SWS and Pulse Asia’s methods or why you’ve yet to answer election surveys

With about 60 million adult Filipinos across the country, many must wonder about their chances to be picked as a respondent in surveys being conducted.

 

Experts interviewed by GMA News Online admit that the chances are pretty slim. One would have better luck in lotto, or even in love, they say in jest.

 

"Yes, napakahirap talaga at napakaliit ng chances ng isang average Filipino na ma-interview," said Leo Laroza, senior survey research and communications specialist for the Social Weather Stations (SWS). 

"Maliit ang chance na mapili ka doon, mas malaki pa ang chances ng isang tao na tumataya sa lotto na manalo kaysa makuha ka bilang respondent sa isang survey," he added.

 

He added: "Sa katunayan, wala pa sa regular staff ng SWS ang nai-interview bilang respondent. Ganoon kaliit ang tsansa."

 

But it's more than just a matter of fate since a systematic approach comes into play when selecting who they interview for their surveys.

 

Both the SWS and Pulse Asia usually go by a sample of 1,200 respondents for their surveys, composed of adults (those of voting age, 18 years old and above).

 

In a piece that she wrote for GMA News Online, Pulse Asia associate research director Pia Bennagen Raquedan said this gives an adult Filipino less than one percent chance to be included as a respondent. 
 

"Ngunit dahil sa ang proseso ng pagpili ng mga respondents sa isang survey ay kadalasang mas mabusisi kaysa sa simpleng random sampling (i.e. may karadagdagang prosesong pinagdaraanan pa), lumiliit pa lalo ang tsansang ito. (Sabi nga, mas malaki pa ang tsansang matagpuan mo ang iyong 'soulmate' kaysa mapili kang respondent sa isang survey)," she said.

 

Selecting the areas

 

According to Laroza, 1,200 is an "internationally accepted sample size" and it's just a matter of using "a systematic way of distributing it."

 

For both SWS and Pulse Asia, the sample size is divided among four major areas in the country: Metro Manila, the rest of Luzon, Visayas, and Mindanao. The 300 respondents per major area will come from 60 barangays or precincts. 

"In one precinct, we get five respondents," Laroza said.

 

The barangays or precincts are selected among 17 cities and municipalities for Metro Manila, and 15 in the rest of Luzon, Visayas, and Mindanao, based on population size.

 

Raquedan noted that at least one barangay must represent every city or municipality chosen for the survey. 

 

The areas are selected through a comprehensive list and street maps, as available from the Commission on Elections and even the National Statistics Authority.

 

'Right coverage'

 

Once the areas to be covered by the survey are determined, the survey firms send out their field interviewers, armed with spot maps for their designated areas, which also indicate the landmark, random start, and interval for selecting the houses.

 

The landmark, the field interviewer's randomly-selected "predetermined starting point," can either be a church, a public school, the barangay hall, or the house of the barangay captain, Laroza said.

 

"Doon sa map na iyon, naka-identify na kung anong corner magsisimula at kung anong random start niya -- pang-ilang bahay ang una niyang pupuntahan," he pointed out, adding that a random start ranges from 1 to 6, determined in the headquarters of the survey firm.

 

The interval, meanwhile, is six households in urban areas and two in rural areas. This means that in urban areas, the respondent will come from every seventh house, and from every third house in rural areas.

 

Raquedan also pointed out that the number of households is not based on the physical structure, but by the number of households in one.

 

"Ang ating binibilang ay sambahayan o household at hindi bahay o building o kung anupamang istraktura. Ibig sabihin, kung sa isang bahay ay may nakatirang limang sambahayan, ang bawat isa sa kanila ay kasali sa pagbilang kahit nna sila ay magkakasama sa iisang bubong," Raquedan said. 

 

For SWS, they also have a "way of walking" in a barangay, which they call "right coverage."

 

"In our case, there's a way of walking around that area when you're counting the houses.... It simply means you point your right shoulder towards the entrance of the houses, or yung madadaanan mo and start counting like that. Ganoon din with multi-dwelling houses, sa pag-akyat halimbawa, ganoon din," Laroza said. 

 

Gender, 'Kish Grid'

 

When a field interviewer ends up in a household, this doesn't mean she can just talk to whoever answers the door or agrees to be interviewed.

 

"Ang bawat questionnaire ay may kaakibat na numero. Ang mga odd-numbered questionnaires ay para lamang sa mga lalaking respondents at ang mga even-numbered questionnaires ay para naman sa mga babaeng respondents," Raquedan said.

 

Both SWS and Pulse Asia also use what is called a probability selection table, also known as a "Kish grid," which helps them determine who will be selected in the survey.

 

In the grid, the field interviewer will take down the names of all the male (or female) members of the household by age, with the eldest first in the list. Each questionnaire comes with a "household number," as determined at the headquarters.

 

The interviewee will be determined by the number where the columns of the last name in the list and the household number meet.

 

'Callback'

 

Another catch in the process: Just because the person isn't there doesn't mean the interviewer will go ahead and select someone else.

 

"For example, napili si Pedro. If Pedro is there, then good for the interviewer... but in some cases, wala si Pedro dahil nasa trabaho, there will be callbacks. Sa urban area, we allow as much as three [times], or as much as the project will allow before mag-decide na i-replace na," Laroza said.

 

He added that for rural areas "na medyo mahirap na transportation, we're only allowed one callback—babalikan usually in a week or within the day."

 

"But if the household will say, 'Itong taong ito, hindi mahahagilap within that time, then we'll have substitution. The substitute household must be similar to the economic class of the household it replaces," Laroza said. -NB/JJ GMA News