Poll Nerd Heaven

0

This is how well-hidden these slides were...Can it really be that Ricardo Sucre’s massive trove of Chávez era polling slides has been hiding in plain sight for all these years, and I didn’t know about it? I’m embarrassed, really…though, in fact, I’m too giddy with excitement to dwell on my embarrassment.

Thanks to the trove, we can do something I’ve always wanted to do: actually put numbers to how well or badly various pollsters have performed over the years. In fact, I’ve already started.  (Some fun early results – the pollsters that came closest to calling the 2010 parliamentary election’s vote split were Consultores 21 and…the much maligned Hinterlaces, which was within 0.1% of the actual vote split way back in June polling…and Keller was farther off in his pro-oppo bias than GIS XXI was on its pro-Chávez bias!)

There’s a ton of raw data in Sucre’s data-dump, so it’s really the sort of thing we should crowd-source – it’s labour intensive scanning each presentation for take-away numbers to create the spreadsheet we’d need for this.

Got half an hour to waste? Help me!

Caracas Chronicles is 100% reader-supported. Support independent Venezuelan journalism by making a donation.

1 COMMENT

  1. “and Keller was farther off in his pro-oppo bias than GIS XXI was on its pro-Chávez bias!”

    Not a fair statement since the Keller poll came in early May and the PIS XXI came in late august.

  2. My old look at such numbers taught me that Consultores XXI was the most reliable followed by Datanalisis, except that Datanalisis always gets wishy washy right before an election and screws up. The rest always inconsistent.

  3. Nice initiative.
    I’m a pollster myself but unfortunately don’t have the time to do the job of putting together the spreadsheet.
    I’d like to say, though, that I know most of the pollsters in Venezuela and, for the most part, it’s a professional crowd. Over a decade ago, we started a group called “Sovimo” (Sociedad Venzolana de Investigadores de Mercado y Opinión). Keller was the leader at the time, but there were also representatives from Datanálsis, Datos, Consultores 21, Mercanálisis, Mercedes Hércules y Asoc. Statmark, etc. -I was Country Manager for Gallup back then-. I can say most of these firms are legitimate pollsters that abide to the international Codes of Ethics for the industry (or at least know about them and claim to follow them), have been polling for many years, and have way too much to lose by falsifying information. Reputation is the industry’s “blood”. You lose it, you’re pretty much dead.
    Having said that, you may encounter variations in poll results due to differences in rigor of the sampling procedures or supervision of the filed crews, among other root causes. Pretty much every pollster in Venezuela (or in Latin America, for that matter) subcontracts interviewers to the same fieldwork shops. So, the interviewers will be either the same or very comparable in most cases. The difference comes in the supervision of fieldwork execution and, again, in the quality of sampling. In order to save cost, a pollster may choose not to enforce (or verify) a good coverage of smaller towns, rural areas or even urban areas that are hard to access. Others may use a sub-optimal sample design (i.e. without a proper stratification by population size, electoral circuit etc.). In sum, there are numerous potential sources of bias (be it sample or non-sample related) that can cause seemingly equal polls to produce different results. Obviously, there are pollsters in Venezuela that are not independent or legitimate. The paradigmatic case is the one led by Jesse Chacón (since when is he a pollster?), but there are others that pop up only around electoral times.

LEAVE A REPLY

Please enter your comment!
Please enter your name here