The University of Massachusetts Amherst

SBS Students Find Increase in Use of Personal Attacks during Presidential Debates

Friday, October 9, 2020

During the most recent Presidential debate, candidate Joe Biden called Donald Trump a “clown” while Trump said there was “nothing smart” about Biden, mocking his academic performance and compliance with CDC recommendations to wear a face mask in public. Was their behavior anomalous, or part of a growing trend in the use of personal attacks in Presidential debates?

Hypothesis: Candidate speech in election debates now features more direct criticism, particularly personal attacks, than in past debates.

A student research team consisting of Spencer Glover, Ryaan Moss, Sarah Perno, Tim Scalona, and Sarah Weintraub studied the language of past election debates to see if personal attacks have escalated over the years. They identified any time when a candidate criticized another person: other candidates participating in the debate, candidates not present (e.g., Trump criticizing Clinton during the Republican primary debate), or other targeted individuals (e.g., other political elites, family members, etc.). The researchers also looked for criticism of broad groups (e.g., the opposing party, “liberals”, “immigrants”) and specific policies or policy proposals, as well as positive calls for bipartisan action. 

The team coded the language used during two primary debates and two general election debates for three presidential Republican nominees: President George W. Bush’s 2000 campaign, Senator John McCain’s 2008 campaign, and President Donald Trump’s 2016 campaign. The group decided to focus on the winning nominee in a single party in order to use their limited research time to focus on the debate messages that were most successful with voters, although they would have liked to be able to look at the language of both Republican and Democratic candidates in all of the debates in recent elections.

The tone of Presidential debates grew significantly more critical over the three campaign cycles, with Trump going on the attack around three times as often (57.5 per general, 25 per primary) as Bush in the 2000 debates (20.5 per general, 9.5 per primary). The upward trend in critical attacks was already in place in 2008 with McCain engaging in more direct criticism (24.5 per general, 16 per primary) than Bush. All three candidates were consistently more likely to be critical in general debates than in primary debates, making about twice as many criticisms of all types when debating their Democratic opponent. 

It may be easy to attribute this trend to Trump’s personal style, as he regularly is critical of individuals on Twitter and in other venues. However, students found his behavior was in keeping with a well-established pattern of Republican primary winners criticizing individuals as much (if not more) than policy positions. Over half of the critical statements made by Bush and McCain during primary and general debates were coded as attacks on individuals. While Trump’s strategy in all debates showed a greater reliance (around 65-75%) on personal attacks, his strategy was not unprecedented; Bush used a similar proportion of individual attacks in the 2000 primary debates. Compared to Trump, Bush was less likely to engage in direct criticism overall and he directed more of his critical comments towards individuals not participating in the debate, but he still was 3 times as likely to criticize individuals as compared to policies.

Trump’s debate participation did depart from established trends in ways that may bode poorly for future campaign rhetoric and political polarization if adopted by future candidates. Trump appears to be an outlier in terms of his use of attacks against minority groups and individuals not participating in the debate. Explicit criticisms of minority groups rarely featured in campaign debate rhetoric prior to Trump: McCain never engaged in minority group attacks, while Bush made only one criticism of a minority group (in a primary debate) in 2000. While still rare, Trump criticized minority groups an average of 2 times per debate in all debates. 

Republican candidates already had some history of criticizing individuals not on the stage, particularly during the primary debates where they presumably spoke negatively about potential Democratic opponents (3.5 times per debate). However, Bush and McCain focused most of the personal attacks on their Democratic opponent during the general debate and only rarely criticized non-participating individuals (2 attacks per debate). Trump greatly upped the frequency of attacks against off-stage individuals (11 per debate) in his debates against Hilary Clinton compared to both his own style in the primary debates and the style of earlier Republican candidates in both primary and general debates.

While attacks during presidential debates have increased, calls for bipartisanship have stayed low in Republican primary debates (around 1 per debate) during this time period and have disappeared in general election debates. Bush made fairly frequent calls (5 per debate) for bipartisanship in 2000, while McCain (1 per debate) and Trump (0 per debate) essentially dropped the issue from their debate rhetoric. 

The team wished they had more time to go further with this project, as they were not able to find out whether Democratic candidates or other Republican candidates who did not secure the official nomination adopted similar debate strategies. They would have also liked to analyze a great number of debates, to ascertain whether or not debate format or moderation affected the overall critical tone of a debate. 

The research team concluded that non-incumbent Republican nominees have been successful with voters even though they have long been willing to openly criticize individuals and not just policies. Moreover, the relative frequency with which candidates engaged in criticism - including individual criticism - was already on the rise prior to Trump’s 2016 campaign. At the same time,   calls for bipartisanship action during debates were rare or non-existent, reinforcing the point that the general public now expects successful Presidential candidates to act in ways consistent with a highly polarized political environment. 


This study was completed by students at the University of Massachusetts Amherst. The UMass Amherst data analytics and computational social science (DACSS) program offers training for graduate and undergraduate students to learn to identify sources of evidence to use in decision-making, design original research, work with large datasets, develop reliable and ethical data management practices, interpret and visualize results of the analysis, and communicate about data and research to decision-makers and general audiences.

All students have the opportunity to conduct independent research. This is a modified summary of a student project.