Is the FBI Obscuring Murder Statistics?

Does realistic crime fiction represent or distort the realities of crime? To find out I took an amateur dive into crime statistics to see what kind of murders are being committed in the contemporary US. (Qualification: I’m no statistician and no criminologist, just an inquisitive person using Google to see what I can find out.) Where do you go to look for accurate crime statistics? The FBI’s website, of course. Visitors are directed to the FBI’s Crime Data Explorer (CDE).

The opening page of this ‘interactive’ site tells you that violent crime for 2023 is down 3%. But when you proceed to homicide data, you’re told that not all law enforcement agencies are reporting, so the FBI’s stats don’t include all homicides. How then do they know for sure it’s down 3%? Anyway, it turns out 63% of US law enforcement agencies reported in 2021, meaning 37% didn’t. They say this percentage is improving, but think about this. 2022 statistics include no, repeat no, data from NYPD, LAPD, the police of Phoenix and San Jose and many other municipalities. Only 10% of agencies in Florida and Pennsylvania reported to the FBI that year and many other states faltered. How accurate are national homicide stats that leave out the two largest urban areas in the country, two large states and many cities and towns? 

The CDE breaks out homicides with charts of victim and offender demographics,  weapon used and victim circumstances. Victim circumstances means in what context did the murder occur? A robbery? A dispute? An auto theft? The majority of these situations are classified as ‘other arguments’ or ‘other’. There is no category for domestic dispute or domestic violence. Gang murders are divided into two categories, ‘gangland killings’ and juvenile gang killings’. Supposedly gangland killings refers to organized crime but as cartels favor using minors and interact with gangs, this could be dicey to determine. Around half the reported murders say the ‘facts provided do not permit determination of circumstances’. 

This is unsatisfying so I check other websites. Non-governmental sites estimate there are over 30,000 gangs in the US with a membership approaching one million, average age 17. It’s estimated they account for 13% of all homicides and 25% of all crime.  Gang murders are clearly a huge problem but don’t go to the FBI for information. By contrast it took me one click to pull up statistics on gang killings in Canada by year, province, city. Of course Canada only had 788 homicides in 2022 so maybe it’s easier for them to keep track.

Surely the FBI tells us how many femicides are occurring in this country. But no, their statistics are patchy and unclear. The CDE provides a chart categorizing the relation between victim and offender but the categories are confusing. For example, victims can be a friend, girlfriend, wife, acquaintance or the victim of someone they know. Who decides whether the dead woman lying on the carpet is the murderer’s girlfriend, friend, acquaintance or somebody they know?  

Add to that a weird ‘boyfriend’ category, which had a quarter the number of murder cases of wives and girlfriends together. What does this refer to? Are they talking about gay males killed by a relationship partner? Or male romantic partners killed by their female partner? If the latter, could it be true, as these statistics suggest, that one of every four murders occurring in the context of a heterosexual relationship has a male victim? Seems unlikely but that’s the way it’s written. 

Compare this to femicide in Mexico. The NIH, PubMed and other sites offer detailed information on annual femicide numbers in that country. Nonprofit sites say the rate of femicides in the US exceeds that of other wealthy countries, 2.2 per 100,000 people. You’d never know this from the FBI’s website with its overlapping subjective relationship categories and no designation for femicides. 

The confusions don’t end here. What about classifying murder victims by weapon? That should be straightforward. But the FBI provides another profusion of categories that don’t make sense. Firearms (22,542) are separated from handguns (32,443), rifles (2116), shotguns (790) and other firearms (1258), not to mention personal weapons (2846) and other (5137) , whatever that means. Aren’t handguns and rifles and shotguns firearms?

Surely annual gun murders in the US should be an easy statistic to find but it turns out to be political. The CDC, Pew Research Center and many nonprofits report this online. But the wording is fraught. Some sites give figures for ‘gun deaths’ or ‘gun violence’ combining gun suicides with gun homicides. Gun rights advocates complain this is done to inflate the numbers and scare people into approving gun restrictions. Murder-suicides confuse this even more, and there are plenty of those: 570 in 2020 causing over 1200 deaths. So no one agrees on how to report murders because they’re worried about the consequences of facing the truth, whatever that is.

As an innocent crime fiction buff I wanted to find out how many and what type of homicides occur in the US so I could judge how well crime fiction represents reality, when it tries. I’m also a regular person who would like to use statistics to help me evaluate the danger or security of my environment. Forget about it, as they say on the streets of NYC where police don’t bother to tell the FBI who got killed. Finding the basics of how many people are murdered every year in the US and who’s doing it to whom is hard to do. If you know something about this, please enlighten us. Or simply let me know what you think.