The difficult art of Venezuelan polling

Whether we deem them worthwhile, there are elections around the corner, and pollsters trying to sell their wares. A couple of weeks ago I did discuss the problem of effective polling in Venezuela, and this week, while I had to drive many hours for work, I had the leisure to think about two recent polls that illustrate this point.

The first one came out last week end, and was from Datanalisis (1). I have expressed my concerns on this polling firm, be it from a likely weakness in their statistical and polling instrument, be it from the rather obvious political interests of its director. Still, that poll, with all due reservations, was interesting.

There was really nothing new. Chavez keeps above 70% though when asking a question about "trusting" Chavez (to rule, I suppose) this number dropped to 56%. The survey tried to look at the self qualification of Venezuelans in a right to left rainbow. A very large majority placed itself from left to center to left, yet an equal majority rejected "the socialism of the XXI century". Much more worrisome would be that barely more than 50% want a democratic form of government while more than 40% do not care as long as their needs are taken care of. Even if the data collection is shoddy, such numbers are really dangerous, even for Chavez.

I did find interesting the little table that accompanied the study. That table divides the social classes (on economic power division) in a not very convincing way. For example only 39 people of the wealthiest sector (A and B) were polled. 39 is enough for good statistics?????? How can they expect, with only 39 interviews, to reach some of the newly chavista rich that are now a common sight? At least in the poorest sector (E), the one that is most likely going to receive the positive effect of the social programs that Misiones (2) are, they interviewed 556 folks yielding the rather surprising result of still 1 in 3 not liking Chavez.

Another interesting detail, though right now of little consequence, is that in spite a strong pro Cuba campaign, Chavez seems to have attracted only a 12% pro Castro support and still has 63% against Castro. A referendum there? No fat chance!!! However it should still be a good weapon for the opposition to use against Chavez, and it speaks of its inefficiency that such numbers do not perturb Chavez high numbers.

The other poll came out in El Universal last Friday. I must confess that I tend to like Keller as a polling company, if anything because on TV its spokesperson is obviously trying to say less than what it could say whereas the one from Datanalisis seems to me to want to try to say more than what his data indicates.

The results this time show that Chavez dropped to 61% from 69% positive opinion a quarter ago while as expected the negatives went up from 22% to 28%. This is actually a plausible shift, correlating with 3 months of corruption charges in PDVSA, the inability to create jobs (the recent alleged increases, except for construction up to a point, seem to be a statistical trick to favor the government as "misiones" recipients are considered as fully employed), the aggravation of personal security perception, etc... There is no trickle down effect here from the sky high oil price: people are realizing that only a few really benefit from them, and suspiciously largely.

Still, after 6 years in office any president would be very happy with a 61%. Where the news are not too good is when the voting intention is examined. In this respect Keller does focus in its polling at likely voters only, which probably explains better the coherence of its numbers. Considering that the survey did not specify a name against Chavez, that "unknown" candidate would already score 41% to 49% for Chavez! Let's not focus on the drop from 60% voting intention early this year: Chavez has gone down to 30% in some 2002 polls to 70% early this year. What is significant here is that one would be led to think that the peripatetic president is starting to tire the electorate, an electorate that seems tired of constant campaigning, constant traveling, etc... as that poll seems to indicate: astoundingly, 71% would like a leader able to unite the opposition in a clear message!

I think that this study in contrasts shows the difficulty in doing good polling. They only seem to agree on the fading positive effects of the misiones on Chavez numbers. One reason for results differing between polls (avoiding discussing the obvious shortcomings of Datanalisis) could be simply that many people just do not want to say they oppose Chavez, in particular in lower classes, duly scared by the effects of the Tascon list. Not to mention that they see with their own eyes that the fastest way to get a Mision benefit is to apply for it while wearing a red T-shirt. Otherwise, how could we explain that still a third of "sector E" is not supporting Chavez after 2 years of the crassest populist vote buying that we ever witnessed?

Another reason is that the Chavez anti-Chavez battle is so emotional that some people seem to be able to love Chavez while being ready to vote against him if some credible offer was advanced. But so far, it seems that the message is not reaching the opposition leadership, at all.


(1) In El Nacional, no link, sorry! Subscription only.

(2) Misiones are chavismo social programs.