Among the many spirited exchanges between Donald Trump and Hillary Clinton over the course of their three debates in 2016, the following from their first encounter was particularly notable:

Clinton: "Crime has continued to drop, including murders."

Trump: "You're wrong. You're wrong. Murders are up."

We're talking concrete figures here. So someone had to be mistaken ... right?

[ad fullwidth]

In actuality, they were both kind of right (or wrong, depending on how you look at it).

To Clinton's point, the violent crime rate, including the murder rate, was very recently at its lowest point in more than 40 years (see the chart below). So in that sense, crime has indeed dropped precipitously and, for the most part, steadily. But, as Trump insistently reminded voters throughout the campaign, there has in fact been an uptick in violent crime since 2014. Between 2014 and 2015, the FBI reported a 3.1 percent increase in violent crime and a 10 percent increase in murders. And the agency's preliminary 2016 data show further increases, an important data point that Clinton omitted in her reference.

For his part, though, Trump neglected to mention that in 2014 the rate of violent crime (murder, rape, aggravated assault and robbery) had dropped to less than half of what it was in 1991. So, while any rise in violent crime, particularly homicides, is certainly cause for concern and deserving of attention, it's worth noting that the uptick in question is being measured against the lowest crime rate in modern history.

All of which underscores why talking about crime rates can turn into such a tricky, emotionally fueled and often misleading exercise. As this Above the Noise video explains, it all comes down to which dataset you look at, what section of it you focus on and what you compare your results to. In other words, if you cherry-pick the data, you can generally hook whatever conclusion you're fishing for.

[Use the scroll bar below to zoom in and view specific data ranges]

And then there's the issue of emotion-driven perceptions and the influence of pervasive crime reporting in the media.

As the Pew Research Center notes, Americans typically believe crime is on the rise, even when that notion is directly contradicted by data. In each of the 21 Gallup surveys conducted since 1989, a majority of respondents said that crime in the U.S. had risen since the previous year, despite the generally downward trend in both violent and property crime rates during most of that time period. And a 2016 Pew Research survey found that 57 percent of registered voters thought crime had gotten worse since 2008, even though both violent and property crime rates declined by double-digit percentages during that period.

Ask a roomful of criminologists to explain what causes dramatic changes in crime rates and you’ll likely get a generous sampling of conflicting answers. From demographic trends and poverty levels to policing strategies and, as Trump proclaims, a general breakdown in law and order, theories abound as to what factors lead to spikes and drops in regional and national rates. It's easy to find correlations. But hard evidence of causation is more elusive.

This is a conundrum of particular concern for places like Chicago, where murder rates have soared in recent years, not to mention other urban areas like St. Louis and Detroit, where rates of murder and other violent crimes have remained stubbornly high for years.

The FBI's annual crime report is the most cited and comprehensive national crime report card, but it's certainly not without holes. It relies on voluntary reporting from more than 18,000 law enforcement jurisdictions around the country, including city, county, university, state, tribal and federal agencies. For the 2015 report, the most recent full-year report to date (data for all of 2016 will be released this fall), 16,643 local agencies -- about 90 percent of the the total enrolled in the program -- submitted data.

Another national crime measure is the annual Bureau of Justice Statistics survey of more than 90,000 households, which asks Americans ages 12 and older if they've been victims of crime in the last six months and whether they've reported those crimes to the police. The 2015 survey found that only about half of respondents who admitted to being victims had actually notified the authorities. Reasons for not reporting included the belief that the police “would not or could not do anything to help” or that the crime was “a personal issue or too trivial to report,” according to BJS.

In other words, the data that are publicly available do not provide a complete picture of crime in America. But they do offer a reasonable approximation. And by nearly every estimate, the overall violent crime rate is higher than it was a few years ago, but substantially lower than it was throughout the 1970s, '80s and early '90s.

California Voter Guide

On Our Watch: Season 2

Beyond the Menu

Videos from KQED Live

Donor-Advised Funds

crimecrime

VIDEO: What's the Deal with the Crime Rate (and Why Can't Anyone Agree on It)?