Polling data-based narratives are foolish

"The viewing attractiveness of election day drama incentivizes networks to try their best to stir up as much drama as possible prior to the election. “Stirring up” can be done easily using narratives based on polling data."


Posted November 2020

By Taj O'Malley

Staff Editor

President Donald Trump’s competitiveness in Michigan, Wisconsin, and Pennsylvania, as well as his strong wins in Ohio and Florida, prove for the second presidential election in a row that polling data-based narratives are nothing but tabloid-esk speculation.

Polls themselves are not entirely useless. They can provide a good understanding of how certain demographics in specific states and counties may vote in an upcoming election. My problem is held against how the polls are reported, and the narratives that are created as a result of that reporting.

Election day (which is now becoming election week) is a dramatic affair. Americans tune into major news networks and watch intently as each state’s votes come in. The unfolding drama rakes viewership up for the networks, which means that advertising companies pay up for screen time during commercial breaks.

“The companies that buy commercial time on the major broadcast networks and cable news channels are anticipating huge interest in election coverage on Tuesday—and a vote-counting process that spills even into the next night,” states The New York Times. “Advertising space between 6 p.m. Tuesday and 2 a.m. Wednesday is sold out at Fox News.”

The viewing attractiveness of election day drama incentivizes networks to try their best to stir up as much drama as possible prior to the election. “Stirring up” can be done easily using narratives based on polling data.

This is why polling data reporting is broad; it emulates election day results and creates the illusion of knowledge about how the election will pan out for the viewer.

Headlines like “Blue wave incoming” and “Biden with lead heading into the election” are extremely misleading. Those headlines make impressions that stick to readers more than the specific data that creates them.

Polling data should be reported, but in a way that lets the audience know that they should be taken with a grain of salt. Ways to do this include having the number of people polled listed with the data, as well as revealing where exactly the polling data is coming from.

I know that elections are exciting and that we all want to know who will win them, but the safest way to know who is going to win is just waiting open-mindedly for the results to come in on election day—or the days that follow.