effective U.S.-backed opposition/media campaigns to capitalize on them. 

Polls leading up to the 2-15-2009 Constitutional Referendum vote (Si is pro-Chavez). The final result was 54.86% in favor to 45.13% opposed, a margin of 9.73%)

Hinterlaces

Intención del voto sobre la enmienda constitucional (%)

Fuente

Fecha de estudio

Fecha de publicación

A favor (SI)

En contra (NO)

Indeciso/ otros

El Universal

Diciembre

10 de diciembre de2008

32,0

62,0

7,0

Unión Radio

Diciembre

6 de enero de2009

32,0

61,0

Unión Radio

Enero

30 de enero2009 de

45

55

El Universal

Enero (25 al 28)

7 de febrero2009de

44

49

7

Datanálisis

Intención del voto sobre la enmienda constitucional (%)

Fuente

Fecha de estudio

Fecha de publicación

A favor (SI)

En contra (NO)

Indeciso/ otros

La Verdad

Noviembre

11 de diciembre de2008

25,5

56,0

15,0

El Nacional

Diciembre

3 de enero de2009

36,20

53,50

9,8

Tal Cual

Enero

22 de enerode 2009

29,7

52,8

Reuters

Enero

28 de enerode 2009

51,5

48,1

GIS XXI

Intención del voto sobre la enmienda constitucional (%)

Fuente

Fecha de estudio

Fecha de publicación

A favor (SI)

En contra (NO)

Indeciso/ otros

ABN

Diciembre

16 de diciembre de2008

53,0

42,0

5,0

ABN

Diciembre

3 de enero de2009

51,7

39,1

ABN

Enero

29 de enero de2009

55,9

40

4,1

YVKE Mundial

Enero (15 al 22)

1 de febrero de2009

52,9

40

7,1

IVAD

Intención del voto sobre la enmienda constitucional (%)

Fuente

Fecha de estudio

Fecha de publicación

A favor (SI)

En contra (NO)

Indeciso/ otros

ABN

Diciembre (12 al 22)

4 de enero de2009

48,07

42,01

9,92

Televen

Enero

18 de enerode 2009

48

43

9

ABN

Enero (13 al 20)

25 de enerode 2009

54

45,9

YVKE Mundial

Enero (22 al 28)

1 de febrerode 2009

47,7

42,2

10

ABNReutersPrensa Latina

Enero/Febrero (28 al 02)

5 de febrerode 2009

47,5

39,5

13

Otras

Intención del voto sobre la enmienda constitucional (%)

Encuestadora

Fuente

Fecha de estudio

Fecha de publicación

A favor (SI)

En contra (NO)

Indeciso/ otros

Keller & Asociados

Unión Radio

Diciembre

15 de diciembre de2008

31,0

68,0

Consultores 21

La Jornada

Diciembre

7 de enerode 2009

41,80

56,20

Ceca

Diario La Verdad

Enero

28 de enerode 2009

38,8

49,3

11,9

Naorinc

ABN

Febrero

6 de febrerode 2009

54

38

8

______________________________________________________________________________________ 

Encuestas nacionales sobre las Elecciones Parlamentarias 2010

Intención del voto por bando político (%)

Encuestadora

Fuente

Fecha de publicación

Oficialismo

Oposición

Indeciso

GIS XXI

Radio Nacional de Venezuela

marzo de2010

32,0

22,0

36,0

GIS XXI

Radio Nacional de Venezuela

mayo de 2010

36,0

23,0

33,0

GIS XXI

Radio Nacional de Venezuela

junio de 2010

44,0

20,0

Hinterlaces

Hinterlaces El Universal

julio de 2010

27,0

28,0

23,0

GIS XXI

Correo del Orinoco

agosto de2010

50,0

44,0

N/D

Keller & Asociados

El Universal

agosto de2010

32,0

46,0

N/D

Hinterlaces

Últimas Noticias

agosto de2010

37,0

41,0

17,0

GIS XXI

Telesur

septiembre de2010

52,6

47,4

N/D

Datanálisis

ABC.es

septiembre de2010

52,0

48,0

N/D

IVAD

Unión RadioDescifrado

septiembre de2010

54,3

45,7

N/D