AnalysisofZeroArmedPerpetrators(ZAP)Dataset

1、Zero Armed Perpetrators:An Examination and Analysis of Program DataLynn K. Harvey, Ph.D.Associate Professor of SociologyDepartment of Social Sciences and Center for Community SafetyWinston-Salem State UniversitySeptember 2005The Center for Community Safety at Winston-Salem State University works in

2、partnership with the Center for Youth, Family, and Community Partnerships at The University of North Carolina at Greensboro as the Project Safe Neighborhoods Research Partner for the United States Attorneys Office, Middle District of North Carolina. These efforts are supported by PSN funding (Award

3、#2002-GP-CX-0220) through the U.S. Department of Justice, Office of Justice Programs. The opinions, findings, and conclusions or recommendations expressed in this document are those of the author and do not necessarily reflect the views of the Department of Justice.Zero Armed Perpetrators:An Examina

4、tion and Analysis of Program DataBackgroundZero Armed Perpetrators (ZAP) is a program initiated by the Forsyth County District Attorney in July 2000 to reduce gun violence in Winston-Salem and Forsyth County. The program is designed to review all firearm cases that come to the attention of the DA

5、9;s office, even those in which a gun is not seized; a report of a threat to use a gun is sufficient to trigger a review. ZAP seeks to identify the most dangerous offenders and develop prosecution strategies that can lead to the removal of those individuals from the streets of the city and the count

6、y.ZAP is coordinated by an individual within the DA's office, with help from an assistant. A meeting to review all recent gun cases is held weekly. Invited to the meeting are representatives of state and federal prosecutors' offices; city, county, state and federal law enforcement agencies,

7、including ATF and ICE; state adult and juvenile probation offices; the city/county school system; the county attorneys office; and a university-based community outreach center. Participants share information about each case and the individuals involved, and a decision is made about what charges to p

8、ursue and whether the case should be taken to state or federal court.Data related to cases reviewed in ZAP meetings are entered into a Microsoft Excel dataset. Included in the dataset are the following fields:· Defendants name, race, sex, date of birth, Social Security number, Pistol number (Wi

9、nston-Salem Police Department database identifier)· Whether the defendant is a convicted felon· Incident Report number (from law enforcement agency)· Charges against the defendant· Docket number· Law enforcement agency involved in the case· Law enforcement officers name

10、· Assistant DA assigned to the case· Date the incident was reviewed by ZAP· Whether the case was taken to federal court; if so, the federal agents name· Whether the case involved domestic violence· Whether guns, ammunition, and/or casings were seized· Type of gun involv

11、ed· Disposition of the court case· Sentence given the guilty partyA copy of the dataset was made available in the Spring of 2005, for two purposes: analysis of the existing data and suggestions about ways to capture data regarding future ZAP cases. This report responds to those requests.Pr

12、oceduresThe structure of the dataset was examined to determine its implications for analysis using SPSS software. Several elements of the structure, while useful for ZAP purposes through visual inspection and Excel capabilities, presented difficulties for SPSS analysis. For example, each charge agai

13、nst a defendant is listed on a separate row of the spreadsheet, each with the defendants name and demographic/identifying information, with rows alphabetized by defendants last name. While this is useful for such prosecutorial concerns as tracking repeat offending by an individual and separating tho

14、se with single charges from those with multiple charges, the demographic characteristics of defendants with multiple charges would skew the overall demographic pattern of defendants in the dataset as a whole, if analyzed using SPSS. To avoid this type of problem, the dataset for analysis was restruc

15、tured to reflect separate incidents, rather than defendant names. Thus, if one incident report involved multiple charges against a defendant, that defendants name and demographic/identifying characteristics were removed from all but one of the records (all but one of the rows with information about

16、that incident). If the same defendant faced additional charges as a result of other incidents, his/her name and characteristics were retained in only one row for each additional incident. Where an incident report number was unavailable, docket numbers were used to determine the number of separate in

17、cidents involved (not a perfect way to make such a determination, but the method used for this incident-based analysis).In 142 rows of the ZAP dataset, the Defendant field does not include a name. The first column, instead, shows such entries as 50-B Seizure, Found Property, Suicide, Police Service,

18、 and Safekeeping. To maintain a focus on incidents with a defendant, these rows were eliminated from the dataset used for this analysis. All other incidents were retained, even when no charges were shown.The original dataset also contains some rows with only defendant name, demographic information,

19、and identification numbers. When these descriptive characteristics were not included for the defendant on rows with incident data, the information was transferred to appropriate rows, so that records of incident data would be complete. The rows with only defendant name and descriptive characteristic

20、s were then eliminated from the dataset used for analysis.Several variables (fields) were added to the dataset used for analysis. First, age of the defendant at the time the dataset was delivered was calculated for each defendant, based on information in the Date of Birth field. This method of calcu

21、lating age is somewhat misleading i.e., it inflates the defendants age, compared to his/her age at the time of the incident, especially since some of the cases occurred several years prior to the delivery of the data. This estimation method was used because the dataset includes no information on the

22、 date of the incident, and the date the case was reviewed in a ZAP meeting is missing for many of the cases. Despite the limitations, however, the age calculation does provide some estimate of the overall distribution of defendant ages, especially since ages were categorized (under 16, 16-17, 18-24,

23、 25-34, 35-44, etc.) in the analysis.Second, the many different charges faced by defendants were categorized into felonies, misdemeanors, and others. Third, the many types of firearms involved in these incidents were categorized as handguns, rifles, shotguns, or others. The added dataset column with

24、 this categorization includes multiple types of weapons (e.g., handgun and rifle) in any incident, where that information was available. In some cases the original dataset showed only “multiple” firearms, and that designation was retained in the new categorization. The final dataset used for analysi

25、s contains data on cases involving 4,420 charges stemming from 2009 different incidents. The analysis focuses on demographic characteristics of defendants in these incidents and whether they were convicted felons, types of charges against them, whether guns were seized, the types of firearms involve

26、d, whether the case involved a domestic violence charge, whether the case was selected for prosecution at the federal level, and the disposition of the court case.ResultsTable 1 presents characteristics of defendants involved in the 2009 incidents. Nine out of ten defendants were males; 62% were bla

27、ck, 26% white, and 11% Hispanic; and three in ten were convicted felons at the time their cases were reviewed by the ZAP team. When the data were delivered for this study, over two-thirds of the defendants were 18-34 years of age, and their average age was 30. Twenty-five percent were under the age

28、of 22, half under 27, and three-fourths under 36.Table 1. Defendant characteristics.SexFemale9.3%Male.90.7%Race/EthnicityAsian. 0.1%Black.62.3%Hispanic11.2%Native American. 0.0%White.26.3%AgeUnder 16 2.3%16-17. 3.6%18-24.35.3%25-34.33.0%35-44.14.9%45-54. 6.5%55-64. 3.1%65 or over. 1.3%FelonNo.70.2%Y

29、es29.8%Characteristics of incidents and charges are shown in Table 2. Almost half (46%) of the charges that could be classified as felonies or misdemeanors fell into the more serious category. At least one gun was seized in 58% of the incidents; 4.6% of incidents were known to have involved domestic

30、 violence. Although not shown in Table 2, the Winston-Salem Police Department (WSPD) was listed as the law enforcement agency in about three-fourths of the incidents.Table 2. Characteristics of Incidents and ChargesGun Seized in IncidentNo.41.9%Yes58.1%Domestic Violence IncidentNo.95.4%Yes 4.6%Charg

31、esFelony.46.1%Misdemeanor53.9%As shown in Table 3, handguns were by far the most likely type of firearm to be involved in these incidents. Seventy-two percent of cases where a gun was identified involved at least one handgun, either alone or with another type of firearm. Only a rifle was involved in

32、 6.3% of the incidents, while only a shotgun was noted in 5.2%. Other types of firearms included BB guns and pellet guns, as well as many other types. In 3.4% of the cases examined, the dataset indicated only “multiple” firearms, which may have included any of the above types.Table 3. Types of Firea

33、rms InvolvedHandgun only.67.4%Handgun and Rifle. 2.7%Handgun and Shotgun 1.5%Handgun and Other Type of Firearm. 0.3%Handgun, Rifle, and Shotgun. 0.2%Rifle only 6.3%Shotgun only. 5.2%Rifle and Shotgun. 0.9%Other Type of Firearm only.12.1%Multiple Unspecified Firearms. 3.4%The vast majority of cases d

34、iscussed by the ZAP team (96.5%) were prosecuted at the state level (Table 4). Only 2.6% of the incidents in the dataset were designated for federal prosecution alone, and one incident was noted as leading to prosecution in both state and federal courts. Sixteen incidents were to be revisited by the

35、 team for a decision at a later date.Table 4. Prosecution DecisionState96.5%Federal 2.6%Both 0.0%Revisit. 0.8%Felony charges were associated with higher rates of federal prosecution than were misdemeanors (1.9% of all felony charges vs. 0.6% of all misdemeanors), and federal prosecution was more lik

36、ely when the defendant was already a convicted felon (4.7% of felons vs. 0.7% of others). All defendants designated for federal prosecution were 18-54 years of age. Eliminating the three defendants in the dataset who were either Asian or Native American allows for a significance test of the associat

37、ion between race/ethnicity (for blacks, whites, and Hispanics) and federal prosecution. Doing so demonstrates a significant association: 1.3% of white defendants, 2.5% of black defendants, and 5.3% of Hispanic defendants were designated for federal prosecution. This finding is likely affected by the

38、 fact that it is illegal for an undocumented immigrant to possess a firearm.A number of these variables are related to one another. For example, 62.2% of felons compared with 43.6% percent of non-felons were charged with a felony in these cases. About 57% of felons compared with about 22% of non-fel

39、ons were involved in cases where a gun was seized. Among black defendants in these cases, 38.7% were felons, compared with 18.7% of whites and 5.8% of Hispanics. About 54% of black defendants were charged with felonies in these cases, compared with 37% of Hispanics and 33% of whites. Hispanic defend

40、ants were most likely (65.5%) to be involved in cases where guns were seized, compared with whites (61.2%) and blacks (54.4%). A gun seizure was not related to whether a defendant was charged with a felony or a misdemeanor: 46.5% of cases in which a gun was seized resulted in felony charges, compare

41、d with 46.0% of cases in which a gun was not seized.Dispositions related to 2482 charges are included in the dataset (see Table 5). A guilty verdict was recorded for 1438 (57.9%) of those, including 62 where the dataset notes that the defendant was found guilty of a lesser charge. Voluntary dismissa

42、ls accounted for 31.5% of the dispositions. Only 4.5% resulted in a not guilty verdict.Table 5. Disposition of ChargesGuilty57.9%Not Guilty. 4.5%Prayer for Judgment Continued 2.6%Voluntary Dismissal.31.5%Other. 3.5%Felony charges, compared with misdemeanors, were more likely to result in a guilty ve

43、rdict (73.5% vs. 46.8%) and less likely to be voluntarily dismissed (24.2% vs. 37.0%). More likely to be dismissed were cases in which there was no gun seizure (32.8% vs. 27.7%) and cases in which some type of gun other than a handgun, rifle or shotgun was seized (most often a BB or pellet gun) (33.

44、3% vs. 26.6%). Guilty verdicts declined steadily with age, from 73% of defendants 16-17 years old to 36.8% of those 55-64; however, 58.8% of the 17 defendants 65 years of age of older were found guilty. None of the other variables of interest in this analysis was significantly associated with a guil

45、ty verdict.Conclusions and RecommendationsAny potential restructuring of the ZAP dataset should be based on a careful consideration of the uses for which it is intended. If the type of analysis in this report is useful, then the kind of dataset restructuring done for this study would be advisable. H

46、owever, the value of such a change must be balanced against the usefulness of the current dataset structure for the ZAP program and other purposes. Judgment on these issues should be made by individuals intimately involved with the program. At the least, it does seem advisable to add fields not cont

47、ained in the current dataset. For example, incident date, the felony-misdemeanor classification, and the categorization of types of firearms appear to be important. Such data could be much more efficiently included at the time of initial data entry than at a later date. All such additions, of course

48、, increase the amount of time required for data entry, so their value, too, must be weighed in the context of time and budget constraints. Some data entry errors are to be expected, but careful consistency in data codes would increase efficiency in conducting analyses. For example, the original data

49、 set used “Y” in some cases and “X” in others as an indicator of a federal case or a domestic violence incident, and several different entries were clearly intended to represent the Winston-Salem Police Department.The overrepresentation of males in this analysis is consistent with arrest figures in

50、the FBIs Uniform Crime Report for 2003: 90.7% of defendants in the ZAP cases were male, compared with 91.8% of those arrested nationally for weapons (not just firearms) violations. Age breakdowns are also similar: the percentage of arrestees 18-24 years of age was 35.3% in the ZAP dataset and 35.7%

51、for weapons violations in the 2003 UCR. The under 18 percentage is larger nationally (23.3% vs. 5.9%) and the 25-34 percentage is larger locally (33% vs. 20.8%), probably due to ZAPs focus on individuals to be prosecuted as adults and the way the ages were calculated in this analysis. African Americ

52、ans are highly overrepresented as defendants in ZAP cases (62.1%), compared with their proportion in the Forsyth County population, according to the 2000 Census (25.9%), and compared with 2003 UCR arrests for weapons violations (36.1%). Whites are underrepresented in local statistics (26.3%), compared with their local population prevalence (69.3%) and national arrest figures for weapons violations (62.1%).Felonies constitute a large proportion (46%) of charges reviewed in ZAP meetings, and guns are seized in almost 60% of the incidents, suggesting that ZAP is an important vehicle for focusing