Research study
It works on everyone else: 89% of US adults say political ads never move their own vote, 74% say they significantly sway other voters, and 73% think less of a candidate who runs a negative ad
89% of US adults say political ads do not influence their own vote at all while 74% say they significantly influence other voters, and 73% think less of the candidate who runs a negative ad (n=200, October 2026).
- Who was asked
- US adults, 200 clones, US
- When
- October 2026 (fielded 20 October 2026)
- How
- Predicted responses from Brox clones, each built from a long-form interview with a real person.
- What changed from the obvious question
- rather than asking whether attack ads work, the 39-question study plan was split into two studies separating how each ad type moves the view of the sponsor and of the target, what makes an ad credible, and how acceptable each tactic is; one question that referred to the previous question was reworded to stand alone, and the follow-up on what crossed the line was asked of everyone and is read among the 142 who had seen such an ad
Section 01: What the study found
- #
89%
say political ads do not influence their own vote at all, yet 74% say ads significantly influence other voters and nobody says they have no effect on others
base 200
- #
73%
think less of the candidate who runs a negative ad, while 99% say their view of the candidate attacked does not change
base 200
- #
95%
say cited sources or fact-checks on screen make a political ad credible, and 174 of the 200 pick specific statistics or policy details; humor comes last at 1%
base 200
- #
76%
call criticizing a voting record very acceptable, while every respondent rules out attacking family, commenting on appearance, out-of-context clips, AI-altered media and fear-based claims
base 200 · contrast
- #
88%
want stricter legal requirements on the truthfulness of political ads and 80% do not trust them at all; on AI-generated content 54% would ban it and 47% allow it only if clearly labeled
base 200
- #
81%
of Democrats want social media platforms to moderate or fact-check political ads, against 29% of Republicans (n=52), 60% of whom say platforms should not intervene
base 52 · Democrats, Part 2 · contrast
say political ads do not influence their own vote at all, yet 74% say ads significantly influence other voters and nobody says they have no effect on others
02Why does everyone think ads work on someone else?
Because people describe their own habits and imagine everyone else's. 89% (n=200) say political ads do not influence their own vote at all, yet 74% say ads significantly influence other voters. The gap sits inside the same people: of the 177 who say ads never move them, 73% think other voters are significantly swayed. They say they delete, mute and bin the ads (73% skip them consistently) and picture other people buying in.
It holds in every group. Under-45s (n=93) are no more willing to admit influence than those 65 and over (n=43): 89% and 88% say none. Self-reports like this understate real influence, so read the result as how people see themselves, not as proof that ads do nothing.
03What does a negative ad actually do?
It costs the attacker and rarely touches the target. After a negative ad, 99% (n=200) say their view of the candidate attacked does not change, while 73% think less of the candidate who ran it and 69% call that candidate significantly less credible. Contrast ads do a little better, mostly by hurting: 36% see the candidate shown unfavorably more negatively, while 19% warm to the one shown favorably.
Policy-based attacks soften the penalty in principle: 83% say they reflect better on the sponsor than personal ones. In practice the cost stays, because 79% of those who say so still think less of a candidate who runs a negative ad.
04Where do Americans draw the line?
At the record versus the person, and at the truth. 76% (n=200) call criticizing a voting record very acceptable and 67% say the same of using an opponent's own past statements. Every respondent rules out attacking family, commenting on appearance, out-of-context clips, AI-altered media and fear-based claims, and every one of the 161 who rated it rules out repeating a fact-checked falsehood. Personal finances are the grey zone: 78% find them at least moderately acceptable, but 4% very.
71% say they have seen an ad that crossed a line, and among those 142 the commonest reason is false information (46%), well ahead of personal attacks (18%). 61% call negative advertising mostly degrading to the process and nobody calls it a legitimate and useful part of campaigning.
05What do people want done about it?
Rules, almost unanimously. 80% (n=200) do not trust political ads to be accurate at all and 88% want stricter legal requirements on truthfulness, 69% significantly stricter. On AI-generated content nobody wants it left unrestricted: 54% would ban it outright and 47% would allow it only if clearly labeled. Age drives that answer: 72% of those 65 and over (n=43) would ban it, against 40% of under-45s (n=93).
Platforms are where the agreement ends. 58% want social media to moderate or fact-check political ads, while 30% say platforms should not intervene, citing bias and who decides what is true.
06Does party change the verdict?
It changes the emphasis, not the verdict. Democrats and Republicans agree on the problem: in Part 2, 73% of Democrats (n=52) and 83% of Republicans (n=52) do not trust ads at all, and 100% and 85% want stricter truthfulness laws. They split on who should enforce it: 81% of Democrats want platforms to step in against 29% of Republicans, 60% of whom say platforms should stay out.
The messenger splits them too. In Part 1, 63% of Republicans (n=49) say the candidate speaking straight to camera makes an ad credible, against 34% of Democrats (n=59), while 36% of Democrats trust endorsements from respected figures against 16% of Republicans. Democrats react more strongly to attacks: 83% judge the sponsor of a negative ad significantly less credible, against 58% of independents (n=91).
Section 07: What they said, and the interview line it rests on
“I would say significantly. For me personally, I try not to pay attention to political ads, mail, or news because it's not going to change the way I feel, and they're mostly full of white lies. But when it comes to other people, they get so engaged in all the politics that they just buy right into whatever they are told.”
Predicted from this line in the interview
“I try not to watch anything politically, ads, paper, news, anything like that. I don't really care. That's it. It's not going to change the way one person feels.”
“Honestly, I view them as significantly less credible because it just feels mean-spirited. I've always believed in treating people fairly and showing kindness, empathy, and respect. When someone resorts to attacking and tearing down an opponent rather than standing on their own merits and focusing on solutions for the people, it goes against my moral compass.”
Predicted from this line in the interview
“I wouldn't try to hurt someone's feelings to make them feel lesser or feel like they're definitely in the wrong.”
“For me, what makes a political ad feel credible is seeing cited sources or fact-checks shown on screen, specific statistics or policy details, and having the candidate speaking directly to the camera. I prefer getting information first-hand directly from the person rather than second-hand gossip, and I need to see real, verifiable facts and sources instead of claims that are just made up.”
Predicted from this line in the interview
“If I want to know something, I go first hand and get information first hand, not second hand. Because that's how you get lies and rumors.”
“For me, an ad crosses the line when it's just outright false information or completely misleading. I place a huge value on honesty and truthfulness, so when a political ad deliberately misrepresents facts, distorts what actually happened, or makes up claims about an opponent that aren't true, that's where I draw the line.”
Predicted from this line in the interview
“I'd want to display my trustworthiness, some commitment to honesty”
“To me, a political ad that is tough but fair is one that sticks strictly to the facts, like a candidate's actual voting record, policies, or decisions they've made in office. If you can point to verifiable evidence and compare records, that's completely fair game. Where an ad goes too far is when it leaves the facts behind and relies on distortions, vague exaggerations, or just tries to manipulate people through pure emotion.”
Predicted from this line in the interview
“For me, it's the facts, not the feeling.”
“Should be completely banned. Honestly, they shouldn't even be allowing that. AI is already getting out of hand and doing too much, and having fake videos, fake voices, or fake pictures of people up in political ads is just lying to folks. People already can't believe what they see.”
Predicted from this line in the interview
“With the introduction of AI, it's gonna get worse. Like, AI is taking all of our jobs. Every single one of them.”
“No, platforms shouldn't intervene. Who is to say who is the authority on what's misinformation and what's not? People should be critical thinkers and make up their own minds rather than having social media companies act as arbiters of truth.”
Predicted from this line in the interview
“Because who's to say who is the authority on what's misinformation and what's not?”
Section 08: How much should you trust this?
What we would defend
- the gap between influence on oneself and on others, in every party and age group
- the penalty a negative ad carries for its sponsor rather than its target
- evidence on screen as the route to credibility, ahead of emotion and humor
- the line between attacking the record and attacking the person, or the truth
- near-total distrust with broad, bipartisan support for stricter truthfulness rules, and the party split on platform moderation
What we would not
- exact levels on norm-like questions, where clone answers bunch: several acceptability items came back unanimous, and 'I judge every ad by the same standard' (96%) and 'no change in my view of the target' (99%) are likely more principled than real behavior
- self-reported influence as a measure of real influence: it understates it
- the 65 and over columns as more than directional: 43 people in each part
- any person-by-person link between the parts: they are separate draws of 200 (69 clones answered both) and no answer is joined across them
The main limitation
everything here is what people say about ads, not what ads do to them: self-reports understate influence and clone answers bunch on questions with a clear social norm
Section 09: How was this run?
| Segment | Clones |
|---|---|
| Part 1: US adults | 200 |
| Part 1: Democrats | 59 |
| Part 1: Republicans | 49 |
| Part 1: independent, other party or none | 91 |
| Part 2: US adults | 200 |
| Part 2: Democrats | 52 |
| Part 2: Republicans | 52 |
| Part 2: independent, other party or none | 95 |
Question wording as fielded
- Part 1. How often do you encounter political advertising?
- Part 1. Where do you most often encounter political ads? (up to 2)
- Part 1. How do you feel about political advertising as a concept?
- Part 1. How much do political ads influence your own vote?
- Part 1. How much do political ads influence other voters' decisions?
- Part 1. Which style of political ad do you find more persuasive?
- Part 1. Do you actively avoid or skip political advertising?
- Part 1. Which elements make a political ad feel more credible? (up to 3)
- Part 1. What tone do you respond to best?
- Part 1. Should political ads focus on policy and issues, or on character?
- Part 1. After a negative ad, how do you see the candidate who ran it?
- Part 1. After a negative ad, how do you see the candidate it attacked?
- Part 1. After a contrast ad, how do you see the candidate shown favorably?
- Part 1. After a contrast ad, how do you see the candidate shown unfavorably?
- Part 1. After a positive ad about a candidate's own record and vision, how do you see them?
- Part 1. Do negative ads based on policy reflect better on the sponsor than ones based on personal attacks?
- Part 2. How do you feel immediately after seeing: Positive ads about a candidate's own vision?
- Part 2. How do you feel immediately after seeing: Policy-focused contrast ads?
- Part 2. How do you feel immediately after seeing: Personal attack ads?
- Part 2. How do you feel immediately after seeing: Fear-based 'if my opponent wins' ads?
- Part 2. How do you feel immediately after seeing: Humorous or satirical ads?
- Part 2. After heavy political advertising, how do you feel about the political process?
- Part 2. Does negative advertising make you more or less likely to vote?
- Part 2. How acceptable is this in a political ad: Criticizing their voting record or policy positions?
- Part 2. How acceptable is this in a political ad: Using their own past statements against them?
- Part 2. How acceptable is this in a political ad: Criticizing their past job performance?
- Part 2. How acceptable is this in a political ad: Questioning their judgment on a specific decision?
- Part 2. How acceptable is this in a political ad: Criticizing their personal finances or business dealings?
- Part 2. How acceptable is this in a political ad: Commenting on their age or health?
- Part 2. How acceptable is this in a political ad: Attacking their family members?
- Part 2. How acceptable is this in a political ad: Commenting on their physical appearance?
- Part 2. How acceptable is this in a political ad: Using out-of-context clips or quotes?
- Part 2. How acceptable is this in a political ad: Repeating a claim fact-checked as false?
- Part 2. How acceptable is this in a political ad: Using AI-generated or altered images, audio or video?
- Part 2. How acceptable is this in a political ad: Exaggerated or fear-based claims about an opponent winning?
- Part 2. What makes an ad mudslinging rather than legitimate criticism? (up to 3)
- Part 2. Is negative advertising legitimate, or does it degrade the process?
- Part 2. Have you seen a political ad that crossed a line?
- Part 2. What was it about that ad that crossed the line? (those who have seen one, coded)
- Part 2. Do you judge negative ads from your own candidate by the same standard?
- Part 2. How should AI-generated or digitally altered content in political ads be treated?
- Part 2. How much do you trust political ads to be factually accurate?
- Part 2. Should truthfulness rules for political ads be stricter than they are now?
- Part 2. Should social media platforms moderate or fact-check political ads?
- Part 2. How often do you fact-check claims made in political ads?
- Part 1 and Part 2. Political views and party identification (asked in both parts so each can be split by politics)
- Part 2. In your own words, what separates a tough but fair political ad from one that goes too far? (coded)
What was not asked, and why
age, gender, income, education and region were not asked: they come from each clone's own record; party identification and political views were asked in both parts so each part can be split by its own respondents' politics
What changed from the obvious question
rather than asking whether attack ads work, the 39-question study plan was split into two studies separating how each ad type moves the view of the sponsor and of the target, what makes an ad credible, and how acceptable each tactic is; one question that referred to the previous question was reworded to stand alone, and the follow-up on what crossed the line was asked of everyone and is read among the 142 who had seen such an ad
More on Public opinion · Advertising channels · Message testing · Public affairs & society
Ask this of your own audience: humans@brox.ai
More in Public affairs & society
All in Public affairs & society-
100% of US adult clones call a secret, sexual AI affair cheating, against 1% when a partner discloses flirting with an AI that is not sexual
100% 100% rate a partner's secret, sexual relationship with an AI companion as cheating (6 or 7 of 7), against 1% when the partner told them about flirting that was not sexual (base 200)
Read -
89% strongly praise weight lost through diet and exercise, against 1% when the same weight is lost on medication
89% 89% give strong praise to Diet-and-Exercise Jordan, against 1% for Medication Jordan, who lost the same 60 pounds (separate samples) (base 200)
Read -
82% of US men would not get a hair transplant, and 98% of those who see their hair thinning rule one out
82% 82% of US men are not open to getting a hair transplant, and none say they are very open (base 200)
Read