The Philippine Online Chronicles

Third Quarter 2009 Social Weather Survey: Noynoy Aquino and Manny Villar top the people's "three best leaders to succeed PGMA in 2010"

A superficial glance at the Social Weather Survey’s (SWS) third quarter report shows that Benigno “Noynoy” Aquino III is leading the polls, garnering 60% of the respondents “votes.”

Every quarter, this uncomissioned national survey asked 1,200 randomly selected Filipino citizens of voting age a number of questions in order to obtain the “people’s” social attitudes and political preferences. (I place people in quotation marks because 1,200 respondents are too small to be properly representative of the Filipino people.) The second and third quarter report included the following question on the people’s preferred candidates for next year’s Presidential elections.

Manolo Quezon remarks that the SWS survey results are useful for identifying which of the candidates are the real contenders for the presidency.

Personally, I think the survey results are useful not in terms of presenting a snapshot of how voters might vote, were elections held today, but rather, who, in the public mind, are the real contenders at this point in time. I’ve suggested in the past that we’re seeing a return, in voter attitudes and orientation, towards seeing the presidential contest as a two party fight, which is a more natural order of things as far as a presidential system with no runoff elections is concerned.

From this perspective, the fight for the presidency is between Benigno Aquino III and Manuel Villar Jr. The other candidates will determine who gets to shave off votes from the leading contenders and who ends up losing more votes to the minor candidates more than the other.

However, the methodology used by the SWS was too disorganized and failed to obtain the actual value voters place on each candidate.

The first thing one might notice is that the percentages of the survey add up to 300% rather than 100%. Philippine Commentary explains that the statistical anomaly is due to the methodology employed by the SWS. It turns out that there were no fixed names of candidates in the survey questions; the interviewers left it up to the respondents to come up with three candidates of their choice. The survey then counts the number of times each candidate’s name is mentioned and divides this by the number of respondents to obtain the percentages.

To obtain a better understanding of how the responses were tallied up, let’s imagine how the survey was conducted. Five people were interviewed and asked to name three candidates. Person A, B, and C each pick "Noynoy, Villar, [Joseph] Estrada". Person D says "[Loren] Legarda, Bayani [Fernando], Estrada". Person E says "Chiz [Escudero], Noynoy, Bayani". The interviewer then counts the number of times each name was mentioned, giving us the following figures (Table 1).

In order to obtain the percentage, each name is divided by 5 (the number of respondents). This leaves us with the following figures (Table 2).

There are two ways the SWS survey could have presented a more comprehensive picture of the election’s real contenders. The first involves a complete change in methodology. Instead of asking the respondents to give three names out of thin air, SWS should have limited the choices to candidates actually running. Then they should have “tiered” the selection by giving a value to each choice.

For instance, the first choice gets three points, the second choice gets two points, and the third choice gets one point. This way, you can easily see the real value of each candidate; it’s highly likely that the respondents’ votes will go to their respective number one choices. Limiting the choices and doing tiered selections also eliminates the problem of survey responses that fall “less than the threshold.” There is no limit to the answers you can get from an open-ended question like “Who do you want to succeed Arroyo?” According to the Philippine Commentary, around 50% of the respondents’ answers went to someone “less than the threshold”, or an unknown third option that could be anyone in the country – Willie Revillame, their Tito Boy, even themselves.

There’s another way to make better sense of the data even with the flawed methodology used by the SWS. Instead of obtaining the percentages by dividing the votes by the number of respondents, they should have divided the votes by the total number of candidates’ names mentioned. In the example given above, the respondents named fifteen candidates. Out of those fifteen names, Noynoy’s name was mentioned 4 times. Divide four by fifteen, and you get 26.67% - Noynoy’s actual value.

Image source: SWS Media Release.

---

Add this page to your favorite Social Bookmarking websites