U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics ---------------------------------------------------------- This file is text only without graphics and many of the tables. A Zip archive of the tables in this report in spreadsheet format (.csv) and the full report including tables and graphics in .pdf format are available on BJS website at: http://www.bjs.gov/index.cfm?ty=pbdetail&iid=4759 ---------------------------------------------------------- ******************* Technical Report ******************* Measuring the Prevalence of Crime with the National Crime Victimization Survey Janet L. Lauritsen, BJS Visiting Fellow Maribeth L. Rezey, University of Missouri - St. Louis ************** Introduction ************** The Bureau of Justice Statistics’ (BJS) National Crime Victimization Survey (NCVS) is an annual data collection designed to gather information about nonfatal personal crimes and household property crimes in the United States. The main purpose of the NCVS is to accurately measure the number and type of criminal victimizations that occur each year to persons age 12 or older. Victimization rates are most commonly used in NCVS reports to describe changes in the level of personal and household crime over time and the levels of crime experienced by different population subgroups. However, prevalence rates also may be used to describe changes in the level of crime over time and differences between subgroups. This report discusses victimization rates and prevalence rates and the value of each type of indicator for understanding criminal victimization. Using NCVS data, the report displays and compares trends in the victimization and prevalence rates from 1993 to 2010. It also analyzes the differences between victimization and prevalence rates for various types of crime and demographic groups using data from the 2010 NCVS. Defining victimization rates and prevalence rates Annual estimates of a population’s risk for criminal victimization can be examined using victimization rates, incident rates, or prevalence rates. Historically, BJS reports using NCVS data rely on victimization rates, which measure the extent to which victimizations occur in a specified population and time period. For crimes affecting persons, NCVS victimization rates are estimated by dividing the number of victimizations that occur during a specified time period (T) by the population at risk for those victimizations and multiplying the rate by 1,000. These victimization rates permit comparisons of crime over time and between population subgroups. Number of victimizations experienced Victimization rate T = by specified population T _________________________________________________________x1,000 Number of persons in the specified population T A victimization rate is typically larger than an alternative incident rate because more than one person can be victimized during a single incident of crime. For example, the armed robbery of two persons would be counted as one incident but as two victimizations. Using a victimization rate rather than an incident rate provides an estimate of the number of victimizations experienced by individuals (or by households in the case of household-based crimes). When the purpose of the rate is to describe the level of crime among persons in the population, victimization rates tend to be preferred over incident rates. Incident rates are more likely to be used when making comparisons to official law enforcement crime data, which generally count crime by the number of incidents rather than the number of victims. Prevalence rates also describe the level of victimization but are based on the number of persons (or households) in the population who experienced at least one victimization during a specified time period. The key distinction between a victimization or incident rate and a prevalence rate is whether the numerator consists of the number of victimizations or the number of victims. Prevalence rates do not take into account the number of victimizations each victim experiences. These rates tell about the risk of experiencing at least one crime in a given period. ************* HIGHLIGHTS ************* * From 1993 to 2010, the decline in violent victimization rates (down 76%) was greater than the decline in prevalence rates (down 63%). For serious violent crimes, the victimization rate decreased 77% and the prevalence rate decreased 66%. * In 1993, 77% of violent crime victims reported that they were victimized one time during the year, compared to 83% in 2010. * The percentage of violent crime victims who experienced two or more victimizations during a year declined from 23% in 1993 to 17% in 2010. In 2010, this 17% accounted for more than half (54%) of all violent victimizations. * Victims of intimate partner violence (21%) were more likely to experience repeat victimization within the year than were victims of stranger violence (9%). * From 1993 to 2010, the decline in total household property crime victimization rates (down 64%) was greater than the decline in prevalence rates (down 48%). * The proportion of household property crime victims who reported two or more incidents during the year decreased from 25% in 1993 to 18% in 2010. In 2010, the 18% of repeat household victims accounted for about 41% of all household property victimizations. * In 2010, 12% of burglary victims reported two or more incidents during the year, accounting for about 34% of all burglary victimizations. **************************************************** BJS Visiting Fellows --------------------- The BJS Visiting Fellows program is intended to facilitate collaboration between academic scholars and government researchers in survey methodology, statistics, economics, and social sciences. Visiting Fellows have the unique opportunity to address substantive, methodological and analytic issues relevant to BJS programs, and to further knowledge and understanding of criminal justice systems operation. Fellows conduct research at BJS or at their home site, use BJS data and facilities, and interact with BJS staff. Pending available funding, the program is normally conducted each year through a competitive solicitation. Applicants are limited to senior-level social science researchers or statisticians in the fields of statistics, survey methodology, mathematics, criminology, demography, economics, behavioral science, and other related fields. They must have an established research record in their field, have considerable expertise in their area of proposed research, and be willing to commit a substantial portion of their time (typically 6 to 18 months) to undertake analyses of existing BJS data. In addition, they must produce a report that both summarizes their analyses and meets BJS publication and data quality standards. For further information, see: http://www.bjs.gov/content/fellows.cfm. **************************************************** While prevalence rates are not often produced using NCVS data, it is possible to do so. For crimes affecting persons, prevalence rates are estimated by dividing the number of victims in the specified population by the total number of persons in the population and multiplying the rate by 1,000. When the same multiplier (1,000) is used, the prevalence rate can be compared to the victimization rate. Prevalence rates can also be based out of 100, which is simply the percentage of the population victimized at least once in a given period of time. Number of victims in a specified population T Prevalence rate T =_________________________________________-x 1,000 Number of persons in the specified population T Victimization and prevalence rates may also be produced for household-based crimes, such as burglary. In these instances, the numerators and denominators are adjusted accordingly to reflect households rather than persons. When the prevalence rate and the victimization rate are estimated on a per 1,000 basis, the trends can show changes in the mean number of incidents per victim over time or the concentration of victimizations per victim. The mean number of incidents per victim is estimated by dividing the number of victimizations by the number of victims. Higher mean values are one indicator of the extent to which repeat victimization occurs. Number of victimizations Concentration = ________________________________ Number of victims Once repeat victims are identified, the percentage of victimizations accounted for by these repeat victims can then be estimated using the following formula: [V – (Y x P)] / V x 100 where V is the victimization rate per 1,000 Y is the proportion of victims with only one victimization P is the prevalence rate per 1,000. ************************************** The value of measuring victimization risk using different rates ************************************** Both victimization and prevalence rates provide information about the level of and risk for crime. Using these together will better inform changes in the crime rate over time. For example, the crime rate may increase because of more victims or victimizations per victim or both. If the crime rate increases because there are more victimizations per victim (an increase in repeat victimization), a person’s risk of experiencing any victimization has not changed, but their risk for repeated victimization if they are a victim has increased. A considerable portion of the victimizations that occur (typically within a single year) are experienced by a small number of repeat victims. ***Footnote 1 See Grove, L., G. Farrell, D. Farrington, and S. Johnson. (2012). Preventing repeat victimization: A systematic review. Stockholm: The Swedish National Council for Crime Prevention***. This suggests that identifying repeat victims may present a unique opportunity to reduce crime rates, as it focuses attention on those who disproportionately experience victimization. However, if repeat victims are unique from other crime victims, focusing on repeat victims may result in crime reductions that are not uniform across all types of crime, victim characteristics, and degrees of crime severity. The extent of repeat victimization and the potential value of focusing efforts on repeat victims can only be known if prevalence rates and victimization rates are distinguished. For instance, some forms of repeat victimization (such as household burglary) tend to occur within relatively short time intervals after the initial victimization. ***Footnote 2 Forrester, D., S. Frenz, M. O’Connell, and K. Pease. (1990). The Kirkholt burglary prevention project. Home Office Crime Prevention Unit Paper 13. London: Home Office***. If underlying patterns of repeat victimization can be determined by examining prevalence rates, efforts focused on reducing future victimization by those recently victimized may result in greater decreases in crime than strategies that are more broadly targeted. It is also valuable to understand and report victimization rates. To a large extent, reports based on NCVS data typically provide annual victimization rates rather than prevalence rates because the NCVS serves as an alternative to annual police-based estimates of crime from the FBI’s Uniform Crime Reports (UCR). ***Footnote 3 Groves, R. and D. Cork., eds. (2008). Surveying victims: Options for conducting the National Crime Victimization Survey. Washington DC: National Research Council***. NCVS-based estimates of the levels and changes in crime are routinely compared to estimates generated using UCR data, which only include incidents reported to or recorded by the police. Prevalence rates cannot be easily compared to police rates of crime because they count unique victims and not the number of crimes. Reporting on changes in the victimization rate is also important because criminal events initiate criminal justice system activities, such as police investigations, arrests, and subsequent prosecutions, sentencing outcomes, and the provision of victim services. Therefore, NCVS victimization rates are capable of providing information that can be used for making key decisions for criminal justice system operations. ***************************************** Victimization rates and prevalence rates for personal violence: 1993–2010 ***************************************** The NCVS measures the violent crimes of rape and sexual assault, robbery, aggravated assault, and simple assault, and includes both completed and attempted crimes. From 1993 to 2010, declines occurred in both the victimization rate and the prevalence rate of violent victimization (figure 1). The violent victimization rate declined 76%, from 79.8 per 1,000 persons age 12 or older in 1993 to 19.3 per 1,000 in 2010.***Footnote 4 Annual victimization rates include series victimizations and count such incidents based on the number of times the victim reported that the victimization occurred, up to a maximum of 10 times. See Methodology for more information.*** During the same period, the prevalence rate for violent crime declined 63%, from 29.3 to 10.8 per 1,000. Since the decline in the victimization rate was greater than the decline in the prevalence rate, the mean number of victimizations per victim also declined, from 2.7 victimizations per victim in 1993 to 1.8 in 2010 (table 1). Over time, both the number of persons victimized and the average number of victimizations each victim experienced declined, with the number of victimizations declining at a faster rate than the number of victims. In 1993, 6.2 million victims experienced 16.8 million victimizations, while in 2010, 2.8 million victims experienced 4.9 million victimizations (table 2). The distribution of violent victimization reported in the NCVS is skewed, as most persons experienced no victimizations and the majority of victims reported one victimization during the year. In 1993, of the 2.9% of the population that experienced at least one violent victimization, approximately 77% reported that they were victimized once, which increased to 83% in 2010 when 1.1% of the population experienced at least one violent victimization. About 23% of violent crime victims in 1993 and 17% in 2010 experienced repeat victimization. Over this period, both the percentage of the population victimized and the percentage of those repeatedly victimized declined. Because of this skew in the distribution of violent victimization, both the modal category (one victimization during the year) and the mean should be considered as statistical estimates of the risk for repeat victimization. It is not fully accurate to say that in 2010 the typical violent crime victim experienced 1.8 victimizations per year in 2010, since 83% of victims experienced only one victimization during the year. More accurately, most victims of violent crime in a given year experienced one victimization, and a small percentage of victims reported repeat victimizations. Though the percentage of victims who experienced repeat violence of any form in 2010 was comparatively small, the 17% who did experience repeat violence accounted for 54% of all of the violent victimizations that occurred in that year. In contrast, in 1993 the 23% of victims who experienced repeat victimization accounted for 72% of the violent victimizations that occurred that year (not shown in table). ***************** Serious violence ***************** The rate of serious violent victimization--rape and sexual assault, robbery, and aggravated assault--declined 77%, from 29.1 per 1,000 in 1993 to 6.6 per 1,000 in 2010 (figure 2). The prevalence rate declined 66%, from 12.6 per 1,000 in 1993 to 4.3 per 1,000 in 2010. Because the victimization rate declined more than the prevalence rate, the average number of serious violent victimizations per victim declined from 2.3 to 1.5 victimizations (table 1). The majority of serious violent crime victims reported experiencing one victimization during the year. In 1993, 16% of serious violent crime victims experienced repeat victimization within the year, compared to 8% in 2010. Repeat victims accounted for 40% of all serious violence in 2010 and 63% of all serious violence in 1993 (not shown in table). ************************ Rape and sexual assault ************************ Similar to other forms of violence, the victimization and prevalence rates for rape and sexual assault among females age 12 or older declined overall from 1993 to 2010 (figure 3). The female victimization rate for rape and sexual assault declined 75% from 1993 to 2010 (from 7.7 to 1.9 per 1,000), while the prevalence rate declined 63% (from 2.6 to 1.0 per 1,000). The average number of rape and sexual assault victimizations per female victim was 2.9 in 1993 and 2.0 in 2010. The proportion of female rape and sexual assault victims reporting one incident during the year was 74% in 1993 and 85% in 2010 (table 1). ******** Robbery ******** The robbery victimization rate declined 73%, from 8.3 in 1993 to 2.2 per 1,000 in 2010 (figure 4). During the same period, the prevalence rate declined 59%, from 3.9 to 1.6 per 1,000. On average, victims of robbery experienced 2.1 victimizations in 1993, compared to 1.4 in 2010. The proportion of robbery victims who reported experiencing one robbery during the year was 91% in 1993 and 97% in 2010. Robbery victims were less likely to experience a repeat victimization during the year than rape and sexual assault and simple assault victims (table 1). ******************* Aggravated assault ******************* The aggravated assault victimization rate declined 80%, from 16.5 per 1,000 in 1993 to 3.4 in 2010 (figure 5). During the same period, the prevalence rate declined 71%, from 7.6 to 2.2 per 1,000. The average number of aggravated assault victimizations per victim declined from 2.2 in 1993 to 1.5 in 2010. In 1993, 86% of aggravated assault victims reported one aggravated assault victimization, compared to 94% in 2010 (table 1). Since aggravated assaults made up the majority of serious violent crimes, the mean number of victimizations and the percentage of victims with one aggravated assault victimization were similar to the estimates for overall serious violence. In 1993, aggravated assault victims were less likely than rape and sexual assault victims--but more likely than robbery victims--to experience a repeat victimization within the year. However, in 2010, a similar percentage of aggravated assault victims experienced repeat victimization as rape and sexual assault and robbery victims. **************** Simple assault **************** The simple assault victimization rate declined 75%, from 50.7 per 1,000 in 1993 to 12.7 per 1,000 in 2010 (figure 6). The prevalence rate declined 62% during the same period, from 18.4 to 6.9 per 1,000. Victims of simple assault experienced an average of 2.8 victimizations in 1993 and 1.8 in 2010. The percentage of simple assault victims who reported one victimization during the year was 77% in 1993 and 82% in 2010. The average number of simple assault victimizations per victim was most comparable to that found for female rape and sexual assault. This was also true for the proportion of simple assault victims who experienced more than one victimization during the year. Since rape and sexual assault had the lowest victimization rate and simple assault had the highest, the similar levels of repeat victimization among the two crime types indicates that risk for repeat victimization is not simply due to high or low crime rates. ************************* Percentage experiencing violent crime ************************* When prevalence rates are computed as a percentage of the population experiencing crime, the levels and changes in the risk for victimization for each of the types of violent crime can be compared from 1993 to 2010. For rape and sexual assault, the percentage was calculated for females age 12 or older. For the other types of violence, the percentages were based on all persons age 12 or older. The percentage of the population experiencing simple assault was consistently more than twice as high as the percentage experiencing any other type of violent crime (figure 7). In 1993, 1.8% of the population age 12 or older experienced simple assault, 0.8% experienced aggravated assault, 0.4% were victims of robbery, and 0.3% of females were victims of rape and sexual assault. In 2010, 0.7% of the population age 12 or older experienced simple assault, 0.2% experienced aggravated assault, 0.2% were victims of robbery, and 0.1% of females age 12 or older were victims of rape or sexual assault. In total, approximately 1% of the population age 12 or older experienced some form of violent victimization at least once in 2010 (not shown in table). **************************************** Violent victimization and prevalence rates for selected subgroups and types of violence during 2010 **************************************** To further illustrate how violent victimization rates and prevalence rates are distributed in the population, rates for specific subgroups and additional types of violence were examined. Comparisons between victimization rates and prevalence rates are important for understanding the level of and risk for different subtypes of violence, such as stranger violence and intimate partner violence. Although differences were found in earlier years, the violent victimization rates among males and females did not differ in 2010, when the rate for males was 20.1 per 1,000, compared to 18.5 for females (table 3). No differences were detected in the prevalence rates for violent crime between males (11.4 per 1,000) and females (10.1 per 1,000). Because there were no statistically significant differences in either the victimization rates or the prevalence rates in 2010, the average number of violent victimizations for male and female victims was also similar. Greater differences in the risk for violence were found across age groups. Persons age 35 or older had significantly lower victimization rates (12.5 per 1,000) than persons ages 12 to 34. The differences in the victimization rates between the younger age groups (i.e., persons ages 12 to 17, persons ages 18 to 24, and persons ages 25 to 34) were not statistically significant. However, prevalence rates suggest that the risk for violence for persons ages 25 to 34 (13.8 per 1,000) was lower than the risk for persons ages 18 to 24 (19.9 per 1,000). The differences across age groups in the prevalence and victimization rates and in the average number of victimizations per victim may reflect the types of violence that persons of different ages are most likely to experience. Overall, the victimization rates for the total population were significantly lower for violence committed by nonintimate family members (1.4 per 1,000) than they were for violence committed by intimate partners (3.0 per 1,000), friends and acquaintances (6.1 per 1,000), and strangers (7.1 per 1,000) (table 4). The prevalence rates of violence committed by a stranger were similar to rates of violence committed by friends and acquaintances, and higher than the rates of violence committed by an intimate partner or relative. The prevalence rates revealed that the highest rate was for violence committed by a stranger (4.7 per 1,000), while the lowest rate was for violence committed by a nonintimate relative (0.9 per 1,000). The differences between the victimization and prevalence rates across these relationship categories showed that the mean number of victimizations per victim was lower for victims of stranger violence than for victims of intimate partner violence. The data also indicated that victims of stranger violence were less likely to experience repeat victimization during the year, compared to victims who knew the offender. For example, 9% of victims of stranger violence experienced repeat victimization in 2010, compared to 21% of intimate partner violence victims who experienced repeat victimization. Even though overall violent victimization rates among males and females were about the same in 2010, males and females experienced different levels of risk based on victim-offender relationship. ***Footnote 5 See Truman, J. (2011). Criminal victimization, 2010. Washington DC: U.S. Department of Justice, Bureau of Justice Statistics, NCJ 235508***. Both victimization and prevalence rates suggest similar patterns in violence against males and females according to victim-offender relationship. Violent victimization committed by relatives and by friends and acquaintances occurred at similar rates among males and females in 2010. However, the rates of stranger violence and intimate partner violence differed between males and females. Victimization rates for stranger violence against males (9.5 per 1,000) were nearly twice as large as the rates for females (4.7 per 1,000). The prevalence rates for stranger victimization were also greater among males (6.0 per 1,000 for males, compared to 3.5 per 1,000 for females). Victimization rates for intimate partner violence against females (4.9 per 1,000) were greater than the rates against males (1.1 per 1,000), as were the prevalence rates for intimate partner violence (2.2 per 1,000 for females, compared to 0.5 per 1,000 for males). The average number of victimizations per year for male and female victims of violence committed by other family members (1.7 for males and 1.6 for females), friends and acquaintances (1.8 for males and 2.0 for females), and strangers (1.6 for males and 1.3 for females) were similar. The average number of victimizations among victims of intimate partner violence was also similar for males (2.3) and females (2.2), although the prevalence rate for females was higher than it was for males. In general, the relationship between age and victimization risk was the same across victim-offender relationship regardless of whether victimization rates or prevalence rates were used (table 5). For example, using both victimization and prevalence rates, persons age 35 or older had the lowest risk for violence by strangers, by friends and acquaintances, and by intimate partners. ***Footnote 6 Intimate partner comparisons excluded persons ages 12 to 17. The 2010 NCVS data are insufficient for producing reliable estimates of intimate partner violence rates among youth ages 12 to 17***. Victimization rates and prevalence rates both showed that persons ages 18 to 24 had the highest risk for violence by strangers, while the risk for intimate partner violence was highest among persons ages 18 to 24 and ages 25 to 34. In addition, the magnitudes of the age differences in risk were similar within victim-offender categories when based on both victimization and prevalence rates. For instance, the victimization rates suggested that persons ages 12 to 17 were 4.7 times more likely to be victimized by friends and acquaintances than were persons age 35 or older, while the prevalence rates suggested that this ratio was 4.6. In other words, victimization rates and prevalence rates revealed similar patterns in the relationship between age and victimization risk for these categorizations of violence. ***************************************** Victimization rates and prevalence rates for household property crime: 1993–2010 ***************************************** The NCVS measures household property victimization, including burglary, motor vehicle theft, and other theft. The total household property victimization rate declined 64% from 1993 to 2010, from 351.8 to 125.4 victimizations per 1,000 households (figure 8). The prevalence rate for household property crime declined 48%, from 171.6 per 1,000 in 1993 to 89.2 in 2010. The decline in the victimization rate compared to the prevalence rate resulted in a decline in the average number of victimizations per victimized household, from 2.1 victimizations per year in 1993 to 1.4 in 2010 (table 6). The differences between the prevalence and victimization rate percentage change estimates for property crime (16 percentage points) were comparable to the differences found for violent crime (13 percentage points). Similar to total violent victimization, the distribution of household victimization was skewed, and the majority of victimized households reported one victimization during the year. In 1993, approximately 75% of victimized households reported one victimization during the year, and in 2010, this proportion increased to 82%. During 2010, the 18% of households victimized by repeat property crime accounted for approximately 41% of all property crime victimizations. In comparison, in 1993 the 25% of households that experienced repeat victimization accounted for 64% of all property crime victimizations (not shown in table). ********** Burglary ********** The burglary victimization rate declined 60%, from 63.9 per 1,000 households in 1993 to 25.8 in 2010, while the prevalence rate for burglary declined 46%, from 36.2 per 1,000 households in 1993 to 19.4 in 2010 (figure 9). The average number of victimizations among victims of household burglary was 1.8 per year in 1993 and 1.3 in 2010 (table 6). A similar proportion of burglary victims experienced one incident in 1993 and 2010. In 2010, approximately 12% of burglary victims experienced a repeat burglary. These victims of repeat burglary accounted for about a third (34%) of all burglary victimizations in 2010 and 52% of all burglary victimizations in 1993 (not shown in table). ********************* Motor vehicle theft ********************* Like other household crimes, both the victimization rate (down 74%) and the prevalence rate (down 65%) showed declines in motor vehicle theft from 1993 to 2010 (figure 10). The victimization rate for motor vehicle theft declined from 19.3 to 4.9 per 1,000 households, while the prevalence rate declined from 12.1 to 4.3 per 1,000 households. The greater decline in the victimization rate compared to the prevalence rate produced a decline in the average number of victimizations per victimized household, from 1.6 to 1.2 motor vehicle thefts per year (table 6). Between 1993 and 2010, the proportion of motor vehicle theft victims who experienced only one incident in the calendar year increased slightly from 92% to 96%. Of the three types of household thefts, victims of motor vehicle theft were the least likely to experience a repeat victimization during the year. ******* Theft ******* The third type of household crime measured in the NCVS was theft, which occurred more frequently than burglary and motor vehicle theft. Like other household crimes, theft exhibited declines over time (figure 11). The victimization rate for theft declined 65%, from 268.6 per 1,000 households in 1993 to 94.6 in 2010. The prevalence rate for theft declined 49%, from 136.7 per 1,000 households in 1993 to 70.1 in 2010. The average number of thefts per victimized household decreased from 2.0 in 1993 to 1.3 in 2010 (table 7). Victims of theft were more likely than victims of other household crimes to experience a repeat victimization. Between 1993 and 2010, the proportion of theft victims who experienced one incident in the year increased slightly, from 78% to 84%. In 2010, 16% of household theft victims experienced a repeat victimization, compared to 12% of burglary victims and 4% of motor vehicle theft victims. ********* Summary ********* From a criminal justice system perspective, it is beneficial to understand the rate at which victimization occurs among a population and the prevalence of victims within that population. Victimization and prevalence rates can help detect whether any change in the criminal victimization rate resulted from a change in the number of victims or the number of victimizations per victim. These rates also help identify which groups are prone to repeat victimization and the risk associated with being victimized. Victimization rate data can be used to assess the needs of the criminal justice system as the occurrence of an event initiates certain activities (e.g., investigations, arrests, and prosecutions). Prevalence rate data may be used to estimate the number of victims requiring services in a given year. This report discusses the use of victimization rates and prevalence rates in the NCVS data, and how both help identify the level of crime and distribution of the risk for violence and household theft across the population. **************************** Violent victimization rates ***************************** When analyzing long-term trends for violent crime, examining both victimization rates and prevalence rates in the NCVS led to similar conclusions about changes in the risk for victimization over time. Both indicators showed large declines in the risk for violent victimization between 1993 and 2010; however, annual victimization rates showed that the magnitude of decline in risk for violence was greater than the decline in annual prevalence rates. For total violence, the difference in the estimate of change between victimization rates (down 76% from 1993 to 2010) and prevalence rates (down 63%) was 13 percentage points. In addition, the victimization rates for violent crime were approximately 2 to 3 times greater than the prevalence rates in 1993 and about 1.5 to 2 times greater in 2010. Examination of the prevalence rates also showed that 83% of the victims of violence experienced one victimization in 2010. The 17% of victims who experienced repeat victimization accounted for 54% of all violent victimizations that occurred that year, indicating that significant reductions in violence could potentially be achieved by preventing repeat victimization. Because the majority of victims (83%) did not experience additional violence in 2010, determining which victims are most likely to experience repeat victimization is a challenge for research and practice. ********************* Rates for subgroups ********************* The comparisons of differences by age and sex in overall violence resulted in generally similar conclusions regardless of whether victimization rates or prevalence rates were used to make the comparisons. However, the distinction between victimization and prevalence rates was most useful when describing rates of violence according to victim-offender relationship. Intimate partner violence victims were more likely to experience repeat violence within a year, compared to victims of stranger violence. Without taking into consideration the challenges associated with targeting intimate partner violence compared to stranger violence, this finding suggests that successfully targeting victims of intimate partner violence to prevent repeat victimization would result in a proportionately greater reduction in violence than targeting victims of stranger violence. The patterns also show that when the average number of incidents per victim was similar across subgroups, such as the average number of violent victimizations for male (1.8) and female (1.8) victims in 2010, these averages may hide differences in the composition of the types of victimizations experienced by each group and variation in the repeated nature of some types of violence (e.g., stranger) across the groups. Therefore, caution must be used when interpreting subgroup averages in the annual number of victimizations per victim, because these averages may be a combination of different kinds of processes and violent events. Age and sex are not the only subgroup comparisons that maybe examined, nor is the victim-offender relationship the only categorization of violence that should be considered when studying victimization and prevalence rates. Further research would need to be completed to better understand where the risk of repeat victimization is greatest. Additional assessments may reveal some subgroups or forms of violence for which there are larger distinctions between victimization rates and prevalence rates. For instance, there may be differences across race and Hispanic origin or across urban, suburban, and rural residents. Variations in the patterns may also be found if violence is categorized by location of occurrence, such as in the home, at school, or in a public location. In addition, this report examines subgroup comparisons using data from the 2010 NCVS. However, the distinction between victimization rates and prevalence rates in subgroup comparisons could be more critical during periods when overall violence rates were higher, such as the early to mid-1990s. ********************** Property crime rates ********************** As with violent crimes, long-term trends in household property crime resulted in similar conclusions regardless of whether victimization rates or prevalence rates were used. Both rates showed large declines in the risk for household property crime between 1993 and 2010, although the percentage change in annual property victimization rates (down 64%) was greater than the percentage change in annual prevalence rates (down 48%). Household crimes exhibited smaller differences between victimization rates and prevalence rates than were found among the violent crimes. In 2010, the victimization rates for household crimes were approximately 1.4 times greater than the prevalence rates, compared to approximately 1.8 times greater for violent crime rates. However, during the same year, the proportion of household property crime victims who experienced one victimization (82%) was similar to the proportion of violent crime victims who experienced one victimization (83%). This indicates that the distribution of violent victimization was more often characterized by a few high-rate repeat victims than the distribution of household property crime victimization. For household property crimes, victimizations were less concentrated. Repeat victims accounted for a lower proportion of total property crime (41%) than violent crime (54%) victimization, which suggests that the factors underlying repeat victimization are different for household crimes and violent victimizations. In addition, this finding suggests that reductions in repeat violence would have a greater impact on the total number of violent victimizations than reductions in repeat property victimization would have on the number of property crime victimizations. ************** Methodology ************** Survey coverage **************** The National Crime Victimization Survey (NCVS) is an annual data collection conducted by the U.S. Census Bureau for the Bureau of Justice Statistics (BJS). The NCVS is a self-report survey in which interviewed persons are asked about the number and characteristics of victimizations experienced during the prior 6 months. The NCVS collects information on nonfatal personal crimes (rape or sexual assault, robbery, aggravated and simple assault, and personal larceny) and household property crimes (burglary, motor vehicle theft, and other theft) both reported and not reported to police. In addition to providing annual level and change estimates on criminal victimization, the NCVS is the primary source of information on the nature of criminal victimization incidents. Survey respondents provide information about themselves (e.g., age, sex, race and Hispanic origin, marital status, education level, and income) and whether they experienced a victimization. Information is collected for each victimization incident about the offender (e.g., age, race and Hispanic origin, sex, and victim-offender relationship), characteristics of the crime (including time and place of occurrence, use of weapons, nature of injury, and economic consequences), whether the crime was reported to police, reasons the crime was or was not reported, and experiences with the criminal justice system. The NCVS is administered to persons age 12 or older from a nationally representative sample of households in the United States. The NCVS defines a household as a group of members who all reside at a sampled address. Persons are considered household members when the sampled address is their usual place of residence at the time of the interview and when they have no usual place of residence elsewhere. Once selected, households remain in the sample for 3 years, and eligible persons in these households are interviewed every 6 months either in-person or over the phone, for a total of seven interviews. Generally, all first interviews are conducted in-person. New households rotate into the sample on an ongoing basis to replace outgoing households that have been in sample for the 3-year period. The sample includes persons living in group quarters, such as dormitories, rooming houses, and religious group dwellings, and excludes persons living in military barracks and institutional settings, such as correctional or hospital facilities, and the homeless. (For more information, see the Survey Methodology for Criminal Victimization in the United States, 2008, NCJ 231173, BJS Web, May 2011.) In 2010, about 41,000 households and 73,300 individuals age 12 or older were interviewed for the NCVS. Each household was interviewed twice during the year. The response rate was 92.3% of households and 87.5% of eligible individuals. Victimizations that occurred outside of the U.S. were excluded from this report. From 1993 to 2010, 1,657 (0.8%) of the total 197,849 unweighted victimizations occurred outside of the U.S. Weighting adjustments for estimating victimization ******************************** Estimates in this report use data from the 1993 to 2010 NCVS data files. These files can be weighted to produce annual estimates of victimization for persons age 12 or older living in U.S. households. Because the NCVS relies on a sample rather than a census of the entire U.S. population, weights are designed to inflate sample point estimates to known population totals and to compensate for survey nonresponse and other aspects of the sample design. The NCVS data files include both person and household weights. Person weights provide an estimate of the population represented by each person in the sample. Household weights provide an estimate of the U.S. household population represented by each household in the sample. After proper adjustment, both household and person weights are also typically used to form the denominator in calculations of crime rates. Annual victimization estimates are derived by accumulating estimates across the two 6-month interview periods. The weights of all crimes reported during interviews in that year are summed, regardless of when the crime occurred. This method provides annual estimates of victimization, the typical measure reported from NCVS analyses. Victimization weights used in this analysis account for the number of persons present during an incident and for repeat victims of series incidents. The weight counts series incidents as the actual number of incidents reported by the victim, up to a maximum of 10 incidents. Series victimizations are similar in type but occur with such frequency that a victim is unable to recall each individual event or describe each event in detail. Survey procedures allow NCVS interviewers to identify and classify these similar victimizations as series victimizations and to collect detailed information on only the most recent incident in the series. In 2010, about 3% of all victimizations were series incidents. Weighting series incidents as the number of incidents up to a maximum of 10 incidents produces more reliable estimates of crime levels, while the cap at 10 minimizes the effect of extreme outliers on the rates. Additional information on the series enumeration is detailed in the report Methods for Counting High Frequency Repeat Victimizations in the National Crime Victimization Survey, NCJ 237308, BJS Web, April 2012. This report also measures prevalence rates of victimization. The prevalence estimate rate for a violent crime is a measure of the number of unique persons age 12 or older that experienced a violent victimization in a given year, regardless of the number or frequency of victimizations. The prevalence estimate rate for household property crime is a measure of the number of unique households that experienced a household property victimization in a given year, regardless of the number or frequency of victimizations. Person victimization weights were constructed to ensure that a person who experienced more than one nonfatal violent victimization within a 6-month period or in both 6-month periods in a given year was counted only once in that annual prevalence estimate. Similarly, household victimization weights were constructed to ensure that a household that experienced one or more property crime victimizations within a 6-month period or in both 6-month periods in a given year was counted only once in that annual prevalence estimate. Not all persons (or households) surveyed in the NCVS in a given year are interviewed twice. Because the NCVS uses a rotating panel design with 6-month recall periods, some persons were interviewed once only during the first half of the year because that was their last intended interview. Others who were interviewed only once during the first half of the year may have moved out of a sampled household, or their second intended interview for that year was not obtained for some other reason. Another group of persons were interviewed once only during the second half of the year because that was their first intended interview, because the persons recently moved into a currently sampled household, or they were missing a prior interview from the first half of the year for some other reason. The assumption underlying the annual prevalence rate estimates is that the annual rate of victimization for persons interviewed twice is similar to the rate of persons interviewed once. This assumption was tested using data from 2009 and 2010 NCVS interviews. Those tests showed that the prevalence rates of those interviewed twice were not significantly different from those interviewed once. The prevalence rates of persons interviewed only in the first half of the year were somewhat lower than the prevalence rates for persons interviewed twice, while persons interviewed only in the second half of the year had prevalence rates that were somewhat higher than those interviewed twice. The average of the rates of the two groups of persons interviewed once was not significantly different from the average of those interviewed twice. Standard error computations ***************************** When national estimates are derived from a sample, as is the case with the NCVS, caution must be taken when comparing one estimate to another estimate or when comparing estimates over time. Although one estimate may be larger than another, estimates based on a sample have some degree of sampling error. The sampling error of an estimate depends on several factors, including the amount of variation in the responses and the size of the sample. When the sampling error around the estimates is taken into consideration, the estimates that appear different may not be statistically different. One measure of the sampling error associated with an estimate is the standard error. The standard error can vary from one estimate to the next. In general, for a given metric, an estimate with a smaller standard error provides a more reliable approximation of the true value than an estimate with a larger standard error. Estimates with relatively large standard errors are associated with less precision and reliability and should be interpreted with caution. In order to generate standard errors around numbers and estimates from the NCVS, the Census Bureau produces generalized variance function (GVF) parameters for BJS. The GVFs take into account aspects of the NCVS complex sample design and represent the curve fitted to a selection of individual standard errors based on the Jackknife Repeated Replication technique. The GVF parameters were used to generate standard errors for each point estimate (such as counts, percentages, and rates) in the report. In this report, BJS conducted tests to determine whether differences in estimated numbers and percentages were statistically significant once sampling error was taken into account. Using statistical programs developed specifically for the NCVS, all comparisons in the text were tested for significance. The primary test procedure used was Student’s t- statistic, which tests the difference between two sample estimates. To ensure that the observed differences between estimates were larger than might be expected due to sampling variation, the significance level was set at the 95% confidence level. Data users can use the estimates and the standard errors of the estimates provided in this report to generate a confidence interval around the estimate as a measure of the margin of error. The following example illustrates how standard errors can be used to generate confidence intervals: According to the NCVS, in 2010, the violent victimization rate among persons age 12 or older was 19.3 per 1,000 (see table 1). Using the GVFs, it was determined that the victimization rate estimate has a standard error of 0.84 (see appendix table 2). A confidence interval around the estimate was generated by multiplying the standard errors by ±1.96 (the t-score of a normal, two-tailed distribution that excludes 2.5% at either end of the distribution). Therefore, the confidence interval around the 19.3 estimate from 2010 is equal to 19.3 ± 1.65 (.84 X 1.96) or 17.65 to 20.95. In others words, if different samples using the same procedures were taken from the U.S. population in 2010, 95% of the time the violent victimization rate would fall between 17.65 and 20.95. In this report, BJS also calculated a coefficient of variation (CV) for all estimates, representing the ratio of the standard error to the estimate. CVs provide a measure of reliability and a means to compare the precision of estimates across measures with differing levels or metrics. In cases where the CV was greater than 50%, or the unweighted sample had 10 or fewer cases, the estimate was noted with a “!” symbol (interpret data with caution; estimate based on 10 or fewer sample cases, or the coefficient of variation exceeds 50%). Methodological changes to the NCVS in 2006 *********************** Methodological changes implemented in 2006 may have affected the crime estimates for that year to such an extent that they are not comparable to estimates from other years. Evaluation of 2007 and later data from the NCVS conducted by BJS and the Census Bureau found a high degree of confidence that estimates for 2007, 2008, 2009, and 2010 are consistent with and comparable to estimates for 2005 and previous years. The reports, Criminal Victimization, 2006, NCJ 219413, December 2007; Criminal Victimization, 2007, NCJ 224390, December 2008; Criminal Victimization, 2008, NCJ 227777, September 2009; Criminal Victimization, 2009, NCJ 231327, October 2010; and Criminal Victimization, 2010, NCJ 235508, September 2011, are available on the BJS website. References ************ Forrester, D., S. Frenz, M. O’Connell, and K. Pease. (1990). The Kirkholt burglary prevention project. Home Office Crime Prevention Unit Paper 13. London: Home Office. Grove, L., G. Farrell, D. Farrington, and S. Johnson. (2012). Preventing repeat victimization: A systematic review. Stockholm: The Swedish National Council for Crime Prevention. Groves, R. and D. Cork., eds. (2008). Surveying victims: Options for conducting the National Crime Victimization Survey. Washington DC: National Research Council. Lauritsen, J. (2012). Violent repeat victimization: Prospects and challenges for research and practice. National Institute of Justice, Research for the Real World Seminar. Available at: http://www.nij.gov/events/research-real-world.htm Lauritsen, J., J. Gatewood Owens, M. Planty, M. Rand, and J. Truman. (2012). Methods for counting high-frequency repeat victimizations in the National Crime Victimization Survey, BJS Web, NCJ 237308. Truman, J. (2011). Criminal Victimization, 2010, BJS Web, NCJ 235508. **************************************************** The Bureau of Justice Statistics, located in the Office of Justice Programs, U.S. Department of Justice, collects, analyses, and disseminates statistical information on crime, criminal offenders, victims of crime, and the operation of justice systems at all levels of government. William J. Sabol is acting director. This report was written by Janet L. Lauritsen, BJS Visiting Fellow, and Maribeth L. Rezey, University of Missouri - St. Louis. Lynn Langton provided statistical review and verification of the report. Jill Thomas edited the report, and Barbara Quinn produced the report, under the supervision of Doris J. James. September 2013, NCJ 241656 **************************************************** ************************************************ Office of Justice Programs Innovation * Partnerships * Safer Neighborhoods www.ojp.usdoj.gov ************************************************ ************************ 8/21/13/JER/11:45am ************************