Featured Post

Love’s Labor Found

Love’s Labor Found August 4, 2019 Eighteenth Sunday in Ordinary Time Luke 12: 13-21 Someone in the crowd said to Je...

Sunday, August 5, 2018

A tale of two surveys

BY ON
SOME people do not believe surveys. Others believe only those whose results they agree with.
Those who doubt surveys argue that it is impossible that the opinion of several hundred Filipinos could truly represent the opinion of the entire country. These people need to be told about the science of statistics, that there is already a well-established body of knowledge that governs drawing generalizations from samples. When a survey is conducted scientifically, which means that proper rules of randomization were followed in drawing up the sample and that the survey instruments were fair and administered in an unbiased manner, it is already scientifically established that the opinion of 1,200 people can represent an entire population of millions.
Surveys have been applied over the years, and the techniques for conducting them have become more and more sophisticated. Surveys are not confined to opinion polling, but even in quality control, where the consumer goods that people buy, especially those that are manufactured in large quantities are subjected to quality checks by measuring and evaluating not an entire batch rolling out of conveyor belts in factories, but only a random sample from such batch.
However, while the people who do not believe any survey can be convinced to change their minds by educating them about the science behind it, those people who believe only the surveys that are consistent with their partisan truths are harder to convince, if not totally hopeless to enlighten. These are people who would ignore, and worse bash, surveys that produce results contrary to their political positions, and would celebrate those that yield numbers that exalt their preferred narratives.
Social Weather Stations (SWS) recently came out with the results of its survey where President Duterte’s satisfaction rating went down by 5 percentage points, from 70 percent in March to 65 percent in June. His dissatisfaction rating, on the other hand, went up by 6 percentage points from 14 percent in March to 20 percent in June. This resulted in a double-digit decline in his net-satisfaction ratings from 56 percent in March down to 45 percent in June. By SWS classification, the drop is by no means interpreted as “bad,” but rather a shift of the President’s rating from “very good” in March to only “good” in June.
A week later, Pulse Asia came out with a much different result. The President’s trust rating went up by 9 points, from 78 percent in March to 87 percent in June, while his performance approval rating went up by 8 points, from 80 percent in March to 88 percent in June. What is even more interesting is that the Pulse Asia survey also revealed an increase in the numbers of Leni Robredo. Her trust ratings went up by 3 points, from 53 percent in March to 56 percent in June, while her performance approval ratings went up by 7 points, from 55 percent in March to 62 percent in June. In fact, the survey revealed an increase in trust and performance approval ratings for then Senate President Koko Pimentel and House Speaker Pantaleon Alvarez. The increase in Pimentel’s ratings was even higher than the President’s, at 11 percent.
As expected, the reactions from both sides of the political divide were different. There was a palpable silence from the ranks of the Duterte supporters on the SWS results, even as there was an eruption of jubilation when the Pulse Asia results were released. Notable also was the muted reaction on the increasing numbers for Robredo; what was instead played up was the “incompetent” label the President gave her, and the mathematical blunder she committed in public. In contrast, while the political opposition talked about the SWS results as an affirmation of the declining satisfaction with the President, they lambasted the Pulse Asia result, with Sen. Antonio Trillanes 4th even accusing Pulse Asia of bias.
The significantly different results of the two surveys have also given ammunition to those who continue to doubt surveys, with some people using the contrasting results as evidence of the unreliability of any survey, and that because different peoples are sampled, then different results are to be expected.
Actually, the difference in trend is not due to the fact that the samples of SWS and Pulse Asia are different sets of people. If they are truly randomly chosen, then they both will represent the trend for the entire population.
One needs to be told that the different results can be properly explained by taking into consideration the element of timing. Opinion surveys are mere snapshots of the state of mind of people at a given time, influenced by events happening before them. Hence, the difference in the results of SWS and Pulse Asia surveys can be explained by the fact that they were taken in two different time periods. Pulse Asia conducted its survey from June 15 to 21, while SWS conducted it from June 27 to 30. Had there been no major event that occurred from June 16 to June 26, a big difference in the trend would have been suspect. However, it was on June 22, a day after Pulse Asia completed its data gathering, that the President uttered his “stupid God” remark, which created much negative publicity in the days that followed. It is highly likely that this caused the President’s satisfaction rating in the SWS survey to shed off 5 points from 70 percent to 65 percent, and his dissatisfaction rating to increase by 6 points from 14 percent to 20 percent.
It is clear that the President’s attacks on religion have had negative impacts on his numbers. He can either ignore it or take it to heart. A decline in satisfaction ratings may personally not bother the President, but in the realm of political reality, lowered satisfaction can spell trouble, more so that a mid-term election is just less than a year from now.
His supporters can conveniently ignore the SWS survey results, and instead celebrate that of Pulse Asia’s as their preferred narrative. However, interpreting survey results, when done by partisans, reflects nothing but the convenient truths seen from biased lenses, which may not be necessarily helpful in recalibrating policy or the conduct of public officials.
An increase in the trust and performance ratings of the President in a survey where the numbers of others, including that of Robredo’s, increased and where the increase in Pimentel’s numbers is even higher than the President’s, cannot be a source of comfort when a later survey shows a decline in net satisfaction. Celebrating Pulse Asia while ignoring SWS may give the illusion that the President is immune from political backlash, when in truth he is not.

http://www.manilatimes.net/a-tale-of-two-surveys/420296/

No comments: