Surveys Are Snapshots, Not Crystal Balls

By Francis Allan L. Angelo

The assertion that Bam Aquino and Kiko Pangilinan’s electoral victories are a “blow” to pre-election surveys rests on a common but flawed misunderstanding of what surveys are and what they are designed to do.

Public opinion surveys, especially those conducted by established organizations like Pulse Asia, Social Weather Stations (SWS), and Octa Research, are not predictive tools. They are snapshots of public sentiment at a particular point in time, not guarantees of election results. This is a fundamental tenet of survey science. Polls measure intention—not destiny.

Each pre-election survey reflects the public mood, name recall, and voter preferences during the fieldwork period, often days or weeks before an election. But public opinion is not static. It is dynamic, shaped by a multitude of factors including late-campaign events, media coverage, endorsements, controversies, the mobilization efforts of grassroots organizing, and the evolving decisions of undecided voters. Candidates like Aquino and Pangilinan likely benefited from last-minute swings—what pollsters call “late-breaking” support—that occurred after the final surveys were conducted or were not fully captured.

Moreover, the margin of error, typically ±2% to ±3% for national surveys, must always be considered. A candidate ranked, for instance, at 14th in a pre-election survey could, statistically, land in the 11th or 12th spot by election day if their support grew within this margin. These margins are not indicative of errors in counting; they are inherent to the sampling process, reflecting natural variations that occur when studying a representative sample of a larger population.

The criticism that a sample of 1,200 or 2,400 respondents cannot represent 68 million voters also ignores basic statistical principles. Properly drawn random samples, when weighted and stratified according to regional population, age, class, and other relevant demographics, provide accurate representations of larger populations. This is why opinion polling is used worldwide—not only in elections but also in economics, health, governance, and market research. In fact, the U.S. National Science Foundation, the World Bank, and even major technology companies like Google and Meta invest heavily in survey-based research. That’s because when done scientifically, surveys are powerful tools to understand attitudes, behaviors, and trends.

More critically, dismissing surveys wholesale because they “got it wrong” in predicting exact electoral placements is akin to blaming thermometers for a fever. The function of polls is to inform, not to anoint winners. That some candidates who ranked lower in earlier surveys eventually triumphed does not invalidate the surveys themselves; it highlights that voters can and do change their minds, often influenced by factors right up to election day.

This understanding leads to recognizing the broader importance of surveys beyond just political races:

• Governance and Public Policy: Surveys are indispensable for governments and NGOs to assess community needs, gauge public satisfaction with services (like healthcare and education), and gather input on policy proposals, leading to more responsive and effective governance. Social Weather Stations (SWS) reports in the Philippines, for example, offer crucial data on poverty, hunger, and public opinion on national issues.
• Business and Marketing: Companies rely extensively on surveys for market research—understanding consumer preferences, testing new products, measuring brand loyalty, and assessing customer satisfaction, all of which inform critical business strategies.
• Healthcare: Surveys in the healthcare sector gather patient feedback to improve care quality and track public health trends, disease prevalence, and health behaviors, informing vital public health campaigns and policies.
• Social Sciences and Academia: Researchers across various disciplines use surveys as a primary method to study social attitudes, economic conditions, cultural shifts, and behavioral patterns, contributing to our collective understanding of society.
• Media: Beyond election polling, media organizations use surveys to explore public opinion on diverse current events and social issues, adding depth and context to their reporting.

As for insinuations of “computer manipulation” of results, these are serious allegations that require concrete evidence and thorough investigation by competent authorities like the COMELEC, not speculative chain messages. Former COMELEC Commissioner Gus Lagman, a noted advocate for electoral transparency who has consistently called for audits and open-source reviews, has not verified nor endorsed the supposed message attributed to him in the original piece regarding ballot shading.

Indeed, for those who, like myself, love to look at poll surveys, juxtaposing them with realities and insights gathered on the ground offers a fascinating way of appreciating the intricate dynamics and nuances of power and politics. This practice underscores that surveys are valuable instruments for analysis, offering a particular lens on public sentiment that, when combined with other observations, enriches our understanding of the electoral landscape.

Rather than vilifying survey firms, a healthier democratic approach is to demand better public understanding of surveys—how they are designed, interpreted, and improved. Survey firms must remain transparent about their methodologies, funding sources, and error margins. But the public must also resist drawing absolutist conclusions from what is, by design, probabilistic data.

Surveys aren’t always right—because people aren’t always predictable. But that doesn’t make them useless. On the contrary, they offer a vital window into the political, social, and economic temperature of a nation, especially one as complex as the Philippines.

In the end, democracy is served not by throwing away thermometers, but by learning to read them better.