Crime victimization survey : 2010 report.

              Oklahoma State Bureau of Investigation
Information Services Division
Data Collection and Reporting Section
Oklahoma Statistical Analysis Center Crime Victimization Survey 2010 Report
August 2011
Introduction
The Oklahoma Statistical Analysis Center (SAC) recently completed the first statewide crime victimization survey. The purpose of the Oklahoma Crime Victimization Survey (OCVS) was to generate statewide crime estimates for crimes in Oklahoma. Since surveying an entire population is not practical, researchers developed estimates based on a sample of the population; estimates calculated from this survey were based on a sample of 200 adults. Next year, researchers plan to conduct another victimization survey with a sample of 1,000 adults.
The results presented below provide valuable insight into the nature of crime in Oklahoma. Additionally, researchers gained a better understanding of Oklahomans’ perceptions of crime and their opinions of the criminal justice system. Overall, Oklahomans feel safe in their community and are satisfied with the efforts of law enforcement. According to estimates, most adults in Oklahoma support capital punishment and believe criminal justice funding should be centered on education and prevention.
Highlights:
 Oklahomans experienced an estimated 500,000 violent and property crimes in 2010;
 61.9% of victimizations were reported to law enforcement;
 Survey estimates of burglary rates were higher than national rates in 2010;
 Oklahomans over 35 were more likely to be burglarized than those under 35;
 Over 50% of vandalisms involved a motor vehicle;
 Approximately $1 billion dollars worth of stolen or damaged property was attributable to property crimes;
 For those crimes with a known suspect, over 75% of victims knew the perpetrator;
 Almost 50% of all crimes occurred in Oklahoma and Tulsa counties;
 Half of all crimes occurred during summer months;
 59% of victims were female;
 77% of known offenders were male;
 Overall, Oklahomans feel safe in their communities;
 Oklahomans are satisfied with law enforcement’s efforts;
 More than half of adults in Oklahoma own a firearm;
 15% of adults in Oklahoma carry a firearm outside their home for protection;
 The majority of Oklahomans (86.9%) support capital punishment;
 Over one-fourth of Oklahomans think criminal justice funding should be directed at prevention and education;
 35% of Oklahomans believe drugs and alcohol are most responsible for crime in their community;
 Over 80% of Oklahomans have taken steps to prevent victimization.
Oklahoma Crime Victimization Survey 2010
2 Oklahoma Statistical Analysis Center
Crime Victimization, 2010
In 2010, adults in Oklahoma experienced an estimated one-half million violent and property victimizations (table 1), including both reported and unreported crimes. An estimated 96,000 crimes were violent (including simple assaults), and an estimated 468,000 were property crimes. In comparison, according to the annual Crime in Oklahoma report, victims reported 18,394 violent and 130,965 property crimes to law enforcement in 2009 (see tables 4 and 5 for crime reporting estimates).
This discrepancy between reported crime, which is captured by traditional summary reporting (UCR), and unreported crime, which is captured by victimization surveys, is a common occurrence for a number of reasons. Many victims do not report the crime to law enforcement. Summary reporting is organized by the hierarchy rule, which means only the most serious crime committed during the crime incident is reported.
Oklahoma crime rate estimates were higher than national estimates reported in the 2009 National Crime Victimization Survey (See Criminal Victimization, 2009, NCJ 231327, Bureau of Justice Statistics (BJS) website, October 2010). However, only burglary rates were statistically significant at the 95% confidence level; this increase in crime is consistent with trends in reported crimes. In 2009, Oklahoma’s reported burglary rate was 10.3, compared to 7.2 reported nationally, and Oklahoma’s reported larceny/theft rate was 22.6, compared to 20.6 reported nationally (See Crime in the U.S., 2009, FBI website, September 2010). No associations in this survey between victimization rates and victim characteristics were found to be statistically significant at the 95% confidence level.
At the 90% confidence level, survey results support two statistically significant associations between victimization rates and victim characteristics:
 Age was associated with burglary rates, with those 35 years of age or older having higher rates of burglary than those younger than 35;
 Residency tenure was associated with vandalism rates, with those who had lived at their residence for 11 years or less experiencing higher rates of vandalism than those who had lived at their residence for longer than 11 years.
Incident Characteristics
In 2010, 46.7% of all incidents in Oklahoma occurred in Oklahoma and Tulsa Counties; in fact, 59.1% occurred in the Oklahoma City
and Tulsa metropolitan areas (defined by county based on population density and on proximity to an urban county – appendix figure 1 and appendix table 1).
Half of all incidents (49.5%) occurred in the summer, with nearly one in four incidents (23.9%) occurring in July alone. Peaks also occurred in April (10.3%) and December (10.1%). Most crimes (54.3%) occurred at night (6:00 pm to 6:00 am) with one-third (32.6%) occurring between 9:00 pm and midnight; 31.1% of crimes were daytime crimes (6:00 am to 6:00 pm).
Nearly half of all crimes (44.4%) occurred “near the victim’s home,” while 32.8% happened “inside the victim’s home.” More than one in ten crimes (11.6%) occurred “in a commercial place.” The majority of victims (72%) did not know if the offender was carrying a weapon during the incident. The offender carried a knife in 3.7% of known incidents, and in another 2.5% of known incidents the offender carried a blunt object.
Table 1. Criminal victimization in Oklahoma, 2010 Crime Type Number of Victimizations Ratea All crimesb 564,254 ~ Violent crimes 96,248 34.1 Property crimesc 468,006 320.5 Motor vehicle theft 50,361^ 34.5 Burglary 139,948 95.8 Larceny/theft 277,697 190.1 Vandalism 119,352 81.7 Total population age 18 or older was 2,821,685 Total number of households was 1,450,460 ~Not applicable. ^Estimate is based on five or fewer sample cases a rates are per 1,000 persons age 18 or older or per 1,000 households b simple assault, motor vehicle theft, burglary, and larceny/theft c motor vehicle theft, burglary, and larceny/theft Oklahoma Crime Victimization Survey 2010
3 Oklahoma Statistical Analysis Center
Violent Crimes
In most cases, the victim was not attacked by the offender during the incident. However, 5.1% of victims reported they were “hit, slapped, or knocked down”; 3.7% were “grabbed, held, tripped, or pushed”; and 2.1% reported that the offender attempted to attack them with a knife or sharp weapon; tried to “hit, slap, or grab them”; threatened them with a weapon; or verbally threatened them with some other attack (respondents listed multiple offenses – all based on fewer than five incidents). After reviewing all reported attacks (including attempted attacks and threats), researchers determined that 9.4% of reported victimizations met the legal definition for assault in Oklahoma.
Property Crimes
In 2010, 24.4% of incidents involved forcible entry or attempted forcible entry of a victim’s home, and 44.0% of incidents involved forcible entry or attempted forcible entry of a victim’s vehicle. In 88.4% of reported property incidents, the offender stole or took items without the victim’s permission.
According to respondents, commonly stolen items included: “personal effects such as clothing or jewelry” (29.0%); “tools, machines, or office equipment” (21.5%); “TVs, DVD players, or other household appliances” (16.5%); “gasoline or oil” (16.1%); “cash” (11.7%); “motor vehicles” (11.4%); “bicycles or parts” (7.7%); and “handguns (pistol or revolver) or other firearm (rifle or shotgun)”(3.6%).
An estimated 52.2% of vandalisms involved damage to the victim’s motor vehicle; another 18.0% involved damage to a “house, window, screen, or door”; and 16.9% involved damage to “trees, shrubs, or fences” located in the victim’s yard or garden. One in four acts of vandalism involved “broken glass” (28.3%), and more than one in ten (10.1%) involved “defacing, such as marring or painting with graffiti.”
Crime Costs
Oklahomans lost an estimated one billion dollars in stolen and damaged property in 2010. Twenty-five percent of all victimizations involved damage to the victim’s property. On average, Oklahomans lost $1,053.00 worth of stolen property per incident of motor vehicle theft, burglary, or larceny/theft.
Victims
One in four (23.8%) Oklahomans reported they were the victim of at least one incident of assault, motor vehicle theft, burglary, or larceny/theft in 2010. Approximately one-third of victims (31.9%)
lived below the poverty line. An estimated 37% were unemployed, and almost half (46.7%) received a high school education or less (see table 3).
Over half of victims (51.1%) lived in urban areas. Over one-third (36.2%) of victims reported that they have lived at their current residence for less than five years.
Victims were predominantly:
 Female (59.0%);
 Married (59.3%);
 At least 40 years old (61.7%);
 White (79.8%); and
 Non-Hispanic (97.8%).
There were no statistically significant differences in demographic characteristics between victims and non-victims, likely due to the study’s limited sample size.
Offenders
There was only one offender in 38% of incidents, two offenders in 21.1% of incidents, and three or more offenders in 3.9% of incidents. In 36.9% of incidents, the victim did not know the number of offenders.
Seventy-seven percent of offenders were male. Also, victims described the offender as “white” in 58.3% of incidents; 17.5% of offenders were described as “black” by the victim. An estimated one in three offenders were 18 to 20 years old, 28.5% were 21 to 29, and 21.7% were 30 or older. According to respondents, 5.8% of offenders were juveniles. In 26.4% of cases, the victim believed the offender was under the influence of drugs or alcohol during the incident. Table 2. Property Loss Values, 2010 Type of loss Dollar Value All types 1,036,365,490 Stolen property 868,186,250 Damaged property* 97,961,293 Vandalized property 70,217,947 Total number of households was 1,450,460 *Includes property damaged in connection with a motor vehicle theft, burglary, or larceny/theft Oklahoma Crime Victimization Survey 2010
4 Oklahoma Statistical Analysis Center
An estimated three out of four victims (75.4%) reported knowing the offender, 12.9% of victims said the offender was a “casual acquaintance,” 20.2% knew the offender “by sight only,” and 42.4% knew the offender “well.” The offender was a relative or in a domestic relationship with the victim in 5.5% of reported incidents (See 21 O.S. § 644 (c)); in 13.8% of incidents the offender was a “friend, ex-friend, or neighbor,” and in another 43.0% of incidents, the offender was a “non-relative who had been previously known in some other way.” One-fourth of offenders (24.6%) were “strangers” to the victims.
Crime Reporting
In 2010, the majority of victims (61.9%) reported the crime to a law enforcement agency (table 4). Overall (both reported and unreported), law enforcement completed a report for 47.3% of crimes.
For reported crimes, victims personally notified the police 77% of the time. In 68% of reported crimes, the police went to the crime scene. The victim went to the local law enforcement agency in another 16.2% of reported crimes. In 15.7% of reported crimes, the victim did not have face-to-face contact with law enforcement.
Once dispatched, the officer’s response time was “less than an hour” in 80% of reported crimes, including a response time of “less than five minutes” in 8.3% of crimes. Officers responded “within a day” for 11.3% of crimes, while in 2.8% of crimes the response time was “longer than a day.” Approximately 6% of victims did not remember how long it took the police to respond.
Once they arrived, officers completed a report 100% of the time. Police questioned witnesses or offenders 59.0% of the time and collected evidence (fingerprints or inventory) 28.1% of the time. During the first response, police officers arrested an offender one percent of the time. Overall, victims reported that an offender(s) had been arrested 10.4% of the time. In another 11.9% of crimes the victim did not know if an arrest had been made
Unreported Crimes
Victims cited several reasons for not reporting crimes to law enforcement; most common was the belief that police would be “inefficient or ineffective” (24.1% of unreported crimes). Approximately 21% of victims considered the crime a “private or personal matter.” Another 20.9% of victims reported they “took care of it [themselves].” The victim did not report the crime in 14.5% of incidents because they considered it to “be a minor or unsuccessful crime resulting in small or no loss”; 11.2% of crimes went unreported because of “lack of proof or inability to identify the stolen property or the offender”; and 10.6% went unreported because the victim “did not want the offender to get in trouble with the law.” In 8.6% of incidents the victim did not report the crime to law enforcement because it had been reported to another official, such as a guard or apartment manager. Table 3. Crime Reported to Police, by Type, 2010 Type of crime Percent Reported 95% Wilson Score Interval All crimes 61.9% 47.4 - 75.0% Simple assault 54.5 14.3 - 91.0 Motor vehicle theft 100.0^ 21.2 - 100.0 Burglary 88.8 62.8 - 100.0 Larceny/theft 37.7 20.5 - 57.4 ^Estimate is based on five or fewer sample cases Table 4. Police Reports Filed, by Type, 2010 Type of crime Percentage resulting in a police report 95% Wilson Score Interval All crimes 47.3% 32.9 - 61.9% Simple assault 25.3 0.0 - 74.6 Motor vehicle theft 100.0^ 21.2 - 100.0 Burglary 77.4 48.7 - 95.2 Larceny/theft 20.6 7.5 - 39.9 ^Estimate is based on five or fewer sample cases Oklahoma Crime Victimization Survey 2010
5 Oklahoma Statistical Analysis Center
Table 5. Characteristics of victims of crime in Oklahoma, 2010
Characteristic
Estimate
95% Wilson Score Interval
Characteristic
Estimate
95% Wilson Score Interval
Sex
Poverty line
Male
41.0%
24.7 - 58.7%
Above
56.9%
37.9 - 74.5%
Female
59.0
40.0 - 76.4
Below
31.9
17.2 - 49.5
Race
Refused
11.3
2.5 - 27.4
White
79.8%
63.2 - 91.5%
Employment
Black
10.2
1.9 - 26.1
Employed
63.5%
45.7 - 79.0%
American Indian
8.3
1.0 - 23.6
Unemployed
36.5
19.9 - 55.6
Other/multiple
1.7
0.0 - 13.6
Education
Ethnicity
No schooling
18.7%
7.5 - 35.1%
Hispanic
2.2%
0.0 - 13.1%
High School or GED
27.9
13.3 - 46.8
Non-Hispanic
97.8
85.5 - 100.0
Vocational or trade school
0.8
0.0 - 12.1
Age
Some college/associate degree
23.8
10.3 - 42.4
18-29
18.0%
7.0 - 34.3%
Four-year college degree
25.9
11.7 - 44.6
30-39
19.6
7.4 - 37.6
Master’s degree or higher
2.8
0.0 - 15.4
40-49
26.9
12.5 - 45.6
Residence type
50-59
16.7
5.6 - 34.2
Urban
51.1%
33.8 - 68.2%
60 or older
18.2
6.5 - 35.9
Rural
48.9
30.5 - 67.5
Refused
0.7
0.0 - 11.7
Tenure (Years at Residence)
Marital status
0-1 years
11.5%
3.1 - 26.4%
Married
59.3%
41.5 - 75.5%
2 years
11.1
2.4 - 27.2
Widowed
13.3
3.5 - 30.0
3-4 years
13.6
3.7 - 30.4
Separated
2.4
0.0 - 14.8
5-6 years
14.0
4.0 - 30.9
Divorced
6.4
0.2 - 20.9
7-15 years
23.6
10.1 - 42.1
Never married
18.6
6.8 - 36.5
Longer than 15 years
26.2
12.0 - 45.0
Oklahoma Crime Victimization Survey 2010
6 Oklahoma Statistical Analysis Center
Perceptions of Crime
For this survey, researchers also asked respondents about their opinions and perceptions of crime. Most respondents reported feeling safe in their community. Overall, respondents were satisfied with the work of local law enforcement. The majority of respondents reported crime was not a problem in their community. They also supported the death penalty and expressed a negative view of offenders.
Perception of Safety
Most Oklahomans feel safe in their community (96.5%). Less than one percent feel “very unsafe” in their community (table 6).
Table 6. How safe do you feel in your community? Very safe 54.6% Somewhat safe 41.9 Somewhat unsafe 2.5 Very unsafe 0.9 --Rounds to less than 0.5 percent
At the 90% confidence level-
 A higher percentage of respondents who were victims of crime in 2010 (62.5%) said they felt “somewhat safe” in their communities compared to non-victims (36.5%), and a higher percentage of non-victims (60.1%) said they felt “very safe” in their communities compared to victims (35.8%);
 A higher percentage of female respondents (51.8%) said they felt “somewhat safe” compared to male respondents (32.2%), and a higher percentage of male respondents (65.3%) said they felt “very safe” compared to female respondents (44.8%).
Perception of Law Enforcement
The majority (89.0%) of Oklahomans rated local law enforcement’s performance as “very good” (47.9%) or “somewhat good” (41.1%), while 4.6% rated their performance as “very bad” (table 7).
At the 90% confidence level –
 Married respondents (53.3%) reported that the performance of law enforcement agencies in their community are “very good,” when compared to unmarried respondents (33.4%);
 Crime Victims (22.9%) were less than half as likely as non-victims (53.9%) to rate law enforcement’s performance as “very good.”
Perception of Crime in Community
Half of those surveyed (53.5%) feel crime is “not a problem” (19.2%) or “a slight problem” (34.3%) in their communities, while almost one in five (16.5%) feel crime is “a serious problem” (table 8). No statistically significant differences were identified for this factor.
Table 8. To what degree is crime a problem in your community? Not a problem 19.2% A slight problem 34.3 A moderate problem 29.1 A serious problem 16.5 Don’t Know 0.9
Capital Punishment
The majority of Oklahomans support capital punishment. Of those surveyed, 63.8% reported that they “strongly support” and 23.1% reported that they “somewhat support” capital punishment; 3.5% of Oklahomans are “strongly against” capital punishment (table 9).
Table 7. How would you rate law enforcement in your community? Very Good 47.9% Somewhat Good 41.1 Somewhat Bad 3.2 Very Bad 4.6 Don’t Know 2.9 Refused To Answer -- --Rounds to less than 0.5 percent Oklahoma Crime Victimization Survey 2010
7 Oklahoma Statistical Analysis Center
Table 9. To what degree do you support capital punishment? Strongly support 63.8% Somewhat support 23.1 Somewhat against 3.9 Strongly against 3.5 Don’t know 4.2 Refused to answer 1.4 Note: Detail does not sum to 100 percent due to rounding.
 White respondents (29.0%) were three times as likely as non-white respondents (9.7%) to “somewhat support” capital punishment (95% confidence level).
Released Offenders
Oklahomans are divided on how they feel about released offenders in Oklahoma. One-fourth (27.3%) feel most ex-offenders are “dangerous or untrustworthy”; one in five (22.4%) feel most are “lazy, a drain on society”; one in five (21.2%) feel most are “normal – no different from anyone else”; and one in seven (13.5%) feel most ex-offenders are “reformed, contributing members of society.” A large percentage of Oklahomans (15.2%) were unsure which of these categories best describes their opinion of ex-offenders (table 10). Table 10. Which of the following statements best describes your opinion of released offenders? Dangerous or untrustworthy 27.3% Lazy, “drain on society” 22.4 Normal, no different than others 21.2 Reformed, contributing members 13.5 Don’t know 15.2 Refused to answer -- --Rounds to less than 0.5 percent
At the 95% confidence level –
 Male respondents (31.3%) were more than twice as likely as female respondents (12.7%) to consider most ex-offenders “normal – no different from anyone else”;
 Employed respondents (27.2%) were more than twice as likely as unemployed respondents (10.3%) to consider most ex-offenders “normal – no different from anyone else.”
At the 90% confidence level –
 Respondents who received post-high school education (35.9%) were twice as likely as those with a high school education or less (18.6%) to consider most ex-offenders “dangerous or untrustworthy.”
Criminal Justice Funding
Respondents were also divided on how criminal justice funds should be allocated in Oklahoma. One-fourth (27.3%) of Oklahomans think criminal justice funds should be allocated to “education and prevention,” 23.5% think “arrest, prosecution, and imprisonment of offenders” should receive the most money and support. One in eight (12.3%) think “treatment and rehabilitation,” and 9.8% think criminal justice resources “should be spent on something else” (table 11). Table 11. Which of the following do you feel should receive the most money and support in the fight against crime? Education and prevention 27.3% Adjudication process 23.5 Stopping the flow of illegal drugs 23.5 Treatment and rehabilitation 12.3 Something else 9.8 Don’t know 3.6
At the 95% confidence level:
Male respondents (35.9%) were almost three times as likely as females respondents (12.3%) to think the most money and support should go to “stopping the flow of illegal drugs into the state/country”;
At the 90% confidence level:
White respondents (26.7%) were more than twice as likely as non-white respondents (10.1%) to think the most money and support should go to “stopping the flow of illegal drugs into the state/country.” Oklahoma Crime Victimization Survey 2010
8 Oklahoma Statistical Analysis Center
Causes of Crime
Over a third (35.1%) of those surveyed think that “drugs, alcohol, or other substance abuse” is most responsible for crime in their community; another third (32.1%) attribute crime to the “decline of family values and discipline.” One in ten (11.0%) Oklahomans think “the economy” is most responsible for crime in their community (table 12). Table 12. Which of the following do you feel most contribute to crime in Oklahoma? Drug, alcohol, or other substance abuse 35.1% Decline of family values and discipline 32.1 The economy 11.0 Street gangs 3.6 Improper sentencing of offenders 3.0 Ineffective anti-gang/drug education 1.4 Lack of adequate law enforcement 1.4 Violent movies, TV, video games 0.9 Guns 0.8 Something else 4.6 Don’t know 5.7 Refused -- --Rounds to less than 0.5 percent
 Married respondents (40.6%) were twice as likely as unmarried respondents (21.8%) to feel the “decline of family values and discipline” was most responsible for crime in their community;
 Unmarried respondents (13.0%) were nearly six times as likely as married respondents (2.3%) to feel “something else” was responsible for crime in their community;
 Female respondents (13.3%) were more than seven times as likely as male respondents (1.8%) to say they didn’t know what was responsible for crime in their community.
Gun Ownership
More than half (54.7%) of adults in Oklahoman keep a gun or guns in their home (table 13). Of these, 73.2% keep a gun or guns “for protection or sporting purposes,” while 15.5% keep a gun or guns “for protection” alone. Nearly one in seven (14.7%) Oklahomans carried a gun outside their home for protection in 2010 (table 14).
Table 14. Which one of the following reasons best describes why you have gun(s) in your home? For protection and sporting purposes 73.2% For protection 15.5 For sporting 8.4 For another reason 2.9
At the 95% confidence level –
 Married respondents (68.1%) were twice as likely as unmarried respondents (34.4%) to have a gun(s) in their home;
 White respondents (58.0%) were twice as likely as non-white respondents (27.4%) to have a gun(s) in their home;
 Those who had lived at their current residence for longer than five years (58.3%) were more than twice as likely as those who had lived at their current residence for five years or less (25.1%) to have a gun(s) in their home; Table 13. Gun ownership Question Yes No Don’t Know Refused Do you keep a gun or guns in your home? 54.7% 42.6 -- 2.3 In the last 12 months, did you carry a gun for protection? 14.7 84.9 -- -- - Rounds to less than 0.5 percent. Oklahoma Crime Victimization Survey 2010
9 Oklahoma Statistical Analysis Center
 Those above the poverty line (66.1%) were more than twice as likely as those below the poverty line (26.6%) to have a gun(s) in their home.
At the 95% confidence level -
 Male respondents (27.3%) were more than three times as likely as female respondents (7.9%) to report carrying a gun outside their home for protection in 2010;
 Respondents who had lived at their current residence for longer than five years (19.6%) were more than three times as likely as respondents who had lived at their current residence for five years or less (5.6%) to report carrying a gun outside their home for protection in 2010.
At the 90% confidence level –
 Respondents whose income and household size put them above the poverty line (23.9%) were three times as likely as respondents below the poverty line (7.8%) to report carrying a gun outside their home for protection.
Crime Prevention Steps
Most Oklahomans have taken deliberate steps to prevent becoming the victim of a crime. More than one in three Oklahomans (38.1%) “installed more secure door or window locks”, while a fourth (29.1%) “purchased a dog”; a fourth (27.3%) “purchased a gun”; a fourth (26.9%) “installed burglar alarms”; and a fourth (24.5%) “installed security lights.” One in ten Oklahomans (10.7%) “purchased pepper spray”, and one in thirteen (7.7%) “carried [an] object for defense.” Nearly one in five (18.4%) Oklahomans have “taken no action” to feel safer from crime (table 15).
At the 95% confidence level –
 Respondents over age 30 (30.9%) were three times as likely as those under 30 (10.4%) to have installed burglar alarms to feel safer from crime;
 Respondents above the poverty line (36.4%) were more than three times as likely as those below the poverty line (11.1%) to have installed burglar alarms;
 Respondents who had lived at their current residence for longer than five years (30.5%) were more than three times as likely as those who had lived at their residence for five years or less (8.9%) to have installed burglar alarms;
 Married respondents (40.4%) were nearly three times as likely as unmarried respondents (14.4%) to have purchased a gun to feel safer;
 Respondents above the poverty line (37.5%) were four times as likely as those below the poverty line (9.5%) to have purchased a gun(s);
 Respondents who had lived at their current residence for longer than five years (33.5%) were more than five times as likely as those who had lived at their current residence for five years or less (6.2%) to have purchased a gun;
 Respondents living in rural communities (41.8%) were twice as likely as those in urban communities (18.9%) to have purchased a dog to feel safer. Table 15. Which of the following steps have you taken to feel safer from crime? Installed more secure door/window locks 38.1% Purchased dog 29.1 Purchased gun 27.3 Installed burglar alarms 26.9 Installed security lights 24.5 Displayed security sticker 13.8 Purchased pepper spray 10.7 Involved in neighborhood watch 10.5 Completed a self-defense course 9.3 Carried an object for defense 7.7 Something else 18.4 Took no action 8.5 Respondents could select more than one action
Oklahoma Crime Victimization Survey 2010
10 Oklahoma Statistical Analysis Center
At the 90% confidence level –
 Female respondents (35.5%) were twice as likely as male respondents (18.2%) to have purchased a dog;
 Respondents with education beyond high school (35.4%) were twice as likely as those with a high school education or less (16.2%) to have installed burglar alarms;
 Respondents over age 30 (15.3%) were three times as likely as those under 30 (4.6%) to have joined a neighborhood watch.
Victim Services Knowledge
Most Oklahomans (87.4%) reported they knew where to go in their community or neighborhood when they need help or services as victims of crime; however, nearly one in eight (12.6%) did not know where to go for help. No statistically significant differences were identified between subsets of the population for this factor.
Conclusion
In 2010, SAC staff conducted the first statewide crime victimization survey. The purpose of the survey was to gain a better understanding of unreported crime in Oklahoma. In the future, researchers plan to conduct another survey with a larger sample size.
Respondents answered questions about their experiences with crime in 2010. Offenders committed an estimated one-half million violent and property crimes in Oklahoma last year. Half of all crimes occurred during the summer months. Survey results indicate that more than half of all crimes occurred at night, and 75% of crimes occurred near or in the victim’s home. Oklahomans lost an estimated one billion dollars in stolen and damaged property in 2010.
According to respondents, over 60% of crimes were reported to law enforcement in 2010. Seventy-seven percent of offenders were male. Victims believed that the offender was under the influence of alcohol or illegal drugs during 24% of crimes. Seventy-five percent of victims knew the offender, and approximately six percent of crimes were related to domestic violence.
Researchers asked respondents about their perceptions of the criminal justice system in Oklahoma. The majority of Oklahomans are satisfied with the work of local law enforcement in their community. Half of those surveyed did not feel crime was a significant problem in their community. The majority of Oklahomans support the death penalty. Respondents were divided in their opinions of where criminal justice funds should be concentrated: one-quarter of Oklahomans believe education should be a priority; another quarter believed the adjudication process should receive more support. More than half of survey respondents reported that they keep a gun(s) in their home; 15% reported they carried a gun(s) outside their home for protection in 2010.
Methodology
Oklahoma residents age 18 and older were eligible to participate in the survey, provided they had an Oklahoma telephone number, and given they had the ability, physical or otherwise, to answer the telephone and complete a survey when called. Excluded groups included armed forces personnel living in military barracks, correctional facility inmates, and other similar populations. However, the results are intended to be generalized to the entire population of Oklahoma.
Questionnaire Design
This study used victimization questions from the National Crime Victimization Survey (NCVS) questionnaires, with some slight modifications to facilitate a telephone interview. The close adherence to NCVS questions allows the results of this study to be comparable to national estimates of crime victimization.
Topic areas identified as of interest to law enforcement personnel in the Oklahoma Law Enforcement Needs Assessment administered in 2010 determined the other questions included in the study. Questions asked in victimization surveys conducted in other states provided the basis for some of these questions (See Maine Crime Victimization Report, 2007 and 2006 Crime Victimization Study: A Survey of Utah Residents).
Sample Design and Size
This study utilized a stratified sample design by area code. Random digit dialing (RDD) was used, with 1,616 numbers generated for area code 405, 820 numbers for area code 580, and 1,564 for area code 918, proportionate to the combined 2009 Census Bureau population estimates for the counties in each area code (for counties covering more than one area code, the dominant area code was used, based on centers of population). The 4,000 generated phone numbers (with unused exchanges for each area code excluded) were randomized and called until the desired sample size was obtained. When generating the sample, no distinction was made between landline phones and cell phones, though cell phone numbers were later identified during data Oklahoma Crime Victimization Survey 2010
11 Oklahoma Statistical Analysis Center
analysis (appendix table 2). Based on prefix designations, 34.7% of generated phone numbers were determined to be for cell phones. All cell phone numbers were dialed by hand, in compliance with federal requirements.
For an outcome where the probability of success is equal to that of failure (50/50), a margin of error of 5% requires a sample size of 384, and a margin of error of 10% requires a sample size of 96. Given that a margin of error no greater than 10% was desired for the statewide victimization estimates and that some inflation of the standard errors was expected due to sample weighting, a sample size of 200 was selected. The final margin of error for percentages reported in the study (except where otherwise reported) is 8.3%.
Review of Human Subjects Research
The sensitive nature of the survey questions in this study necessitated oversight by the OSBI Institutional Review Board (IRB). The five member Board reviewed the proposed survey questionnaire and methodology. IRB members determined there would be no more than minimal risk to study participants and approved the project.
Data Collection
Following IRB approval, SAC personnel at OSBI Headquarters in Oklahoma City conducted telephone interviews during February 4 through March 31, 2011. Typically, calls were made on weekday mornings and evenings, and interviews lasted between approximately 10 and 20 minutes. Interviewers entered interview responses into a Microsoft Access database created for the OKCVS and noted dispositions for each phone number. SAC staff calculated the response rate for the survey (appendix table 2) from the noted final dispositions in accordance with American Association for Public Opinion Research (AAPOR) standards (See Computation of Response Rates below). Interviewers called phone numbers in blocks of 50, with all 50 numbers in each block tried a minimum of five times at varying times of day (at least twice in the morning and at least twice in the evening).
Computation of Response Rates
The computation of response rates for this study was performed in accordance with AAPOR guidelines using RR3 (See Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys, AAPOR website, revised 2011). Completed interviews were defined as those interviews where respondents answered all questions pertaining to victimization in 2010. The proportion of eligible numbers among all numbers where the eligibility status was known determined what proportion of numbers of unknown eligibility would be considered eligible, with landline phones and cell phones treated separately. The overall response rate was calculated as the number of completed interviews (as defined above) divided by the number of eligible phone numbers (with the projected numbers of eligible landline and cell phone numbers of unknown eligibility included).
Estimation Procedure
In order to estimate statewide rates of victimization, responses were assigned two weights – one for personal crimes and one for property crimes. Weights were applied using the inverse of respondents’ probability of selection, inflating the responses so the survey participants represent themselves and portions of the overall state population. Weights were also applied to account for respondents’ age, race, and sex. Selection probability and demographic distribution were based on 2009 Census Bureau general estimates and 2010 Census county population estimates (See American Community Survey, 2009, and 2010 Census tables, Census website) in eight geographic regions of Oklahoma, based on population density, proximity to urban counties, and geographic area of the state (figure 1 and table 1). For personal crimes, a weight for the number of adults in a household was also factored into the overall weighting to account for those in the household who were not interviewed. Raking (or sample balancing) was performed to combine the five weights for personal crimes and four weights for property crimes experienced by study participants using a Statistical Analysis System (SAS version 9.1.3) macro (See To Rake or Not To Rake Is Not the Question Anymore with the Enhanced Raking Macro, ABT Associates website, 2009).
Series Victimizations
Series victimizations are defined as six or more similar, but separate, crimes that the victim is unable to recall individually or describe in detail to the interviewer (See Survey Methodology for Criminal Victimization in the United States, BJS website, 2007). The National Crime Victimization Survey has found such victimizations tend to inflate victimization counts when included with non-series victimizations. Series victimizations are therefore routinely excluded from NCVS tables and are counted as a single incident in the generation of NCVS estimates. In order to maximize the use of collected data while addressing the potential for inflated estimates, the same practice is followed in this study, with each series of victimization counted as a single incident and included Oklahoma Crime Victimization Survey 2010
12 Oklahoma Statistical Analysis Center
with the non-series victimizations in the generation of population estimates.
Accuracy of Estimates
A certain amount of sampling and random error occurs in all surveys. For this survey, a maximum unweighted standard error of about 7% was selected, meaning the true percentage of Oklahomans being victimized may vary from 7% below to 7% above the unweighted estimates. Sample weighting introduces additional variation accounted for by incorporating the design effect into the standard errors (See Computation and Application of Standard Errors below).
Nonsampling error in the estimates results from a number of factors. Recall bias occurs when respondents have difficulty remembering the exact details of an event of interest, such as the date of victimization. As a result, some of the crimes reported in this study may not have occurred during 2010. Certain crimes, such as crimes involving violence or sexual abuse, are reported less frequently than other crimes, resulting in underestimates of the true prevalence of these crimes. Also, the decision of some individuals not to participate in the survey may make the estimates less generalizable to the overall population.
Despite these limitations, faced by all victimization surveys, the estimates generated in this study provide measures of victimization rates in Oklahoma that may be compared with other states and the nation. These estimates are the product of extensive efforts by SAC staff to maximize the quality and accuracy of the data collected.
Computation and Application of Standard Errors
Standard errors for the victimization estimates were calculated using the Taylor series expansion method in SAS (appendix table 3). For estimates of proportions of subsets of the sample, 95% confidence intervals were calculated using the Wilson Score Interval due to the small sample sizes and often extreme estimates.
In this study, data were weighted in order to ensure the characteristics of the sample closely matched those of the overall population of Oklahoma. Such weighting of the data generally results in inflation of the standard errors, which is called design effect. There are a number of ways to calculate design effects that result from weighting. For this study, the design effect was calculated using the formula
DEFF = n*Σ(wi)2/(Σwi)2,
where n is the number of respondents in the stratum being considered and wi is the weight assigned to the ith respondent. The design effect was multiplied by the sample variance in the calculation of standard errors not generated in SAS (specifically, the Wilson Score Intervals). This is equivalent to multiplying the margin of error of a confidence interval by the square root of the design effect (the design factor).
Definitions of Population Subsets
Subsets of the sample (by age, sex, education, minority status, poverty status, marital status, employment status, urban or rural residency, and years at current residence) were checked for statistical differences in victimization rates and crime perceptions responses.
Age was dichotomized by whether a respondent was younger than age 30 or 30 years of age or older. Education was dichotomized by whether a respondent had a high school education or less or if they had some higher education. Marital status was dichotomized by whether a respondent was married or unmarried. Urban or rural residency was based on the Rural-Urban Commuting Area (RUCA) codes developed by the Economic Research Service of the U.S. Department of Agriculture. Zip codes with a “metropolitan” RUCA code were considered urban, while zip codes with “micropolitan,” “small town,” or “non-core” RUCA codes were considered rural. Tenure was dichotomized by whether or not a respondent had lived at their residence for longer than five years.
Oklahoma Crime Victimization Survey 2010
13 Oklahoma Statistical Analysis Center
Figure 1. Eight regions of Oklahoma, defined by population density, proximity to urban (population density of 1,000 persons per square mile or more) counties, and geographic location, 2010
Oklahoma Crime Victimization Survey 2010
14 Oklahoma Statistical Analysis Center
Appendix Table 1. Urban, suburban, and rural regions of Oklahoma and probability of selection
Regiona
Counties
Typeb
Probabilityc
1
Oklahoma
Urban
0.192
2
Tulsa
Urban
0.161
3
Canadian, Cleveland, and Pottawatomie
Suburban
0.117
4
Creek, Muskogee, Payne, Rogers, Wagoner, and Washington
Suburban
0.114
5
Adair, Cherokee, Craig, Delaware, Lincoln, Mayes, Nowata, McIntosh, Okfuskee, Okmulgee, Osage, Ottawa, Pawnee, and Sequoyah
Rural
0.113
6
Atoka, Bryan, Choctaw, Coal, Haskell, Hughes, Johnston, Latimer, Le Flore, Marshall, McCurtain, Pittsburg, Pontotoc, Pushmataha, and Seminole
Rural
0.092
7
Beckham, Caddo, Carter, Comanche, Cotton, Garvin, Grady, Greer, Harmon, Jackson, Jefferson, Kiowa, Love, McClain, Murray, Stephens, Tillman, and Washita
Rural
0.129
8
Alfalfa, Beaver, Blaine, Cimarron, Custer, Dewey, Ellis, Garfield, Grant, Harper, Kay, Kingfisher, Logan, Major, Noble, Roger Mills, Texas, Woods, and Woodward
Rural
0.082
a See appendix figure 1.
b Counties with population density of 1,000 persons per square mile or more were considered urban, counties with population density between 70 and 1,000 persons per square mile were considered suburban when adjacent to an urban county or another suburban county, and counties with population density of 70 persons per square mile or less were considered rural. Groupings were made by geographic location
c Based on 2010 Census county population counts (See 2010 Census tables, Census website) Oklahoma Crime Victimization Survey 2010
15 Oklahoma Statistical Analysis Center
Appendix Table 2. Sample disposition (N=4,000)
Total numbers dialed
Disposition
Landline
Cell
Interview
93
109
Completea
92
108
Complete (interview record lost)
1
1
Eligible, non-interview
102
173
Refusal or break-off
80
131
Callback (never completed)
22
42
Unknown eligibility, non-interview
439
261
Always busy
150
24
No answer
68
5
Telephone answering device
65
96
Call-blocking or no voicemail
27
55
Unattempted
129
81
Not eligible
1978
845
Out of sampleb
13
24
Fax/data line
58
1
Non-working/disconnected
1709
771
Nonresidence
191
39
Language barrier
7
10
Response rate (RR3)c
34.8%
aAll questions on victimization in 2010 completed
bDid not live in Oklahoma in 2010 or child cell phone
cNumber of complete interviews divided by number of eligible numbers plus 8.97% of landlines and 25.0% of cell phones of unknown eligibility based on rates of eligibility in known data (See Computation of Response Rates above)
Appendix Table 3. Standard errors of victimization estimates
Estimate
Standard error of the numbers
Standard error of the rates
All crimesa
94,739
~
Violent crimes
39,715
14.1
Property crimesb
86,013
58.9
Motor vehicle theft
36,216
24.8
Household burglary
47,051
32.2
Larceny/theft
68,185
46.7
Vandalism
31,800
21.8
All types of loss
380,842,340
~
Stolen property
369,397,747
~
Damaged propertyc
86,751,368
~
Vandalized property
32,563,680
~
Note: Standard errors were calculated using the Taylor series expansion method in SAS version 9.1.3. These standard errors are for estimates in tables 1 and 2, and they should be multiplied by 1.971956 to form 95% confidence intervals
~Not applicable
aIncludes simple assault, motor vehicle theft, burglary, and larceny/theft
bIncludes motor vehicle theft, burglary, and larceny/theft
cIncludes property damaged in connection with a motor vehicle theft, burglary, or larceny/theft. Oklahoma Crime Victimization Survey 2010
16 Oklahoma Statistical Analysis Center
Appendix Table 4. Demographic characteristics of study participants (N = 200)
Characteristic
Count
%
Characteristic
Count
%
Race
Sex
White
160
80.0
Male
105
52.5
Black
17
8.5
Female
95
47.5
American Indian
14
7.0
Children in Household
Asian
3
1.5
None
115
57.5
Two or more races
6
3.0
One
43
21.5
Ethnicity
Two
26
13.0
Hispanic
5
2.5
Three or more
15
7.5
Non-Hispanic
195
97.5
Refused
1
0.5
Age
Adults in Household
18-29
41
20.5
One
43
21.5
30-39
36
18.0
Two
120
60.0
40-49
41
20.5
Three
27
13.5
50-59
29
14.5
Four or more
9
4.5
60-69
24
12.0
Refused
1
0.5
70-79
19
9.5
Household Income
80 or older
7
3.5
Less than $10,000
21
10.5
Refused
3
1.5
$10,000-20,000
21
10.5
Marital Status
$20,000-30,000
22
11.0
Married
112
56.0
$30,000-40,000
19
9.5
Widowed
12
6.0
$40,000-50,000
18
9.0
Divorced
25
12.5
$50,000-75,000
27
13.5
Separated
2
1.0
$75,000 or more
41
20.5
Never married
45
22.5
Don’t know
10
5.0
Refused
4
2.0
Refused
21
10.5 Oklahoma Crime Victimization Survey 2010
17 Oklahoma Statistical Analysis Center
Appendix Table 4 (continued). Demographic characteristics of OCVS 2010 study participants (N = 200)
Characteristic
Count
%
Characteristic
Count
%
Education
Years At Current Residence
Less than High School
16
8.0
Less than five
88
44.0
High School or GED
69
34.5
Five or more
111
55.5
Vocational or trade school
4
2.0
Refused
1
0.5
Some college or associate degree
51
25.5
Employment
Four-year college degree
46
23.0
Law Enforcement or Security
4
2.0
Master’s degree
11
5.5
Medical Profession
15
7.5
Doctorate, law, or medical degree
2
1.0
Mental Health Services
3
1.5
Refused
1
0.5
Retail Sales
12
6.0
Region of Residence*
Teaching Profession
14
7.0
Oklahoma County
33
16.5
Transportation
7
3.5
Tulsa County
40
20.0
Something else
80
40.0
Oklahoma City Suburbs
35
17.5
Unemployed
64
32.0
Tulsa Suburbs
22
11.0
Missing data
1
0.5
Northeast Oklahoma
25
12.5
Employer
Southeast Oklahoma
15
7.5
Private company or business
85
42.5
Northwest Oklahoma
9
4.5
College or university
5
2.5
Southwest Oklahoma
21
10.5
Federal government
6
3.0
Residence Type
State or local government
19
9.5
Urban/Metropolitan
120
60.0
Self-employed
18
9.0
Rural/Micropolitan
36
18.0
Unemployed
64
32.0
Rural/Small town
22
11.0
Don’t know
2
1.0
Rural/Noncore
22
11.0
Missing data
1
0.5
Note: Detail may not sum to 100% because of rounding.
*Regions defined by county based on population density, proximity to an urban county (population density of 1,000 persons per square mile or more), and geographic location (appendix figure 1 and appendix table 1). Oklahoma Crime Victimization Survey 2010
18 Oklahoma Statistical Analysis Center
This project was supported by Grant No. 2010-30177-OK-BJ awarded by the Bureau of Justice Statistics. The Bureau of Justice Statistics is a component of the Office of Justice Programs, which also includes the Bureau of Justice Assistance, the National Institute of Justice, the Office of Juvenile Justice and Delinquency Prevention, and the Office for Victims of Crime. Points of view or opinions in this document are those of the author and do not represent the official position or policies of the United States Department of Justice.
Furthermore, the resources and personnel required to complete this project were provided by the Oklahoma State Bureau of Investigation. Points of view or opinions in this document are those of the author and do not represent the official position or policies of the Oklahoma State Bureau of Investigation.
The Oklahoma State Bureau of Investigation is recognized by the Bureau of Justice Statistics as the Statistical Analysis Center (SAC). The SAC collects, analyzes, and disseminates justice information; these functions are located within OSBI’s Office of Criminal Justice Statistics.
The following individuals wrote this report, under the direction of David Page, Information Services Division Director; Linda DeArman, Administrative Programs Officer; and Angie Baker, SAC Director:
Carl Grafe, Statistical Research Specialist
For more information, please contact:
Oklahoma State Bureau of Investigation
6600 North Harvey
Oklahoma City, Oklahoma 73116-7912
(405) 848-6724
http://www.ok.gov/osbi/