The Polling in Perspective

0

Over on his excellent Venezuelan Poll Nerd blog, YV Polis,  political scientist Iñaki Sagarzazu puts the latest C21 numbers in perspective. The real take-away from me, is in his last slide:

Yes, Chávez leads Capriles in almost all recent polls, but there’s a big split between some fly-by-night, Johnny-Come-Lately or plain-old-fake pollsters, and more established polling organizations. Basically all the supersized, 25+ point Chávez leads come out of maletín-based pollsters.

In fact, I’d argue that that slide actually understates the split. Iñaki counts Varianzas as a a new pollster, but they’ve been around longer than most, it’s just that their polls used to be branded Evans/McDonough.

Take them as an established pollster, and the bankruptcy of the chavista fake pollster operation would be even more visible than it is in that chart.

I hope I get to collaborate closely with Iñaki in the next few months on this stuff. He has the quant skills and the obsessiveness it takes to really drill into the murky world of Venezuelan polling. More to come…

1 COMMENT

  1. I don’t mean to be a party-pooper, but an almost 50-50 split will not get rid of Chavez. Too close for comfort, and even a Capriles win by a small margin will most likely result in cries of recount and various accusations of cheating by the Chavistas. I mean who wants to lose a job full of perks and unlimited kleptocracy? This is a doozy, and unless Capriles gives a decisive spanking to Don Chavez I fear not much will change.
    I hope I’m wrong.

    I’m a long time reader, first time poster by the way.

    • Recall Jack, that the party has barely gotten started. You are correct in stating that 50/50 gets Capriles nothing, but the trend is going in the right direction.

      Last thing Capriles wants is to peak too early.

    • If Leopoldo López does his job and lands signed copies of every acta de totalizacion (polling-station specific tally sheet) and every acta de auditoria (hand-audit tally sheet) – there’s just no practical way to cheat numerically without it being very, very obvious. Maybe they’ll try very very obvious, but I don’t think it can stick.

      • Hate to be negative about this, but… ‘very, very obvious’ has been the modus operandi of this government for a looong time! Plus, they can back up any VVO maneuver with instant-like-a-Polaroid rulings from the TSJ. Finally, they will denounce any international condemnations of the election tampering with cries of: foreign meddling, sovereignty, imperialism, and all that crap. Anywho, that’s what I would do in their place.

      • Very very obvious can stick. The Chavistas have global support. The liberals (one we might generally expect to defend democracy) would just turn a blind eye. I don’t see any way that they can resolve the problem unless those who orchestrated an obvious polling cheat just came out and admitted to it.

  2. I’m sick of reading news about the polls and Chavez’s lead. For me, they are extremely demoralizing, a freaking bucket of cold water in my face. I don’t know if the pollsters are wrong or biased but for some reason or other I always go back to the constitutional referendum when Chavez first proposed indifinite reelection. I remember Luis Vicente Leon stating in various occasions to the media that Chavez was going to be defeated and indeed he was. Why not believe him this time? The guy’s good at his job.

    Don’t take me as a cynic, or a defeatist. I believe in the work of the polling companies (the recognized ones, to be precise) because at least in Venezuela they’ve had a good historical track record. You really think they’ve sold themselves to the devil?

    With all this said, I just hope for the sake of our country, that they are wrong this time.

    • Again: standard disclaimers apply. Polls measure public opinion as it was at the time when the poll was made – they’re not predictions, they’re not meant to be predictions, they can’t predict what public opinion will be like at the time of the election.

      This is a basic fact, but people seem to struggle with it…

      • Alex,
        There are still three months to go before the election, no need to be demoralized. The trend is what counts.

        • Dear Sir, Unfortunately, it’s not the trend that counts; it’s the CNE, and that, under the watchful eye of the Vene-Cuba optic cable. In the stark shadow of that situation, all the polling results seem to be more of a red herring, being maintained primarily with a view to giving some air of verisimilitude to the results to be declared.

        • Agreed. I think the trend is on our side. Also, there’s that “Loch-ness cancer” of his (many talk about it, yet no one can factually confirm it) that may, or may not, work in our favor.

      • People struggle with it due to the significant variables from poll to poll:

        the size and location and diversity of n
        the questions asked
        whether it’s raining or not
        whether people are running off to an appointment, or to sit down at the table to eat
        el fastidio factor
        who commissioned the poll ..

        Enfin, un sinfin de cosas.

      • Be CCS the venezuelan DailyKos or RealClearPolitics, where you can track averages of polls in the front page. That´d be sweet

          • Estoy totalmente de acuerdo — más que una tremenda contribución a la crítica de la política venezolana, por fin tenemos el efecto lupa sobre las encuestas. Mil gracias y bienvenido, Iñaki!

  3. Guys, these polls are starting to drive you crazy. Here’s a sound recommendation: look at the trend in voting spreads in elections deciding the presidential office since 1998. Anything which strays too far from that trend isn’t to be trusted.

    I’m going for 60-65%.

    • Agreed. The local political-climatology is the best (and cheapest) indicator of future outcomes. So 63%-62% is possible. However the Marine forecast can change everything, luckily due to a cooling North American Id the chances of a batallion landing team are vanishingly small.

    • Yes, you could be right and I, as an ignorant PSF, think you are. In numbers around 9 million votes+ for Chavez and about 5 million+ for Capriles.

      Today is a great day – Full Membership of Mercorsur (the details don’t matter as today is celebration day) and Globovision paid its fine of Bs.F. 9.3 million for breaking the broadcasting laws. I am pleased that Globovision is finally seeing sense and obeying Conatel.

      Sunday will be an even better day when I see el Comandante on the tarima in Maracay in the opening of his electoral campaign!

      Enjoy it guys. You only have to keep up the self delusion for another 100 days until the bucket of icy water hits you.

      • Boy, who needs Radio Rochela??? Maybe they will televise “El Comandante” in Maracay lying down, but then turn the image 90 degrees to make him look like he’s standing up??

  4. It may be too soon to judge that latest data point. Most other points on the line are resulting from several measurments. The last time a measurement was this close to the red line, the blue line was well below because of the other measurements. If, ignoring this latest point, we look at the trends of the lines previous to that latest point, it almost seems like the blue line has been declining for quite a while, and the red line on the rise, at least in the historical pollsters’s graph.

  5. Your comment about Varianzas is correct. This pollster directed by Rafael Delgado Osuna has been around for a long time.
    Quote from analitica.com, Oct. 2005
    http://www.analitica.com/va/politica/opinion/8865452.asp
    “No podemos obviar dentro de estos estudios el trabajo laborioso y sostenido en el tiempo por Rafael Delgado Osuna, profesor de Sociología Electoral de la UCV y ex Director de Estadísticas Electorales del otrora CSE e igualmente los trabajos y estudios de sondeos de Alfredo Keller y asociados.”
    Now, it would be interesting to see the above Iñaki graph adjusted with the fact that Varianzas is a “Encuestadoras Historicas”

Leave a Reply