As the ink dries on the recent political agreement, surveys are revealing a nail-biting competition between Donald Trump and Kamala Harris, both at the national level and within the critical seven swing states. Of late, it has been challenging to find surveys that suggest other outcomes, leading some to suspect that pollsters are converging their results to align with their peers, a phenomenon known as ‘herding.’
The term ‘herding’ denotes the tendency of polls to yield outcomes that line up closely with other polls, particularly as a campaign comes to its final stages. A scenario where less methodologically sound pollsters display ostensibly positive results by altering their polls to echo those from more robust polling firms is common. This trend, however, usually does not fare well when they are left to function without guidance from the higher echelon polls.
In essence, the fear of being labeled the author of a pre-election poll that stands out as an egregious misfire is unappealing to anyone. This might explain why some of the most drastic variations sometimes come from ‘stronger polling firms’ who aren’t troubled about potential dents in their reputation for accuracy.
The procedure of how this ‘herding’ phenomenon takes place raises several questions. Suppose we disregard the possibility that pollsters are just visually scanning the field to mirror the outcomes of others? After acquiring the poll data, pollsters have to determine whether there’s a need to modify or ‘weight’ the collected data. This is because the demographics of those who participated in the poll may not match with those who didn’t participate.
This process typically answers four foundational queries, namely: Does the respondent’s demographic profile in terms of gender, age, education, and race reflect that of the electorate? Post demographic adjustments, does the political leaning of respondents align with that of the general electorate? Which of the respondents will actually cast their votes? And lastly, should the pollster take the acquired data at face value?
The way that pollsters address these interrogatives can result in as much as an eight-point differential in who’s leading in the race. The relatively minor variations visible in the 2024 general election polls could be the strongest evidence to suggest that these queries are being addressed collectively – albeit without direct influence favoring Trump or Harris.
Moving the spotlight away from polls, it’s understandably making pundits and supporters of both presidential nominees restless as they seek clues indicating the possible winner in the impending tight election race. From each camp, there are some who are too focused on the deceptive allure of early voting data.
Because of the numerous unknown elements associated with identifying these early voters over time and the uncertainty around whether their quick votes would’ve been cast later, one can conveniently twist the early voting data to validate any argument they wish to put forth. There’s also an obsession with the subjective measures of ‘passion’ which could be an important consideration but only if it is wide-reaching beyond the guaranteed-to-vote persons and can rouse others (a ‘lethargic’ vote holds the same weight as an ‘exuberant’ one).
Another key area that comes into play is the breadth and effectiveness of the last-minute advertising and initiatives to get out the vote. Advertising efforts most often cancel each other out, and the efforts to encourage voter participation are usually too obscure to hold any meaningful bearing.
Lastly, there are some who analyze recent trends in relation to the country’s objective state, particularly spikes in macroeconomic data. However, this approach is saddled with two hurdles. Firstly, perceptions about the economy are usually established long before the actual Election Day. Secondly, voters’ contemporary perceptions regarding various topics bear little relation to factual evidence.
In spite of all evidence, substantial portions of the electorate hold beliefs that the economy is plummeting, that a nation-wide crime wave is underway, and that heartland communities are reeling from the influx of illegal immigrants engaged in criminal activity and unlawful voting. In such an environment, the voter’s tendency to carefully scrutinize statistics to assess America’s wellbeing is remarkably low.
If the polls eventually turn out to be flawed, it is almost assured that we will witness a surge of post-election ignorant rhetoric. In this case, agitated individuals will float arguments that we should discard all objective indicators of how an election unfolds and instead trust our instincts, ‘gut reactions,’ and personal biases.
As flawed as they may be, polls along with economic indicators and data pertaining to crime or immigration serve us far better than resorting to the cynical partisan chatter, false news, and disinformation, which are not only misleading but also perpetuate as more people fall for them.
As we have seen in the past, it’s a small step from dismissing polls to disregarding actual election results. Hence, even with their imperfections, reliance on neutral indicators such as polls is far better than a reliance on prejudiced intuition and gut feelings, which are misleading and self-sustaining.
In an era dominated by information overload, fact-checking and relying on objective evidence is of paramount importance. With early voting, poll herding, and the question of statistical accuracy challenging our conventional understanding of election trends, it is important to maintain a critical stance.
While it may be true that these objective methods of collecting election predictions are far from perfect, they are still key tools which, complemented by contextual knowledge and analytical thinking, can help us understand an election’s trajectory and potentially predict its outcome.