MANILA, Philippines -- The data that public opinion survey firms like Social Weather Stations (SWS) and Pulse Asia (PA) foist on the Filipino people can never be classified as statistics.

As I quoted from the dictionary in my last week’s column, statistics is “a branch of mathematics dealing with the collection, analysis, interpretation and presentation of masses of numerical data.”

SWS and PA collect numerical data by means of a purported nationwide field survey of a “multistage probability sample of 1,200 representative adults 18 years and above,” or 300 in each of the areas of Metro Manila, rest of Luzon, Visayas and Mindanao. The sample in each area is further divided into economic class A, B, C, D and F, or 60 respondents per class.

From these figures, the results of the survey can never be called statistics since there is the total absence of “masses of numerical data.”

Besides, how can SWS and PA claim that they conduct nationwide surveys? If they conduct one interview in each city and town, the 900 interviews in the provinces come up to only about 48 percent of the cities and towns.

US mathematician G. H. Gallup (after whom the Gallup poll is named) was the one who first came up with the premise that you do not have to take a whole bowlful to find out how a soup tastes. A teaspoonful will do just as well.

But SWS and PA claim that the results of their 1,200 interviews are the views of the entire Filipino people. This is pure hogwash. The sample is so small that to extrapolate the results as reflective of the views of the entire Filipino people is mathematically indefensible.

In other words, 87 million Filipinos make one enormous bowl and 1,200 interviews will not qualify as a teeny-weeny drop of soup. Even Gallup will agree.

Then, SWS and PA make matters even worse. By breaking the sample into the five economic classes or 60 respondents in each class, the sample is reduced to almost vanishing point. The survey can easily be manipulated. The results are absolutely meaningless.

Also, the homogeneity achieved by limiting the respondents to those 18 years of age and above is lost, especially when the respondents are asked questions on a wide variety of concerns. Adding the results becomes a mathematical impossibility. It is like adding apples, oranges, guavas, santol and calamansi.

Finally, when the survey results are released to the media, SWS and PA should divulge the name of the person who commissioned the survey. The name is very material.

If, for example, the survey is to find out whether people want the President to resign, it matters a lot whether Malacañang or an opposition senator paid for the survey. Besides, I am told, a survey costs about half a million pesos to conduct.

Lately, SWS says that its surveys are noncommissioned; subscribers pay for the cost. If you believe this, you will believe anything.

PA stubbornly refuses to reveal the name of the person who commissioned the survey, claiming that the name is immaterial. This is true only if the survey is for the private consumption of the client. For example, a manufacturer may ask PA to conduct a survey to find out if a new product will sell. The results will remain a trade secret of the client.

SWS and PA should clean up their acts. Then, their surveys would be more credible. And I do not have to write columns like this.

Previous columns:
Statistics - Part 1 – 01/01/08
What will the Senate do with Senator Trillanes – 01/26/07
Rebellion – 12/12/07
Memo to Senator Enrile on antitrust – 12/05/07