Could political rhetoric be hazardous to public health? Quite possibly, according to a new University of Pennsylvania analysis of the last presidential campaign.

During the 2016 campaign, cities that hosted rallies experienced 2.3 more assaults than average on the days that Donald Trump held campaign events there, according to the study, published Friday in the journal Epidemiology.

The researchers noted that news reports cited violent incidents at some Trump campaign rallies, along with language that could be described as colorful. Some of the rally comments attributed to Trump in the study include anti-protester utterances like "I'll beat the crap out of you," "I'd like to punch him in the face," and "Maybe he should have been roughed up."

By comparison, Democratic candidate Hillary Clinton's rallies were not found to be linked to any increase in assaults. In the months preceding the Nov. 8 presidential election, media accounts of Clinton's campaign on numerous occasions included the word boring, an adjective that  was not known to be applied to her opponent.

While the increased number of altercations in the cities on the days of Trump rallies was not proved to be a direct result of his comments, the researchers believe that the difference in language between the candidates may have had an effect on public behavior.

"The language of our leaders matters," said lead study author Christopher Morrison, a postdoctoral fellow in Penn's Center for Clinical Epidemiology and Biostatistics. "We have an association between a candidate who was openly promoting violence at events and increases of violence on days of the events. It's a very clear indication that the language our leaders use can affect the mood of the nation and affect people's behavior."

Epidemiologists explore the causes of health problems, including threats to public safety, injury, and disease. The Penn researchers said they believe their study – an effort to understand some of the causes and motivations of violent behavior — to be the first of its kind: "We know of no other empirical studies that investigated violence at a population level associated with previous U.S. presidential rallies," the article states. The study did not require outside funding.

The researchers focused on cities with more than 200,000 people where criminal incident data were available online. Through online searches, they identified data for 31 Trump rallies in 22 cities and 38 Clinton rallies in 21 cities. They counted the incidents for those days, as well as other days for the cities. Since weather can be a factor in crime rates, the researchers considered temperature and precipitation in preparing the study.

In the case of Trump rallies, the additional assaults may have occurred at or around campaign events or elsewhere in the cities, but the authors said they were likely the result of "emotional states" transmitted through news reports or social media.

No additional assaults were associated with the days and locations of Clinton campaign events. That included some rallies in Philadelphia.

The effect of political speech on a populace is worthy of additional study, the authors said.

"Given the ongoing reports of violence at political gatherings in the U.S. (e.g. March 4, 2017, in Berkeley, Calif., where pro- and anti-Trump protesters clashed)," they wrote, "whether and how political rhetoric that normalizes or promotes violence affects violence at the population level appears to be an important area for further research."