Understanding St. Louis: Total Crime Index and Crime-Ridden Neighborhoods

Tweet about this on TwitterShare on Facebook0Share on Reddit0Print this pageEmail this to someone

Crime and crime reporting shape our perception of St. Louis. A recent murder on a Saturday afternoon in the city's busy Central West End neighborhood has brought more attention to crime in the city. In the wake of the murder, other armed robberies were retroactively reported. Some have even noticed that not all murders are treated equally by the local press. In the first six months of 2012, there were 70 homicides reported in the City of St. Louis. The city is roughly on pace to equal the 143 homicides in 2009 and 144 in 2010, outpacing the 113 recorded last year.

The local press has taken notice of the increase in crimes committed with guns, rising number of homicides, and yes, the CWE murder (two individuals have been apprehended and charged with the crime). The mayor and St. Louis Metropolitan Police Department have held press conferences. Alderman Antonio French held an open house for the 18-camera crime surveillance center he introduced in the 21st Ward. The news coverage wasn't all favorable. "I resent @fox2now's description of my neighborhood as "crime-ridden". Both O'Fallon and Penrose have lower crime indexes than Downtown." French wrote on his Twitter account.

It seems like a good time to ask: What is the "crime index", what does it tell us about crime in our neighborhoods and is it fair to label the O'Fallon and Penrose neighborhoods (or others) as "crime-ridden"? We examined crime statistics released by the SLMPD for the first six months of 2012 and population data from the 2010 Census. Statistics cannot provide a comprehensive explanation of crime and the information here should be considered as simply one more piece of information to consider for a larger conversation regarding our city neighborhoods and relative crime and safety.

One simple, but key, takeaway is that the crime index provides a very highly distorted view of crime in our neighborhoods. In 2004 the FBI discontinued using a total crime index, the sum of eight different crimes reported in standardized form by jurisdiction across the nation. Their concern was that a raw count overemphasized larceny-thefts (almost 60 of the crimes reported) and diminished the focus on crimes against persons, such as aggravated assault, murder and rape. When you see crime rankings today, they based on these numbers. Some weight crimes, using a point scale to assign a number and ranking to neighborhoods and cities.

Looking at the 79 City of St. Louis neighborhoods using the simple sum of crimes reported, the crime index, we see that Downtown, Dutchtown, Downtown West, Tower Grove South and the Central West End lead. Are these the most "crime-ridden" parts of the city, to be avoided at all costs? Of course not. They sit at the top of the total crime index because they are the city's most populous neighborhoods (Dutchtown, Tower Grove South and Central West End) and the most visited (Downtown and Downtown West).
stl neighborhood crime rankings - total crimes

The graph shows central and south city neighborhoods at the top, followed by several north side neighborhoods, but in a broad sense, neighborhoods from across the city can be found from top to bottom. But again, the neighborhoods with the greatest number of reported crimes are also the largest neighborhoods, while the same trend can be seen in reverse with the city's smallest enclaves.

Considering the number of crimes per 1,000 residents provides more perspective. The central city Kings Oak neighborhood reported just 12 crimes counted for the crime index, but it also is home to just 180 residents. Ranked 79th by the crime index, Kings Oak rises to 20 of 79 with 67 crimes per 1,000 residents. There are other things to consider as well, such as the number of non-resident employees and visitors to city neighborhoods. When considering crimes per 1,000 residents in Koskiusko (3,929), Near North Riverfront (230), Downtown (201), Riverview (181) and Downtown West (152) neighborhoods, the rates are inflated by either a very low resident population in a large geographical area, or an influx of visitors (10s of thousands per day in the case of Downtown and Downtown West). Those five neighborhoods are excluded from the graphic below for this reason. This same affect also likely impacts Forest Park Southeast and Mark Twain/I-70 Industrial, but perhaps to a lesser extent.

Looking at the city's 79 neighborhoods by crimes per 1,000 residents, we start to see that south, central and north city neighborhoods are less mixed. South and central city neighborhoods occupy the bottom 17 spots (fewest crimes per 1,000 residents) and 22 of the bottom 23 (of 79). North and central city neighborhoods occupy 19 of the top 20 spots (most crimes per 1,000 residents), with the 14-resident south city industrial neighborhood of Koskiusko sitting far and away at the top.
stl neighborhood crime rankings - crimes per 1K residents

How significant was the change in relative position of city neighborhoods when considering crimes per 1,000 residents instead of simply the sum of all reported crimes? The Central West End and Bevo Mill each dropped 54 places (of 79 total). The CWE may be the most surprising. Although it is generally considered safe, more than 15,000 employees and visitors to the Washington University/Barnes-Jewish Hospital medical campus enter the neighborhood daily. Lindenwood Park dropped 49 spots. Outside the industrial neighborhoods of Koskiusko, Near North Riverfront and Riverview, the small central city neighborhoods of Tiffany, Lasalle, Kings Oak and Botanical Heights climbed the most. The graphic below lists neighborhoods in order of greatest to least number of total crimes reported, with bars and numbers showing how far the neighborhood moved on the list when considering crimes per 1,000 residents.
stl neighborhood crime rankings - change in total vs per 1k resident

The takeaway here is that there is an incredible amount of movement when thinking of crime in relation to the number of residents in a neighborhood. Average movement on the list was more than 24 spots. All of this was meant only to highlight perhaps the most simple, but also common abuse of crimes statistics – the total crime index. Returning to our jumping off point, was Alderman French correct? The O'Fallon (18th, 253) and Penrose (12th, 388) neighborhoods do have a much lower crime index than Downtown (1st, 743). But if it hasn't been made clear, that's probably the worst measure of crime we can use.

On the more subjective issue of whather O'Fallon and Penrose should be considered "crime-ridden"? Not by any reasonable measure. Considering crimes per 1,000 residents, Penrose is 30th of 79 – lower than many central and north side neighborhoods and five south side neighborhoods. O'Fallon, at 55th, has fewer crimes per 1,000 residents than The Hill and Tower Grove South.

Adding the measure of neighborhood population improves our understanding of crime statistics. But what about the type of crimes and our perception of safety? The FBI uniform crime report includes both crimes against people and crimes against property. In many neighborhoods, the high number of larceny-thefts skew the crime index and the number of crimes per 1,000 residents. The prevalence of particular types of crimes shape perceptions of safety. Next we'll look at how the 79 City of St. Louis neighborhoods compare when looking at crimes against people and crimes against property.

A follow-up post expands on the issues here. See: Understanding St. Louis: Total Crime Index, Violent Crimes and Property Crimes in City Neighborhoods

Tweet about this on TwitterShare on Facebook0Share on Reddit0Print this pageEmail this to someone
  • Kevin Farrell

    According to the Partnership for Downtown (and using their boundary lines which encompass the city’s Downtown/Downtown West/Columbus Square and Carr Square neighborhoods) the average daily population Downtown is over 125,000. This includes the 85,000 workers, 13,500 residents, 5,000 hotel guests and 25,000 event attendees/ day visitors. When you look at crime stats based on the average daily population, the numbers look considerably different than basing it on the 13,500 residents.

  • Greyson

    Alex: you must not have much interaction with the immigrant population because it is a known fact that they are less likely trust police & report crimes. I work with that population every day & hear the most horrific crimes in their communities & absolutely nothing we can say can convince them to go to the police. Furthermore, you have to be aware that the ILLEGAL immigrant population has a long history of avoiding the police & not reporting crimes. Many immigrants come from countries where the police are the MOST corrupt & criminal so to try & convince them OUR police are “the good guys” is an uphill battle indeed. I am actually shocked to hear you think this is likely untrue. Spend some time talking with the immigrant population & you might get a surprising education.

    • Alex Ihnen

      I find that a “known fact” is often less than a “fact”. I get your point here and I absolutely believe/know/have read that illegal immigrants are less likely to go to the police. Perhaps we can surmise that first generation legal immigrants are reluctant as well? I’d like to see a study on the issue (I’m sure there are some out there) though. And relative to high-crime neighborhoods in STL City, where there are very few immigrants, my guess is still that crimes are less under-reported. An example: it would stand to reason that quite a few car break-ins, assaults, burglaries, etc. in high crime areas, where the perpetrator and victim may both be involved in illegal activity, would go unreported.

      • El_D_El

        I know I’m late to the conversation here, but just wanted to point out the use of the offensive and inaccurate term “illegal immigrant.” Human beings are not illegal, and in fact being in the country without a visa is not a criminal offense. This term is used to discriminate against and dehumanize undocumented immigrants. Can we all please try to stop using it?


  • Greyson Winters

    I think the point is well made about neighborhoods where residents are more likely to report crimes as areas with high immigrant populations ie Bevo & the Bosnian community are typically distrustful of police and not likely to report non-homicide crimes. I would also like to see a City vs County analysis. As a lifelong City native I get so tired of people trash talking our City. I have only been a victim of crime 3x in my life and all 3x happened in the County! The recent news reports about multiple car breakins in Glendale made me laugh because all of the victims admitted to leaving their vehicles unlocked! Really? Glendale isn’t Mayberry, folks. I just wish the City had more people willing to become engaged & demand better from our City leaders & politicians. No child should have to grow up where stray or intentional gunfire hits them inside their homes or at the playgrounds. Where is the outrage, people? If you aren’t outraged than you aren’t paying attention. And finally, something needs to be done about the police & media’s purposeful non-reporting of many incidents of crime in this City. One recent example of intentional nonreporting was the shooting at the Mansion House Apartments garage in broad daylight in a very busy part of Downtown. STL on a Saturday. The Post said that the police REFUSED to comment on the victim, suspects or motive. Really? A man in his 50s gets shot in the face in broad daylight & we don’t have the right to know about this? And many more crimes don’t even get a mention that happen in this City. Why? And why does no one seem to care?

    • Alex Ihnen

      Not sure I understand – I haven’t seen anything that would suggest that immigrant populations are less likely (let along “not likely”) to report crimes. If I had to guess, I would say it’s the opposite – or at least that those communities are more likely to report crimes than our city’s highest crime areas – which are not immigrant communities.

  • stlhistory

    Uniform crime reports for county municipalities are available online from 2010 at the FBI website (the 2011 report is preliminary and so is not finished). You can produce the total index crime by adding the total of violent and property crimes listed here (it also lists population for convenient rate computation):

    • Alex Ihnen

      Thanks for the link. Perhaps when first 1/2 2012 numbers are compiled, we’ll compare St. Louis suburban communities with city neighborhoods. The numbers are interesting – will put a couple items on Twitter.

  • onecity

    Another statistic it would be useful to cross with these: to what degree are crimes reported in each neighborhood? For instance, if one neighborhood reports 99% of all crimes and another reports x%. I don’t know if the police dept or others have a tool for estimating that, as it is obviously in the realm of pure speculation unless good studies on the subject exist. So Neighborhood A and B could have the same per capita crimes, but if Neighborhood A reports 50% and B reports 100%, A has 2x the crime rate in theory.

    • Roger Mexico

      Northside does seem to underreport property crimes. In northside (STLPD districts 5-8) 23% of total crimes were against persons but in south side (STLPD districts 1-3) only 12% were against persons. If one assumes that crimes against persons and property scale together, this suggests southsiders are more likely to call in property crimes. So I think a better way to compare crimes across neighborhoods would be to just look at the crimes against persons and not the entire UCR (persons + property).

    • Bob Cummings

      The US Dept of Justice regularly analyzes unreported crimes. The reviews are based on telephone surveys in various jurisdictions.

      Link to 2012 info:

  • Chippewa

    This is an extremely interesting analysis. I was admittedly surprised to see Forest Park Southeast so high on the list, and it is nice to see places like O’Fallon Park so low.

  • St. Louis City Resident

    Great analysis! Thank you for the cool charts and thoughtful discussion. I frequently make some of these points to my coworkers and to others who live outside the City. However, it is still unnerving to hear gunshots from the direction of Gravois Park (which we frequently do).

    • Alex Ihnen

      It would be great to be able to compare these numbers to county municipalities, but those numbers aren’t nearly as easy to come by. What we would undoubtedly find is that many city neighborhoods are as safe or safer than some county municipalities (they’re also often the same size, or larger). The difficulty in compiling crime stats across the metro region is another drawback of our fractured political region.

      • Dani Pizzella

        You can get the crime numbers from each muncipality. I had to do compare the numbers per capita for a college journalism class. I probably still have the numbers around but they would be from 2009. I clearly remember, however, afton having a higher crime rate than lindenwood park (where I lived at the time).

  • Paul Hohmann

    Another basic factor in the total crimes chart is that not the the neighborhoods that show up at the top of the list for the most part are the largest neighborhoods in area of land. Here’s the official map: http://stlouis-mo.gov/government/departments/public-safety/neighborhood-stabilization-office/neighborhoods/neighborhood-maps.cfm

    • Paul Hohmann

      Alex, this comment form really sucks because you can only see 2 lines of everything you are tying!

      • Alex Ihnen

        Paul – I’ve checked on several different browsers (Chrome, Firefox, IE) and haven’t had that issue. When typing, two lines appear, but the box expands as I continue typing my comment. What are you using?

      • john w.

        draft everything in Wordpad or Word first, then just paste it into the box.

    • Alex Ihnen

      Right – this was mentioned in the post.

  • Scott Ogilvie

    The problem for Downtown, or any area that has a lot more employment than residents, is that crime will always look higher on a per-capita basis than it really is. The other problem is that 79 neighborhoods is probably a data set that’s a little too fine. A neighborhood like Ellendale gets skewed way up the ladder simply because people shoplift from K-Mart, and so on. A really interesting stat, which we don’t have available, is how likely a resident of a neighborhood is to be a victim of a crime.

    • Alex Ihnen

      Good points. What we would really want to know for each crime is where the victim resides. It may be that downtown residents are very unlikely to be victims of crime – we don’t know. It would also be good to know the number of non-resident workers and visitors. This could be measured in “resident hours”, or some measure that could approximate how many people are present in a neighborhood such as downtown at any given time.