Third of three parts
The most fitting way to describe the opinion polling of SWS and Pulse Asia is to say that they have twisted and kneaded Philippine public opinion into a pretzel. Their polling is so contorted, you cannot separate the pulse of public opinion from the purse of election campaigns.
The contest between pulse and purse oozes from every pore of their surveys. Sadly, it’s usually the purse that gets the upperhand as far as the pollsters are concerned.
By rushing headlong to pop the question for whom a poll respondent will vote for in the May 2016 election, SWS and Pulse Asia push voters to skirt the process of weighing the candidates on their merits and their stand on issues. They skew public attention toward areas that are trivial and insignificant.
By pushing a horse-race model of the election, the pollsters pander to the desire of survey clients to exact the littlest advantage from the surveys, to keep big-money campaign donors interested, and to generate campaign propaganda.
Random sampling and question wording
These are weighty issues in the art or science of polling because wrong decisions can lead a pollster to compromise the integrity and accuracy of his/her opinion polls.
The key to the accuracy of opinion polls is, firstly, the technique of random sampling, which operates on the principle that every one should have an equal probability of being selected as part of the sample.
The science of polling involves estimation; a sample can represent the population with only a certain degree of confidence. The level of confidence is known as the sampling error, which depends on the size of the sample. In order to obtain results that are within sampling error, researchers must follow proper sampling techniques.
Accurate representation of the population, not the number of responses, is the most important feature of a public opinion survey. Indeed, as polling techniques have advanced over the last 50 years, typical sample sizes have been getting smaller, not larger.
Second, and just as critical to the accuracy and integrity of the opinion poll, is the wording of the survey question.
Pollsters can get the results they want by altering the wording of the question. Sometimes subtle changes in question wording can produce dramatic differences.
One famous example often cited in US case studies is that one month before the start of the Gulf War in 2003, the percentage of the US public who thought the US should go to war was 18 points higher in the ABC/Washington Post poll than in the CBS/New York Times poll. The difference was attributed to a few words in the survey question.
The key point to watch out for is that the questions in a survey are fair and unbiased, before one makes too much of the results. The good or the harm that polls do depends on how well the data are collected and how thoughtfully the data are interpreted.
SWS and Pulse Asia have tampered with the polling process in all kinds of ways – in the preparation of samples, in the wording of survey questions, to say nothing of the actual conduct of interviews.
I believe the great disservice of SWS and Pulse Asia in the runup to the 2016 elections is their insistence on conducting voter preference polls and treating the politics of 2016 as a horse race, instead of conducting favorability surveys to feel the pulse of public opinion with respect to prospective or declared candidates.
Needed: a favorability survey
To correct the distortions inflicted on the 2016 electoral process, SWS and Pulse Asia should conduct a favorability survey to measure the favorability ratings of the various candidates, and desist from matching them up until the candidates’ field is finalized and the campaign begins in February.
This will bring Philippine polling up to the level of what is happening in the US today.
One polling analyst, Paul. Lavrakas, has explained the favorability rating survey as follows:
“A favorability rating is a statistical indicator that is produced from data that typically are gathered in political polls. These ratings indicate whether the public’s overall sentiment toward a politician is favorable (positive), unfavorable (negative), or neutral. Journalists often report favorability ratings as part of their coverage of political campaigns and elections.
“A favorability rating about a politician is calculated by using data gathered in so-called approval questions. These questions ask poll respondents whether they “approve or disapprove of X,” where X typically is the name of a politician. The favorability rating for that person is calculated by subtracting the proportion of those interviewed who say they disapprove.”
In such a survey, poll respondents will be asked one question concerning each candidate, to wit : “Do you approve or disapprove of Candidate X (name), or his program of government?”
The response will be tabulated on whether it signifies Approval, Disapproval, or No Opinion.
At the conclusion of the interviews, the responses will be tabulated, charted or ranked as may be desired.
The pollster will then interpret the polling data to find out what they cumulatively mean for the 2016 election.
The value of this survey to the public is easy to see. The public will see how favorably or unfavorably individual candidates are perceived by voters.
Individual campaigns can of course commission their own private polls, without depending on SWS or Pulse Asia.
A statistically valid indicator
Like most surveys that predict public opinion, the favorability rating is subjective. Many unscientific approval rating systems exist that skew popular opinion. The favorability rating, however, is generally accepted as a statistically valid indicator of public perceptions of a candidate during an election season.
Once the election campaign is off and running, it becomes time to ask voters who they will vote for if the election were held today.