Conclusion
Examining serious violent offending through data gathered from victims provides a picture that is both similar to and different from that emerging from police arrest data. Not surprisingly, the volume of offending is much greater in the NCVS data than is indicated in the UCR arrest data. Depending on whether adult co-offending is excluded or included, offending rate estimates from the NCVS data are 9 to 15 times greater than the arrest rate estimate from the UCR data. Both sources of data show race and gender disproportionality in serious violent offending. Male juveniles offend at much higher rates than female juveniles. Black juveniles offend at much greater rates than white juveniles. The amount of gender disproportionality, however, is greater in the NCVS data than the UCR data, while the racial disproportionality is greater in the UCR data than the NCVS data. In other words, victims indicate that there are more male offenders (relative to female offenders) and more white offenders (relative to black offenders) than appear in police records.
|
Adjusting NCVS Data As noted in the text, the methodology employed in the National Crime Victimization Survey (NCVS) was changed radically in 1992. As a result, data from the pre-1992 period, when the survey was known as the National Crime Surveys (NCS), must be adjusted to make them comparable to the more recent data. An adjustment of the data is possible because the old NCS and new (NCVS) designs were run in parallel for 18 months. Data from this overlap period can be used to compute the ratio of estimates from the post-1992 and pre-1992 designs. This ratio, in turn, can be used to adjust all the pre-1992 data to make them comparable to the post-1992 data (Kindermann, Lynch, and Cantor, 1997; Rand, Lynch, and Cantor, 1997; Lynch and Cantor, 1996). The Bureau of Justice Statistics (BJS) has produced a ratio adjustment for major crime classes. BJS cautions, however, that this adjustment factor may not be suitable for adjusting trends for subpopulations (such as juveniles) or specific estimates other than victimization rates. Consequently, adjustment factors for offender trends have been specifically developed for the analysis presented in this Bulletin. These adjustment factors were produced using the same methodology employed by Lynch and Cantor (1996) to determine the adjustments BJS used for gross victimization trends. Specifically, offending rates were estimated using the NCS and the NCVS data from the 1992 overlap. Significance tests were performed on the differences between the rate estimates from the old and new designs. When significant differences were found, further tests were employed for specific subpopulations. If, for example, the rates from the NCVS were significantly different from the NCS for both males and females, the significance of this difference was tested to determine whether specific ratio adjustments should be applied for each gender. Whenever significant differences between the designs were found, adjustment ratios were produced and applied to each subgroup. If these differences were not significant, then the same adjustment ratio was applied to all subgroups. The following are the adjustment ratios used in this Bulletin for offending rates overall, by gender, and by race. Overall. The adjustment ratios for overall juvenile offending including adult co-offending are as follows: rape, not significant (ns); robbery, ns; aggravated assault, ns; simple assault, 2.02; total violence, 1.58. The adjustment ratios for juvenile offending excluding adult co-offending are as follows: rape, 1.65; robbery, ns; aggravated assault, 0.91; simple assault, 1.74; total violence, 1.29. By gender. The adjustment ratios for male juvenile offending including adult co-offending are as follows: rape, 3.05; robbery, ns; aggravated assault, ns; simple assault, 1.42; total violence, 1.41. The adjustment ratios for male juvenile offending excluding adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, ns; simple assault, 1.69; total violence, 1.68. The adjustment ratios for female juvenile offending including adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, ns; simple assault, 1.42; total violence, 1.41. The adjustment ratios for female juvenile offending excluding adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, ns; simple assault, 1.69; total violence, 1.68. By race. The adjustment ratios for white juvenile offending including adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 1.48; simple assault, 1.79; total violence, 1.59. The adjustment ratios for white juvenile offending excluding adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 1.90; simple assault, 1.68; total violence, 1.53. The adjustment ratios for black juvenile offending including adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 0.63; simple assault, 1.99; total violence, 1.11. The adjustment ratios for black juvenile offending excluding adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 0.59; simple assault, 2.06; total violence, 1.53. The adjustment ratios for other-race juvenile offending including adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 2.29; simple assault, 3.75; total violence, 2.26. The adjustment ratios for other-race juvenile offending excluding adult co-offending are as follows: rape, ns; robbery, ns; aggravated assault, 1.0; simple assault, 3.55; total violence, 3.18. Although this adjustment procedure is useful, it must be used with caution. The 18-month overlap period does not provide many data for estimating adjustment ratios. The NCVS sample was not expanded during this period. Half the usual NCVS sample completed the pre-1992 instrument and half the post-1992 instrument. As a result, the reliability of the estimates from the overlap period is less than that of the NCVS generally. This means that the adjustment ratios are likely to be unstable for smaller population subgroups and rare offenses. Moreover, the accuracy of the adjustment ratios is likely to decrease the farther back in time one goes from the overlap period. For these reasons, this Bulletin presents trends only back to 1980, and these trends are limited to large classes of violent crime and major population groups. More detailed trends were produced for the post-1992 period (see "Trends in Offending Rates Since 1993"), when a consistent design was used for the full NCVS sample. |
The trends in serious violent offending for juveniles are broadly similar in the two data sources. Both show decreases in serious violent offending until about 1986–87 and increases thereafter until around 1994, when sharp decreases began and continued until 1998. Both race and gender disproportionality declined between 1980 and 1998 in both the UCR and NCVS data, as the overrepresentation of male offenders and the overrepresentation of black offenders decreased.
This examination of offending using NCVS data confirms the general picture of offending derived from UCR arrest data. At the same time, the greater racial disproportionality and the lower gender disproportionality found in the arrest data raise some questions. It is tempting to infer that police policy is the likely cause of these differences, specifically, that the police tend to arrest black offenders more often than white offenders or female offenders more often than male offenders. However, recent research (Pope and Snyder, in press) finds no support for these contentions, at least as they apply to serious violent crimes.
Other explanations must be considered. One factor may be differential reporting of crimes by victims. For example, if victims of crime by black offenders are more likely to report their victimization to the police, then the arrest rate for black offenders will be higher even if the police treat every case about equally. Other factors may be differences in the scope, definitions, and procedures used in police record systems and victim surveys. The UCR arrest rates, for example, include arrests for events that are not asked about in the NCVS, such as commercial robbery, which in the UCR arrest data is not separated from noncommercial robbery and thus cannot be removed from the dataset to allow equal comparison with the NCVS data. If these events involve black offenders more than white offenders, then this could account for the greater racial disproportionality in the arrest data relative to the offending data. Similarly, the fact that larger urban places (which have disproportionately large black populations) report more reliably than smaller places in the UCR may lead to the overrepresentation of black juveniles in the UCR arrest data. If analysts adjusted the UCR arrest data to take the missing information into account more systematically, the differences between the police and victim survey data in race and gender disproportionality might disappear.
All of these explanations—and others— must be pursued to understand the apparently discrepant pictures of offenders emerging from the UCR and the NCVS. The good news is that most of these alternative explanations for the observed differences can be examined with the UCR and NCVS data. Although beyond the scope of this analysis, it is possible, for example, to develop and apply adjustments in the UCR to account for the overrepresentation of large population places. Then this explanation for black overrepresentation in the UCR data would be eliminated.
Finally, this examination of victim data offers a view of co-offending that is not often obtained (or obtainable) from police data. The majority of juvenile violent offending is done in groups. About one-third of this group offending involves adults. Male juveniles are more disposed to group offending than female juveniles, and black juveniles much more than white juveniles. The declines in the juvenile offending rates for serious violent crimes between 1994 and 1998 were greatest for juvenile co-offending, especially for blacks. It is unclear why the drop in juvenile co-offending was so steep for blacks compared with whites.9 Possible explanations for this dramatic decrease in violent co-offending among black juveniles should be explored empirically to enhance the understanding of the recent drop in overall crime, with particular attention to the disparate declines in lone and group offending.
| Previous | Contents | Next |
 |
|
| Trends in Juvenile Violent Offending: An Analysis of Victim Survey Data | |