"Don't Go, Vote" Arrives as an Ad: A "Smartphone Ad" Stood in Front of the Polling Station — Evidence That Social Media Micro-Targeting Lowered Voter Turnout

"Don't Go, Vote" Arrives as an Ad: A "Smartphone Ad" Stood in Front of the Polling Station — Evidence That Social Media Micro-Targeting Lowered Voter Turnout

When you open your smartphone, there's shopping, videos, updates from friends, and advertisements. We think we're choosing "what we want to see" every day. But are we really the ones making the choices? A study has emerged that answers this question from an uncomfortable angle.


To get straight to the point, SNS advertisements do more than just encourage people to "go vote." They can quietly deliver the choice of "not going to vote" to targeted individuals. And this has been linked to an actual decrease in voter turnout.


Tracking whether individuals saw "voter suppression ads"

The crux of this study is not the usual speculation of "such ads were circulating on SNS," but rather the accumulation of data showing "this person actually saw that ad." The research team recruited over 10,000 participants to closely match the demographic composition of the U.S. voting-age population. For six weeks leading up to the November 2016 presidential election, participants installed an app that recorded the ads they viewed.

 
Furthermore, they cross-referenced state voting records to link "ads seen" with "whether they actually voted." This design verified ad exposure and voting behavior using "verified voting records," rather than self-reported surveys.


The effect seems small, but it's significant enough for elections—voter turnout drops by about 1.9%

The results were clear. Participants who encountered voter suppression messages on Facebook were about 1.9% less likely to vote compared to those who did not (with some reports indicating an average of 1.86%).

 
When you hear 1.9%, some might think, "Isn't that just a margin of error?" However, in 2016, the outcomes in some states were extremely close. The researchers also noted, "It may seem small, but the margins were also small in many states in 2016."

 
Extrapolating nationwide, there is even an estimate that voter suppression ads may have kept "about 4.7 million people" away from the polls.
A "few percent change" can be fatal in a democracy. Especially in presidential elections, where narrow margins in swing states can determine the outcome.


"Who received them" was also biased: Heavily targeted at non-white individuals in swing states

Another important point is the bias in ad distribution. Voter suppression ads were designed to reach non-white populations of voting age in swing states more frequently. The study found that non-white users in swing states received voter suppression ads four times more than their white neighbors.

 
Moreover, compared to "white-majority areas in non-swing states," minority community users received voter suppression messages "about ten times" more frequently.

 
The numbers indicate more than just "differentiated ad placement." Messages were being precisely delivered to groups that are politically influential and historically more likely to distrust the system or face voting barriers.


The content was "Boycotting is the strongest protest"—channeling anger into "abstaining from voting"

The typical rhetoric of voter suppression ads is more sophisticated than blatantly shouting "Don't vote." The study found many ads claiming that "boycotting the election is the most effective protest against politicians."

 
The insidiousness of this tactic lies in its ability to use genuine emotions like anger and disappointment with politics as fuel. "Nothing will change anyway," "They're all the same"—there are always some who feel this way. Pretending to empathize with these sentiments, the ads ultimately only stop voting behavior.


Moreover, such ads do not necessarily spread and spark debate like news articles. Ads are primarily individually targeted, so even family members living in the same area may see different content. It's hard for those around to see who chose "not to vote" and for what reasons.


The "opacity of the source" seen in the background: Not entities reporting to the FEC

Another issue highlighted by the study is the transparency of advertisers. It was reported that the voter suppression ads in question were not purchased by entities disclosing political activities or reporting to the FEC.

 
In other words, recipients find it difficult to discern from the ad screen who is delivering the message to them.


Furthermore, in the context of congressional investigations, it was shown that many of the ads identified by the research team were purchased by Russia's "Internet Research Agency (IRA)." Targeting terms like "Martin Luther King, Jr." were used to ensure they reached non-white voters.

 
The study's main focus is not party criticism but the point that the "structure of the information environment" shakes the foundation of democracy. If it's unclear under whose responsibility, with what funds, and by what standards voter suppression ads are distributed, the same thing can happen repeatedly.


"Digital voter suppression"—a replay of history happening behind ad distribution

Historically, voter suppression has been carried out through violence, intimidation, and institutional design. However, in the digital space, there's no need to stop people at the entrance of polling stations. It's enough to provide "reasons not to go" on a smartphone screen.


The authors of the study point out that such messages are designed for communities that already have inherent distrust in the government or elections, positioning it as a "modern version" of historical voting access issues.

 
From "physically blocking access to the ballot box" to "psychologically breaking the will to go to the polling station." Technological advancements have updated even the forms of oppression.


Reactions on SNS (Emerging "divergence of viewpoints")

This study, from the moment it was published, has been discussed on SNS not only as "academic" but also in terms of "system design" and "platform responsibility." The reactions have split into three main directions.

1) "Finally, 'actual measurements' are out"—Evaluation from the research and policy community

Posts focusing on the novelty of the study (tracking ad exposure at the individual level and cross-referencing with voting records) stand out. The authors themselves emphasize the stronger impact on non-white residents in swing states, promoting the study as one that highlights "blind spots" in voting rights, election systems, and political funding.
The fact that it demonstrated using "behavioral data × public records" rather than "speculation" is seen as solidifying the foundation of the discussion.

2) "The transparency of ads is broken"—Calls for rule establishment

On SNS, the point that "the fact that we don't know who issued the ad itself is a problem" is also strong. Reports repeatedly explain the point that entities not reporting to the FEC, etc., were issuing ads, and that in digital spaces, it's possible to precisely target "weak links."
Such perceptions easily connect to policy discussions on "the effectiveness of ad libraries," "the scope of disclosure obligations," and "countermeasures against foreign intervention."

3) "Is 1.9% small? Or is it fatal?"—Debate over the interpretation of impact

On the other hand, on SNS, there are always reactions like "If the effect is small, isn't it too much fuss?" However, the researchers point out that considering the narrow margins in states in 2016, it's not small, and that extrapolating nationwide gives an estimate of about 4.7 million people.
This point is also a communication issue of how society shares the reality that a few percent shift can change the outcome, as elections are "majority rule."


So what should we be wary of?

What makes this study alarming is that voter suppression ads do not rely solely on "absurd lies like conspiracy theories."


Anger, disappointment, political distrust. The optimization of ad distribution touches on emotions that already exist in society. The words of boycott arrive with a face of justice. Moreover, the recipients are precisely chosen, invisible to those around them.


Countermeasures do not end with "let's stop using SNS." As researchers suggest, disclosure of advertisers, enforcement of regulations, and the context for recipients to "see who is targeting what" are necessary.

 
Before increasing security at the entrance of polling stations, we might need to protect the "invisible entrance" within our smartphones.


Source URL

Related Articles