Could the responses of 1,200 respondents in Pulse Asia’s survey on senatorial preferences this early in the May 2013 electoral game reflect the sentiment of 80-90 million Filipinos? Could the preferences of some 1,200 respondents be extrapolated to mean “Filipinos’ Senatorial Preferences for the May 2013 Elections” as the title of the survey said?
Consider the text of the Pulse Asia statement about the survey:
“If the May 2013 elections were held in early November 2011, fifteen individuals would have a statistical chance of winning a senatorial seat. Most of the probable winners are either former or incumbent members of the Senate. Emerging in the top spot is Senator Francis G. Escudero (65.6%), with Senator Loren Legarda (58.9%) in second place. Meanwhile, Transportation and Communications Secretary Manuel A. Roxas II (43.0%) is in 3rd-4th places. Also in 3rd place (with his lowest showing being 5th place) is Senator Alan Peter S. Cayetano (40.3%). Completing the top five is former Vice-President Noli de Castro whose overall voter preference of 34.8% puts him anywhere from 4th to 8th places.
Sharing 5th to 12th places are San Juan City Representative Joseph Victor Ejercito (30.4%), Senator Gregorio B. Honasan (29.6%) and Cagayan Representative Juan Ponce Enrile, Jr. (29.5%). Senator Aquilino Martin Pimentel III (29.4%) and Justice Secretary Leila M. de Lima (29.4%) land in 6th to 12th places. Other probable winners are Senator Antonio F. Trillanes IV (28.7%), former Senator Juan Miguel F. Zubiri (26.9%), Aurora Province Representative Juan Edgardo M. Angara (24.3%), former Senator Ana Madrigal (24.0%) and former Senator Richard J. Gordon (22.1%). At best, these individuals would find themselves in 6th to 12th places but their lowest statistical rankings – 14th to 20th places – would put them out of the winners’ circle. Less than one in ten Filipinos (5.2%) does not have/refuses to name any preferred candidate for the May 2013 senatorial elections.”
The survey is, in effect, equating the preference of a mere 1,200 to that of the Filipino people. That’s a lot of extrapolation.
Come to think of it, would another set of 1,200 respondents answer the same way today? How many sets of 1,200 people are there in 80 million? One could have 66,667 completely different sets. Could it have been safer to say that the survey reflected the sentiments of only 1,200 Filipino voters?
I came to ask these questions because a number of friends I talked with recently had a different set of preferred senatorial candidates. I could devote some time to ask 1,200 people but I am not in the same business as Pulse Asia or any other survey agency that will make their own surveys soon.
But wide publication of such surveys would influence others and create a trend. After the bombardment of regular speculative conclusions from such surveys come May 2013, these survey agencies would then claim they were right all along.
- Lies, damned lies, and statistics 2: extrapolation (filipspagnoli.wordpress.com)
- Support from the Philippine South? (politika2013.wordpress.com)