Chapter 2 SRS Overview
One of the first, and most important, questions I think people have about crime is a simple one: is crime going up? Answering it seems simple - you just count up all the crimes that happen in an area and see if that number of bigger than it was in previous times.
However, putting this into practice invites a number of questions, all of which affect how we measure crime. First, we need to define what a crime is. Not philosophically what actions are crimes - or what should be crimes - but literally which of the many thousands of different criminal acts (crimes as defined by state or federal law) should be considered in this measure. Should murder count? Most people would say yes. How about jaywalking or speeding? Many would say probably not. Should marital rape be considered a crime? Now, certainly most people (all, I would hope) would say yes. But in much of the United States it was not a crime until the 1970s (Bennice and Resick 2003; McMahon-Howard, Clay-Warner, and Renzulli 2009). What about marijuana possession, an act that is illegal nationally and in some states, but legal elsewhere?
Next, we have to know what geographic and time unit to measure crimes at since these decisions determine how precise we can measure crime and when it changed. That is, if you are mugged on Jan 1st at exactly 12:15pm right outside your house, how do we record it? Should we be as precise as including the exact time and location (such as your home address)? Out of privacy concerns to the victim, should we only include a larger time unit (such as hour of the day or just the day without any time of day) or a larger geographic unit (such as a Census Tract or the city)? And what about when there are questions about geographic jurisdiction such as a local police and sheriff overlapping in their patrol area? Or when a crime happens over the course of several hours (e.g. a length burglary or kidnapping)? Should we count the start time, the end time, or somewhere in the middle?
The final question is that when a crime occurs, how do we know? That is, when we want to count how many crimes occurred do we ask people how often they’ve been victimized, do we ask people how often they commit a crime, do we look at crimes reported to police, crimes charged in a criminal court? Each of these measures will likely give different answers as to how many crimes occurred. And what about crimes where a victim reports it and the police investigate and decide that it did not occur (e.g. victim reports an act which was not actually illegal)? Or where the police say a crime occurred by the local District Attorney declines to prosecute?
The FBI answered all of these questions in 1929 when they began the Uniform Crime Reporting (SRS) Program Data, or SRS data for short. Crime consists of eight crime categories - murder, rape, robbery, aggravated assault, burglary, motor vehicle theft, theft, and simple assault - that are reported to the police and is collected each month from each agency in the country. So essentially we know how many of a small number of crime categories happen each month in each city (though some cities have multiple police agencies so even this is more complicated than it seems). These decisions, born primarily out of the resource limitations of 1929 (i.e. no computers), have had a major impact on criminology research. The first seven crime categories - known as “Index Crimes” or “Part 1 crimes” (or “Part I” sometimes) - are the ones used to measure crime in many criminology papers, even when the researchers have access to data that covers a broader selection of crimes than these.12 These are also the crimes that the news uses each year to report on how crime in the United States compares to the previous year. The crime data actually also includes the final crime, simple assault, though it is not included as an index crime and is, therefore, generally ignored by researchers - a large flaw in most studies that we will discuss in more detail in Section 3.1.2. In fact excluding simple assault is such a large issue that doing so has led us to exclude the most common violent crime from most government, news, and academic reports, leading to an undercount of violence in this country by approximately 50%.
In this chapter we will provide a brief overview of the datasets that make up the FBI’s Summary Reporting System (SRS) which you can think of as the “aggregate” data collection that is part of their Uniform Crime Reporting (UCR) Program Data. These datasets are primarily focused on aggregate data, mostly having monthly counts of crimes without many details about any particular crime. In comparison, the other half of the UCR collection, the National Incident-Based Reporting System (NIBRS) data is focused entirely on providing detail on every individual crime reported to the police. You can actually convert NIBRS data to SRS in most cases - and FBI does for agencies that report only NIBRS data - but there is no way to go from SRS to NIBRS.
2.1 Crimes included in the SRS datasets
SRS data covers only a subset - and for the crime data, a very small subset - of all crimes that can occur. It also only includes crimes reported to the police. So there are two levels of abstraction from a crime occurring to it being included in the data: a crime must occur that is one of the crimes included in one of the SRS datasets (we detail all of these crimes below) and the victim must report the incident to the police. While the crimes included are only a small selection of crimes - which were originally chosen since at the time the SRS was designed these were considered important offenses and among the best reported - this is an important first step to understanding the data.
SRS data should be understood as a loose collection of data that seeks to understand crimes and arrests in the United States. There are seven datasets in the SRS collection that each cover a different topic or a subset of a previously covered topic: crimes, arrests, property crimes specifically, homicides specifically, police officers killed or assaulted, arson specifically, and hate crimes.13 In this section we will go over the crimes included in the two main SRS datasets: Offenses Known and Clearances by Arrest and (which I like to call the “crime” dataset) and the Arrests by Age, Sex, and Race (the “arrests” dataset). These are the most commonly used SRS datasets and the stolen property and homicide datasets are simply more detailed subsets of these datasets. The hate crime dataset can cover a broader selection of offenses than in the crimes or arrests data, so we will discuss those in Chapter @ref(hate_crimes).
2.2 A summary of each SRS dataset
The SRS collection of data can be roughly summarized into two groups: crime data and arrest data. While there are several datasets included in the SRS data collection, they are all extensions of one of the above groups. For arrest data, you have information about who (by race and by age-gender, but not by race-gender or race-age other than within race you know if the arrestee is an adult or a juvenile) was arrested for each crime. For crime data, you have monthly counts of a small number of crimes (many fewer than crimes covered in the arrest data) and then more specialized data on a subset of these crimes - information on homicides, hate crimes, assaults or killings of police officers, and stolen property.
Each of these datasets will have its own chapter in this book where we discuss the data thoroughly. Here is a very brief summary of each dataset which will help you know which one to use for your research. I still recommend reading that data’s chapter since it covers important caveats and uses (or misuses) of the data that would not be covered below.
2.2.1 Offenses Known and Clearances by Arrest (1960 - present)
The Offenses Known and Clearances by Arrest dataset - often called Return A, “Offenses Known” or, less commonly, OKCA - is the oldest and most commonly used dataset and measures crimes reported to the police. For this reason it is used as the main measure of crime in the United States, and I tend to call it the “crimes dataset.” This data answers the most basic questions about crimes: how many occurred? If you see crime data referenced in a news or academic article it is usually this data. This data also includes the number of crimes solved (though with a weaker definition of “solved” than you may think) and the number of crimes in which the agency concluded did not actually occur which they call an “unfounded” crime. This data has the monthly number of crimes - for a select group of crimes types - that occurred in an agency, as well as how many the police investigated and decided did not occur, and the number “cleared” by an arrest. The data uses something called a Hierarchy Rule which means that in incidents with multiple crimes, only the most serious is recorded - though in practice this affects only a small percent of cases, and primarily affects property crimes.
2.2.2 Arrests by Age, Sex, and Race (1974 - present)
The Arrests by Age, Sex, and Race dataset - often called ASR, or the “arrests data”, or sometimes the “Arrests by Age, Sex, Race, and Ethnicity though this is really misleading since most years do not even report ethnicity data - includes the monthly number of arrests for a variety of crimes and, unlike the crime data, breaks down this data by age and gender. This data includes a broader number of crime categories than the crime dataset (the Offenses Known and Clearances by Arrest data) though is less detailed on violent crimes since it does not breakdown aggravated assault or robberies by weapon type as the Offenses Known data does.
For each crime it says the number of arrests for each gender-age group with younger ages (15-24) showing the arrestee’s age to the year (e.g. age 16, age 17) and other ages grouping years together (e.g. age 25-29, 30-34, “under 10”). It also breaks down arrests by race-age by including the number of arrestees of each race (American Indian, Asian, Black, and White are the only included races) and if the arrestee is a juvenile (<18 years old) or an adult. The data does technically include a breakdown by ethnicity-age (e.g. juvenile-Hispanic, juvenile-non-Hispanic) but almost no agencies report this data (for most years zero agencies report ethnicity at all) so in practice the data does not include ethnicity. As the data includes counts of arrestees, people who are arrested multiple times are included in the data multiple times - it is not a measure of unique arrestees.
2.2.3 Law Enforcement Officers Killed and Assaulted (LEOKA) (1960 - present)
The Law Enforcement Officers Killed and Assaulted data, often called just by its acronym LEOKA (“LEE-OH-KUH”), has two main purposes.14 First, it provides counts of employees employed by each agency - broken down by if they are civilian employees or sworn officers, and also broken down by sex (male and female are the only options). And second, it measures how many officers were assaulted or killed (including officers who die accidentally such as in a car crash) in a given month. The assault data is also broken down into shift type (e.g. alone, with a partner, on foot, in a car, etc.), the offender’s weapon, and type of call they are responding to (e.g. robbery, disturbance, traffic stop). The killed data simply says how many officers are killed feloniously (i.e. murdered) or died accidentally (e.g. car crash) in a given month. The employee information is at the year-level so you know, for example, how many male police officers were employed in a given year at an agency, but do not know any more than that such as if the number employed changed over the year. This dataset is commonly used as a measure of police employees and is a generally reliable measure of how many police are employed by a police agency. The second part of this data, measuring assaults and deaths, is more flawed with missing data issues and data error issues (e.g. more officers killed than employed in an agency).
2.2.4 Supplementary Homicide Reports (SHR)
The Supplementary Homicide Reports dataset - often abbreviated to SHR - is the most detailed of the SRS datasets and provides information about the circumstances and participants (victim and offender demographics and relationship status) for homicides. For each homicide incident it tells you the age, gender, race, and ethnicity of each victim and offender as well as the relationship between the first victim and each of the offenders (but not the other victims in cases where there are multiple victims). It also tells you the weapon used by each offender and the circumstance of the killing, such as a “lovers triangle” or a gang-related murder. As with other SRS data, it also tells you the agency it occurred in and the month and year when the crime happened.
One important point of clarification: this is not the number of murders, though it does track that. This data also includes the number of homicides that are manslaughter by negligence (e.g. children playing with a gun, hunting accident) and justifiable homicides (i.e. not criminal). So be carefully when speaking about this data. It is murders but not only murders so you want to speak precisely.
2.2.5 Hate Crime Data (1991 - present)
This dataset covers crimes that are reported to the police and judged by the police to be motivated by hate. More specifically, they are (1) crimes which were (2) motivated - at least in part - by bias towards a certain person or group of people because of characteristics about them such as race, sexual orientation, or religion. The first part is key, they must be crimes - and really must be the selection of crimes that the FBI collects for this dataset. Biased actions that do not meet the standard of a crime, or are of a crime category not included in this data, are not considered hate crimes. For example, if someone yells at a Black person and uses racial slurs against them, it is clearly a racist action. For it to be included in this data, however, it would have to extend to a threat since “intimidation” is a crime included in this data but lesser actions such as simply insulting someone is not included. For the second part, the bias motivation, it must be against a group that the FBI includes in this data. For example, when this data collection began in 1991 crimes against transgender people were not counted so if a transgender person was assaulted or killed because they were transgender, this is not a hate crime recorded in the data (though it would have counted in the “Anti-Lesbian, Gay, Bisexual, Or Transgender, Mixed Group (LGBT)” bias motivation which was always reported).
2.2.6 Property Stolen and Recovered (Supplement to Return A) (1960 - present)
The Property Stolen and Recovered data - sometimes called the Supplement to Return A (Return A being another name for the Offenses Known and Clearances by Arrest dataset, the “crime” dataset) - provides monthly information about property-related offenses (theft, motor vehicle theft, robbery, and burglary), including the location of the offense (in broad categories like “gas station” or “residence”), what was stolen (e.g. clothing, livestock, firearms), and how much the stolen items were worth.15 It also includes robberies so is really the “stuff taken during a crime” dataset than the dataset about property crimes. The “recovered” part of this dataset covers the type and value of property recovered so you can use this, along with the type and value of property stolen, to determine what percent and type of items the police managed to recover.
Like most other SRS datasets this is at the agency-month level so you can, for example, learn how often burglaries occur at the victim’s home during the day, and if that rate changes over the year or differs across agencies. The data, however, provides no information about the offender or the victim (other than if the victim was an individual or a commercial business 16). The value of the property stolen is primarily based on the victim’s estimate of how much the item is worth (items that are decreased in value once used - such as cars - are supposed to be valued at the current market rate, but the data provides no indication of when it uses the current market rate or the victim’s estimate) so it should be used as a very rough estimate of value.
2.2.7 Arson (1979 - present)
The arson dataset provides monthly counts at the police agency-level for arsons that occur, and includes a breakdown of arsons by the type of arson (e.g. arson of a person’s home, arson of a vehicle, arson of a community/public building) and the estimated value of the damage caused by the arson. This data includes all arsons reported to the police or otherwise known to the police (e.g. discovered while on patrol) and also has a count of arsons that lead to an arrest (of at least one person who committed the arson) and reports that turned out to not be arsons (such as if an investigation found that the fire was started accidentally). This is essentially the Offenses Known and Clearances by Arrest data but only for arsons. The data even follows the same definitions for categories such as what counts as a cleared or unfounded crime. The primary additional variable is the estimated damage in dollars caused by the arson.
References
Arson is also an index crime but was added after these initial seven were chosen and is not included in the crimes dataset (though is available separately) so is generally not included in studies that use index crimes.↩︎
There is also a human trafficking dataset though this is a newer dataset and rarely used so I will not cover it in this book.↩︎
This data is also sometimes called the “Police Employees” dataset.↩︎
It also includes the value of items stolen during rapes and murders, if anything was stolen.↩︎
based on the location of the incident - “bank”, “gas station”, etc.↩︎