March 26, 2013

2013 elex: SWS survey beatdown

You already know it's election season already, right? Election season won't be complete without our election surveys. SWS, Pulse Asia, even the likes of StratPolls and the Center, and whatever "scientific" survey I can get my hands on. Our pollsters aren't plentiful, and they only release monthly surveys on election seasons, so there aren't that many surveys we can dig data upon; in fact, most of the time, the media only reports surveys published by SWS and Pulse Asia.

What's the point of this post, you ask? It's about time to aggregate all of these surveys into one survey of surveys, or the "poll of polls", one figure you can use to represent all surveys within a certain time period. What separates a non-scientific survey from supposedly "scientific" ones? It lies in the sampling. In non-scientific surveys, like phone-in surveys, Facebook and internet polls and the like, not everyone (or every voter) has an equal chance of participating. For example, those who do not wish to join the survey, or those who are physically (did not watch TV show that encouraged viewers to vote) or technologically capable (does not have Facebook) would not have been given a chance to participate. This is opposed from scientific surveys where everyone (or every voter) supposedly has an equal chance of participating.

There lies the caveat: sampling is the heart and soul of surveys. If the population was sampled wrongly, such as getting a "heavier" sample than what would've been recommended from places where a candidate particularly strong, this will distort the figures of the survey. That, and what happens between from data gathering, to publication.

But aren't surveys condition the minds of the voter? In a perfect world, it shouldn't. In fact, the only minds the surveys should be conditioning are the campaigns per se. Surveys are meant to determine where the candidate stands; this allows the candidates' campaigns to strategize on what gimmick to do.

Now, for the dirty stuff: SWS just released their March 2013 survey. They interviewed 1,200 people in a matter of two days throughout the country. Some notes:

  • Many are harping the sharp drop of Chiz's numbers. From 62 to 48; in fact, SWS implies that Chiz is now at 4th, as his name appears after JV. Chiz's numbers have been constant at the 61% to 62% range, and consistently second behind Loren.
  • Alan's numbers were statistically identical (down by 1%) from the previous survey, but he jumped from 3rd to 2nd due to Chiz's massive drop in the latter's preferences.
  • One thing constant in the surveys is the separation between #2 and #3 in the survey. It still holds true now, although it is now reversed: Alan is #2 and Chiz is now tied for #3. JV's numbers went up (6%) from the last SWS survey, but he jumped from 9th-10th to 3rd-4th. This comes after a massive drop for his preferences on February vs. his January numbers, which are his personal high. This also means those ranked from 4th to 13th are separated by the thinnest of margins.
  • Honasan, who is running a guerrilla campaign, jumps back into the Magic 12 after losing ground in February. Honasan is traditionally an independent, and doesn't participate in "air wars", but joined UNA this year as a full candidate.
  • Bam's numbers are unchanged at 42%, interestingly, 42% of the voters voted for Noynoy in 2010. Is this the peak of his support?
  • MTRCB's numbers saw a noticeable drop (8%) that saw her ranking fall from 5th-6th to 11th. Amongst the three erstwhile UNA guest candidates, only Loren Legarda, and her army of Facebook commenters, managed to hang on to their respective positions; even Loren's preferences herself went down by 5%.
  • JunMag's preferences are steadily going up. Though its rise is not as spectacular as Bam's, Nancy's, or MTRCB's, the son of the former beloved president is within the margin of error of the #12 candidate.
  • Every since the campaign started, Migz's numbers have been free falling. From 47% in January, 39% last month and 35% now.
  • The Team Patay label on Risa seems to have worked on her favor as her preferences are rising steadily, from 21% in December to 29% now. The 13th-place candidate is at 39% and unlike in 2010, she might not make it even to the margin of error, if her 2010 voters will vote for her again in 2013. In senatorial elections in presidential years, the 13th placed candidate ("the quota") usually gets >30% of the vote, as opposed to midterms, where the "quota" is <35%. (For a more detailed explanation, see my PEX post.)
  • UNA's tailenders, Tingting, Ernie and Mitos, who have to move if they want to win. Tingting has to show the big guns in Mikey Cojuangco and even Dudut Jaworski, or even the Big J himself. Ernie and Mitos would also have to step up to catch the big boys. Tingting and Mitos have been constant with their preferences at the mid-teens, while Ernie's preferences drop with a tie with Hontiveros at 25% on January.
  • Bro. Eddie's and Hagedorn's preferences are constant at the 10%-15% range. Everyone whose numbers are below Maceda's are stuck in a muck.
  • Teddy's numbers are hovering in the range of votes received by Liza Maza and Satur Ocampo in 2010. Perhaps that's the peak of the support of Left in this country for the meantime. If Teddy wants to win, he has to attract support outside his base.
  • It's a massive battle for scrubs between Ang Kapatiran, DPP and the independents below 10%. These haven't broke through above 5% in any survey (except for JC de los Reyes in a StratPolls survey where he got 12%). It remains to be seen if AKP's numbers would go up due to the publicity garnered by the Team Buhay/Team Patay campaign.
In the coming days: the aggregate poll of polls. All of the polls in one handy statistic you shouldn't use.

No comments:

Post a Comment

Comments are absolutely not moderated. Comments are displayed immediately once posted. Comments can be only be removed by the author (if signed in to a Goggle or OpenID account) or if requested by someone else with good reason.