Are attack ads effective? Here's when they don't work
04 May 2020
By Ram Gopal, Niam Yaraghi, Darrell West and Ram Ramesh
A little girl innocently counts petals on a flower. When she reaches nine her face freezes and a countdown interrupts as the camera slowly zooms into her eye. When it reaches zero there is a nuclear explosion and the archetypal mushroom cloud erupts.
This might not have been the first attack advert but the ‘Daisy ad', as it became known, played a major role in the Democrats successfully characterising Republican rival Barry Goldwater as a dangerous man who could spark a nuclear war and so helped Lyndon B Jonson win the 1964 US Presidential race.
Attack ads have grown in prominence ever since and have become a staple diet of elections, with social media now allowing political parties to do A/B testing to see which negative message resonates with voters.
It is now even commonplace outside of elections. According to Wesleyan University’s Wesleyan Media Project US President Donald Trump spent $5 million in October 2019 in an attack ad blitz across digital and TV on Joe Biden’s alleged corruption and on the Democrats for launching an impeachment investigation against him.
There were a record number of attack ads in the US 2018 mid-term elections, with Wesleyan Media Project counting 569,000 negative adverts airing on TV, 69 per cent of all the political ads seen.
Such paid political adverts are not allowed on TV in the UK, but on social media it is a different story, with the Conservative party reportedly spending £1 million on Facebook with attack ads aimed at Labour leader Jeremy Corbyn in the 2019 General Election.
Political parties’ focus groups must be telling them they work, but we have found that in certain circumstances attack ads can have the complete opposite effect and instead fuel support for the party or policy under fire
When the attack ad does not have a lot of information then people are curious to find out more. And as long as there is plenty of facts and data online to help them, they move in the opposite direction, becoming more knowledgeable and supporting the policy under attack instead.
That is what happened when we studied the attack ads released by the Republican Party’s candidates in the 2014 mid-term elections, when the number one target for their adverts was ‘Obamacare’ or the Affordable Care Act (ACA).
Using a statistical model we were able to use data from the Wesleyan Media Project on the number of TV attack ads on Obamacare in each state and compare that with the enrolment rate to the ACA. When we controlled for other variables such as demographics, the political make-up of the state and the amount of people with health insurance, we found the attack ads actually increased enrolment.
To find out why this was, we undertook a randomised controlled trial involving 300 people. Dividing them into three groups, one was shown positive TV adverts on ACA, the second negative ads and a control group received no ads.
We then presented the groups with four topics and asked them to pick one to test their curiosity for more information on ACA. Finally we gave them a quiz on ACA to measure their knowledge on the subject, having giving them a test before the trial to assess their baseline understanding on Obamacare.
Why attack ads can be counterproductive
The results showed that those who saw the attack ads were not only more curious about ACA and more knowledgeable than the control group, they also outperformed the people who watched the ads supporting ACA.
Watching the attack ads made people 0.53 units more curious than the control group, while those who were shown the positive coverage were 0.32 units higher. And in terms of knowledge, the attack ads group were 1.55 higher than the control group with the positive group just 0.76 units higher.
And when it came to supporting ACA, the attack ad group was 0.49 units higher than those who watched no ads, but for those who were shown the positive ads about Obamacare there was no statistically significant difference between them and the control group.
We tested these results against another subject, a less controversial one. The Common Core education standards policy details what standards students throughout the US should reach in English and maths at each school grade. But we found a very similar pattern consistent with attack ads arousing more curiosity from viewers which leads them to gather more knowledge.
This pattern is explained by George Lowenstein’s Information Gap Theory, which postulates that curiosity is born when there is a gap between “what we know and what we want to know”, so we seek out more information to fill that gap. Thus, for organisations looking to combat attack ads be sure to put relevant facts and verifiable information within easy reach of the public, ie direct online traffic to helpful websites.
Attack ads work when it is very difficult for people to find out any more information on the subject or to determine any real facts. But if there is a lot of data on the subject online, attack ads will only serve to push people to these helpful websites as their curiosity has been piqued.
Repeatedly attacking a concept and portraying it as a negative phenomenon without providing supporting information triggers individuals to feel a knowledge-gap between what they know and what they want to know.
And our trials show that once a person’s sense of curiosity has been triggered, this in turn will lead to higher knowledge and subsequent adoption of the service or product actually under attack.
The anti-ACA advertisements neither mention the details of the ACA nor provide any individually relatable reasons for their attack. Rather it seems that their producers assumed the public already agreed that the ACA was inherently bad and so there was no need to present them with more arguments against it.
Republicans fired out these ads attacking the ACA in an attempt to win votes, but instead an unintended side-effect saw voters seek more information about it. And with a wealth of easy-to-understand information on the web, once this was digested enrolment numbers were in fact boosted - not something Republican candidates would have envisaged or been happy about.
Attack ads will no doubt feature heavily in the US Presidential elections taking place in November 2020, but those on the attack need to be careful they are not actually pushing voters in the opposite direction.
Yaraghi, N., West, D., Gopal, R. D. and Ramesh, R. (2020) "(How) did attack advertisements increase affordable care act enrollments?", PLoS One, 15, 2, e0228185.
Gopal, R., Hidaji, H., Kutlu, S., Patterson, R. A., Rolland, E. and Zhdanov, D. (2020) "Real or not? Identifying untrustworthy news websites using third-party partnerships", ACM Transactions on Management Information Systems.
Loewenstein, G. (1994). "The psychology of curiosity: A review and reinterpretation." Psychological Bulletin, 116(1), 75-98.
Ram Gopal is Professor of Information Systems Management and teaches Digital Finance, Blockchain and Cryptocurrencies on MSc Management of Information Systems & Digital Innovation. He also lectures on Digital Leadership on the Executive MBA and Executive MBA (London).
Niam Yaraghi is Assistant Professor of Operations and Information Management at the University of Connecticut.
Darrell West is Vice President and Director of Governance Studies at the Brookings Institution.
Ram Ramesh is Professor of Management Science and Systems at the University of Buffalo.
For more articles like this download Core magazine here.