Measuring victimization: the National Crime Victimization Survey and reporting to police

Linh Tran has a question before the lesson starts. In LES-102 the class learned that police statistics miss a great deal of crime, the so-called dark figure. 'If most crime is never reported,' she asks, 'how does anyone know how much there is?' Brooks says that is exactly what today's lesson is about, and that the answer matters for victims as much as for researchers. Every unreported crime is a person who decided, for some reason, that calling the police was not worth it. Understanding those reasons is the first step toward changing them.

Two ways of counting

There are two main ways to count crime. Police-reported statistics count incidents that come to the attention of the police and are recorded. In the United States, police agencies report to the FBI's Uniform Crime Reporting program, which since 2021 has collected data only through the National Incident-Based Reporting System, known as NIBRS. Police data are detailed and timely, cover every community and include crimes against businesses and homicides. But they count only what victims report and police record.

Victimization surveys go around the police entirely. They ask a large random sample of people whether they have been victims of crime during a recent period, whether they reported it, and what happened afterward. Because they ask people directly, they capture crimes that never reached the police. They also ask about the effects of crime, contact with victim services, and fear and perceptions of safety. Their limits are the mirror image of police data: they rely on people remembering and being willing to disclose, they cannot count homicides or crimes against businesses, and their samples are too small to say much about a single city such as Port Ardent.

The National Crime Victimization Survey

The Bureau of Justice Statistics has run the National Crime Victimization Survey since 1973, when it began as the National Crime Survey. Households are selected at random and stay in the sample for three and a half years, and every person aged 12 or older in the household is interviewed every six months, in person or by telephone. The survey asks about rape and sexual assault, robbery, aggravated and simple assault, personal larceny, household burglary, motor vehicle theft and other theft. It does not cover homicide, crimes against businesses or victims younger than 12.

The 2023 results are a useful snapshot. The rate of violent victimization was about 22.5 per 1,000 people aged 12 or older, similar to the 2022 rate, and there were about 13.6 million property victimizations. Only about 8 in every 100 victims of violent crime received help from a victim service provider other than the police.

How much is reported

The most important figures for police are the reporting rates. The 2023 survey found that almost half of violent victimizations were reported to police, and about 30 in every 100 property victimizations. Reporting rates vary from year to year and by crime type, and some estimates, especially for rape and sexual assault, rest on small samples and have wide margins of error.

Why do people not report? The surveys ask, and the answers are consistent across years and countries. The most common reason is that the crime seemed too minor to be worth reporting, or that nobody was hurt. Many victims of property crime believe the police could not or would not do anything, for example that stolen property would never be found. Some victims deal with the matter another way, for instance through a school, an employer or the family. Some consider it a private matter. Others fear reprisal from the offender, do not want the offender, often someone they know, to get into trouble, or feel ashamed or worried that they will not be believed. For sexual assault and intimate partner violence, these last reasons are much more common.

Brooks asks the class which of those reasons the police can influence. Several, the class decides. Victims who believe the police will not take them seriously, will not do anything or will treat them badly are making a prediction based on what they have heard or experienced. Every interaction an officer has with a victim, and every story that victim tells others afterward, shapes those predictions for the whole community. Amara Okafor points out that reporting also depends on practical things: whether a victim can report online, whether someone speaks their language, whether they can report at a hospital or a victim services agency instead of a police station.

Why reporting matters, and why it is not always the goal

Reporting matters because the police cannot investigate what they do not know about, and patterns such as a series of break-ins or repeated assaults by the same person stay invisible. It also matters to victims directly, because many forms of help depend on it. The Ohio Victims of Crime Compensation Program requires that the crime be reported to law enforcement and that the victim cooperate with the investigation.

Brooks adds an important caution. Victim services exist for victims, whether or not they report. A woman who has been sexually assaulted may want medical care, counseling and safety planning without a police investigation, and she is entitled to that choice. Many hospitals offer sexual assault nurse examiner programs that can collect and store forensic evidence so that a victim can decide later whether to go to police. Pressuring a victim to report can repeat the loss of control that the crime itself caused. The goal is to make reporting possible, safe and worthwhile, not to make it compulsory.

What surveys say about the effects of crime

Victimization surveys do more than count incidents. They also ask victims how the crime affected them, and the answers are a useful reminder that even crimes the justice system treats as minor can have real consequences. Victims commonly report feeling angry, upset, confused or frustrated, and a substantial share report feeling more fearful or more cautious afterward, having trouble sleeping, or worrying about their safety at home. Victims of violence are more likely than victims of property crime to report lasting effects, and a minority of violent crime victims describe symptoms consistent with post-traumatic stress, which the class will study in the next lesson. Many victims also report practical costs: time off work, medical bills, repairs, replacing documents and cards, and hours spent dealing with insurers and banks.

Ruth Lindqvist's losses illustrate the point. Her insurer paid for most of the stolen tools and the broken door, but she paid a deductible of 1,000 dollars, lost two days of sales while the shop was closed, and spent most of a week on the claim. She also began to wake at night when a car passed the shop, and asked her nephew to come with her when she locked up. None of this appears in the police statistics, which record a single break and enter, cleared by charge.

Who the numbers miss

Brooks warns that even surveys miss some victims. People without a fixed address are hard to sample, as are people living in shelters, hospitals, prisons, group homes and long-term care facilities, where some of the most vulnerable people live. People who do not speak the survey's languages, or who distrust government callers, may refuse to take part. Children are covered only from age 12 in the NCVS, and very young children not at all. Some people do not recognize what happened to them as a crime, especially when the offender is a partner, family member or caregiver. For these groups, researchers rely on special surveys, service records from shelters, hospitals and child protection agencies, and studies done in partnership with community organizations. An officer who assumes that the official figures describe everyone may underestimate the harm in the very communities that need the most help.

Estimating the dark figure locally

National surveys cannot describe Port Ardent directly, but they can help analysts estimate how much crime the city's police statistics might be missing. If police recorded a certain number of break and enters, and surveys suggest that about half of break and enters are reported, the true number could be roughly twice the recorded number. Linh Tran uses this idea in the program below. She is careful to call the results rough estimates: the national reporting rates may not fit Port Ardent, and some reported incidents are never recorded.

# Rough estimate of total incidents in Port Ardent from recorded counts and survey reporting rates
recorded = {"break and enter": 588, "robbery": 61, "assault": 1240, "theft from vehicle": 702}
# assumed share of incidents that victims report (illustrative, based on national surveys)
report_rate = {"break and enter": 0.50, "robbery": 0.48, "assault": 0.40, "theft from vehicle": 0.30}
population = 98000
total_est = 0
for crime, n in recorded.items():
    est = round(n / report_rate[crime])
    total_est += est
    print(f"{crime:19} recorded {n:5}  estimated {est:5}  unreported {est - n:5}")
print("recorded total:", sum(recorded.values()), "estimated total:", total_est)
print("estimated incidents per 1,000 residents:", round(total_est / population * 1000, 1))

break and enter     recorded   588  estimated  1176  unreported   588
robbery             recorded    61  estimated   127  unreported    66
assault             recorded  1240  estimated  3100  unreported  1860
theft from vehicle  recorded   702  estimated  2340  unreported  1638
recorded total: 2591 estimated total: 6743
estimated incidents per 1,000 residents: 68.8

The estimate suggests that thousands of incidents in a year never reached the police. Desjardins points out what that means at the front counter: for every victim who calls, there may be one or two more who decided not to. The ones who do call are, in a sense, giving the police a chance to prove that calling was worth it.

Reading victimization statistics carefully

Brooks gives the class four habits for reading victimization statistics. First, check what is being counted: incidents, victims or households, and whether the figure is a rate or a count. A rise in the number of incidents can simply reflect population growth. Second, check the period and the population: a survey of people 15 and older says nothing about younger children. Third, look at the margin of error, especially for small groups and rare crimes; a change from one survey to the next may not be real. Fourth, remember that a rise in police-reported crime can mean more crime, more reporting, or changes in how police record incidents. A campaign that encourages victims of intimate partner violence to come forward should increase police-reported numbers, and that would be a success, not a failure.

Priya Sandhu, who spent five years as a bank fraud analyst, adds a fifth habit from her own work: some crimes are badly measured by both methods. Fraud victims often report to their bank, not the police, and many never realize they were defrauded at all. The FBI's Internet Crime Complaint Center receives hundreds of thousands of complaints a year, but researchers believe most online fraud is never reported to anyone. Grace Whitford almost did not report hers. She told Brooks later that she felt foolish, and that her first thought was that she should not have to tell anyone.