Wednesday, October 14, 2009

Why SWS Presidential Survey Does Not Add Up To 100% But 300%

Facebook messages have proclaimed a new King of the SWS Presidential Survey:

At Midfield: Noynoy (60%), Villar (37%), Erap (18%), Chiz (15%), Gibo (5%)
TOTAL: 135% 
Juan Country (Philippine Commentary): Noynoy (60%), Villar (37%),  Estrada (18%) Esucdero (15%) Roxas (12%) Teodoro (5%) Lacson (4%) Fernando (1%) Teodoro (1%)
TOTAL:  153%

But savvy readers are already scratching their heads and wondering why these numbers don't add up to 100% as one might expect and instead exceed 100%.

This is not quite the kind of survey that people think it is, because above two bloggers have not made any kind of mathematical or arithmetical errors and have faithfully posted information found elsewhere on Twitter (@Leah Navarro) and other sources.   But the simple fact is the percentages reported by SWS should not add up to 100%. They should add up to 300%.  And therein lies a statistical tale of great and grave interest to us all...

What SWS does not emphasize however, is that in this very important survey, they appear to "throw out" or at least not report ABOUT HALF OF THE RESPONDENT DATA, which they do mention in a very tiny note on their website:  SWS does not report any candidate names by less than 0.14% of the respondents.  Confused?  Read on...

Every three months, the Social Weather Stations conducts its regular, uncommissioned Quarterly National Survey. SWS asks 1200 randomly selected voting age Filipinos, a variety of questions, including those that usually get reported as hunger and poverty statistics, social attitudes, and of course political preferences. Since September 2007, the Social Weather Stations has included the following question in its Quarterly survey as part of its public opinion polling on the 2010 national elections:
"Under the present Constitution, the term of Pres. Arroyo is up to 2010 only, and there will be an election for a new President in May 2010. Who do you think are good leaders who should succeed Pres. Arroyo as President? You may give up to three names."




SOURCE: Social Weather Stations 2nd Quarter 2009 Press Release

Above Chart summarizes over two years of SWS polling. The raw data (all of it) is contained in the Table below:



Please notice that about ONE HALF of all the data, and by far the largest percentage of the choices goes to someone named "LESS THAN THE THRESHOLD", which SWS pegs at 0.14% (if you carefully comb through the graphics in the link I supplied to SWS above you can find this!)  My take is that when the respondents get to that third possible choice, they end up naming their father, mother, wife, uncle, mayor, themselves! or some other unknown that together with everybody else's third choices usually make up to half of the SWS data!

But as it should, the data does add up to 300% in all columns.   I have a theory I cannot prove, that it is actually the second placer in these  SWS polls that represents a kind of plurality choice.  Noynoy has come out of nowhere (less than 0.14 percent!) to take a whopping --but sympathy confounded--60%.

But the claimed Margin of Error of plus or minus 2.5 percent is SPURIOUS, as one can easily verify that the reported percentages do not add up to 100%! Since there is no fixed menu of candidates given in the question, and the respondents were free to name up to three choices, this form of statistical survey cannot be conventionally analyzed and its accuracy ascertained using the methods that are applicable to more conventional surveys. 

The more realistic measure of how accurate this series of polls is likely to be can be better seen in the numbers of one particular "CANDIDATE" that is never mentioned in the SWS spin of its own data: namely that of NONE or DON'T KNOW, which an inspection of the above chart will show actually competes with the leaders of the survey, and in fact, apparently topped the First Quarter SWS survey. But this fact never made it to the headlines.

Indeed, I would submit that the REAL STATISTICAL ERROR in the SWS survey is at least equal to the  NONE or DON'T KNOW percentage.

11 comments:

Anonymous said...

SO what SWS says is that 60% of ignorant pinoys like a mama's boy with weak character, 37% like money bags who is bent on buying the presodency, this is followed by a plundering hedonist at 18%, Lets not forget the click snake oil salesman that has entranced our gulible youth all the while working for Pacman and his chinese financial wizard. I dont even want to mention the others.. Now where's that line to the canadian embassy?......

Anonymous said...

Sorry I meant the young but SLICK snake oil saleman who is preying on our youth.. AMong the candidates He scares me most.. id someone move my CHIZZZ????

manuelbuencamino said...

DJB,

Tama ba if I read the survey as Noynoy's name appeared as one of three choices of 1080 of the 1800 respondents polled and Vilar got 666 and so forth down the line?

Deany Bocobo said...

mb,
i haven't seen the sws writeup only the bloggers's posts. is it up on their website na? But what does it mean when a survey's numbers add up to 300 percent do you think?

Remember they did not ask, "If the election were today, who would you vote for as President."

This is just the nomination phase of a Beauty Contest, di ba?

Deany Bocobo said...

Anon: meowwwr !

manuelbuencamino said...

DJB,

That's why I was asking if it could be read as x number of 1800 wrote the following names. Kasi nga tatlong pangalan ang pwedeng isulat ng bawat repondent. Maybe SWS should have reported the results in terms of numers instead of percentages?

Deany Bocobo said...

MB,
It's the spin being spun about what the numbers mean.

They certainly don't mean that Noynoy would get 60% of the vote if the election were held today, because that isn't the question they asked.

Also, you don't think Cory's death and funeral might have something to do with this?

Let's just be straight. This is not a poll that measures presidential preference. No one votes for three presidents, so what does the survey mean??

Jesusa Bernardo said...

Amazing survey! Last month Noynoy had 50%, now it's 60%. I bet next month, SWS will release poll results showing Noynoy's number up to 70%. Perhaps, the month after it would be 80%....

Anonymous said...

To get the statistical graph of the Best Leader to succeed PGMA in 2010. (From 2007-2009). Hence these are the presidentiable aspirants who thought will run on 2010 according to themselves. Then Noynoy came last September 2009. That is is why he was included for their quarterly tabulation.

Deany Bocobo said...

MB,
The answer to your arithmetic question is another question: were all the mentions received by noynoy first place mentions?

Anonymous said...

Poster is obviously supporting "sucha loser" Gibo or other "loser" candidates. Whatever your argument is - you can't deny the fact that surveys "almost always!" predicts the actual election outcome with only slight variation sometimes. Kung pinalalakas mo lang loob mong mga supporter ng mga loser candidates - don't waste your time. Bigo este Gibo won't even make it top top three as long as he sympathizes with Gloria.