"Victimization data supplement UCR data by providing information about crimes not reported to police, but they are still not able to capture the entire dark figure of crime. Victimization surveys are, of course, not able to capture offenses where there is no identifiable victim, such as public order offenses like drug offenses, gambling, disorderly conduct, trespassing, public drunkenness, and prostitution. Other excluded crimes include murder, bank robbery, and nonresidential economic crimes such as tax evasion, nonresidential burglary, possession of stolen property, and employee theft. Thus, since a large majority of offenses involve public order and nonresidential economic crime, victimization surveys are able to provide information about only a small subset of crimes, primarily violent and sex crimes, personal theft, motor vehicle theft, and residential burglary. Mosher, Miethe, and Phillips (2002) caution that using victimization data has four inherent problems: 1. Victimization surveys cover only a small range of crimes. 2. Victimization surveys are based on sample data rather than population counts, subjecting them to distortion from sampling error and sampling bias. 3. Victimization surveys are based on victims’ perceptions without independent confirmation. 4. Question wording and technical elements of the survey, including the use of different procedures over time, make it difficult to compare victimization rates over time. These problems do not diminish the value of victimization data, given that official UR data have their own flaws. Victimization data are important in minimizing the dark figure of crime and necessary to understanding criminal behavior. A first step in studying any form of criminal behavior should involve consulting both UCR and victimization data (and any other available data) for crime categories in which data are available."