Philippine Economy Talk

[Ry_Tower]

Yes sample sizes of 1000 to 1500 are accurate enough to predict for our entire population.

As long as the sampling is done scientifically.

I myself was demoralized with this survey. But it is what it is.

Sent from my SM-N975F using Tapatalk

Guess they did the survey in his baluarte pa din. That much they can control, they know where Pduts have a strong support base from last elections' data. 91% approval rating? Wtf.

Sad but true reality nga. [emoji17]

Sent from my SM-N960F/DS using Tapatalk

[jut703]

Sa mga ayaw maniwala sa science of the statistics:

Martin Consing - The latest Pulse Asia survey reports a... | Facebook

TLDR: 1,200 sample size is sufficient to get within 95% confidence and 3% margin of error.

Survey notes also confirms that it was done nationwide and not just in Duterte's bailiwicks.

Ulat ng Bayan September 22 Technical Details – Pulse Asia Research Inc.

The latest Pulse Asia survey reports a 91% approval rating for Rodrigo Duterte and a 50% approval rating for Leni Robredo using a sample size of 1,200. Some people may be upset at this.

For a population of 109.6 million Filipinos, is that a sufficient sample size? What about a working age population of 72.8 million? What about the sample's distribution across regions?

Well, it depends on what confidence level and margin of error you're willing to settle for. Suppose at a confidence level of 95%, you have 91% of respondents in your sample saying they approve of Duterte. Then you can be 95% certain that if you asked the whole population, 91% of them would say they approve of Duterte. Typically, confidence level is expressed as a Z-score.

Margin of error is how many percentage points your survey results may differ from the real population's values that you're willing to tolerate. Suppose at a 95% confidence level and a 3% margin of error, you have 91% of respondents saying they approve of Duterte. Then you can be 95% certain that 88-94% of Filipinos approve of Duterte.

I've attached here a formula commonly used for deriving the minimum sample size given these variables. At a confidence level of 95%, a margin of error of 3%, and a population size of 109.6 million, a sample size of 1,200 is enough. This also requires that the sample be collected via random sampling, which Pulse Asia did according to their technical notes for this survey (link below).

BUT the region subsamples are too small. From the technical notes, we see that the region subsamples adding up to that 1,200 are not actually good enough when considering them individually, and the corresponding margins of error are quite high, over 20% in some cases (32% in CAR). More respondents should have been collected per region. However, it's possible that there were problems with data collection, such as travel restrictions due to lockdown, hence the small subsample sizes. This is understandable.

Personally, I would have preferred a representative sample of the Philippines at a 99% confidence level and 1% margin of error collected via random sampling which would probably result in decent region subsamples as well, but this would mean surveying over 16,000 Filipinos in total, which would be costly.

So why am I saying all this? Because it's important to understand where the numbers come from. They're not magic. They're attempts to capture reality with some degree of certainty given certain limitations, and this should be kept in mind when we read survey results. Do 91% of Filipinos really approve of Duterte, and only 50% approve of Robredo? Perhaps, perhaps not, but it's unwise to dismiss a statistic just because we don't like it.

Sent from my SM-N970F using Tapatalk

[uls]

because the educated middle class doesn't agree with the survey results they're doing mental gymnastics to explain it

or pagdudahan ang legitimacy ng survey

but the explanation is really simple -- the masses do not think like the educated middle class

-

don't shoot the messenger (Pulse Asia) just coz you don't like the message

[uls]

di nga magets ng mga edukado why the masses love watching other people live their daily lives

-

the masses are large enough in number to make a bunch of jolog vloggers famous and rich

kaya ng masa gawin celebrity ang isang tulad ni Mimiyuuuh

mahirap ba paniwalaan kaya ng masa suportahan ang isang presidente tulad ni Duterte?

[Deestone]

di nga magets ng mga edukado why the masses love watching other people live their daily lives

-

the masses are large enough in number to make a bunch of jolog vloggers famous and rich

kaya ng masa gawin celebrity ang isang tulad ni Mimiyuuuh

mahirap ba paniwalaan kaya ng masa suportahan ang isang presidente tulad ni Duterte?

Lol at mimiyuuh is he or she that famous now? Just saw him 1 or 2 times in lazada advertisement.

Sent from my SM-N910C using Tapatalk

[dr. d]

di nga magets ng mga edukado why the masses love watching other people live their daily lives

voyeurism is alive and well.

[uls]

Lol at mimiyuuh is he or she that famous now? Just saw him 1 or 2 times in lazada advertisement.

Sent from my SM-N910C using Tapatalk

oh yeah he/she is famous

[jut703]

Lol at mimiyuuh is he or she that famous now? Just saw him 1 or 2 times in lazada advertisement.

Sent from my SM-N910C using Tapatalk

Famous enough to be tapped as an influencer for multinational FMCGs. [emoji6]

Sent from my SM-N970F using Tapatalk

[Monseratto]

[akosijhay]

^ Kung pang 20 na tayo sa pinaka maraming case ng may covid, sigurado apektado ang tourism naten. At dahil tourism ang isa sa mga bumubuhay ng economy naten, paano mababayaran ang 9.6 Trillion Pesos na utang?

Kawawa tayong middle class hehe.