Study of Summary and Requirement Sent to Police Departments

Attachment 11: Study Summary and Requirements Sent to Police Departments

BJS LETTERHEAD

DATE
Police Chief Doe
4660 Main Street
Local Town, STATE, XXXX
xxxxx
Dear Police Chief Doe:
The US Bureau of Justice Statistics is currently revising the procedures used for the National Crime
Victimization Survey (NCVS). The NCVS is a national survey on that collects information on the
frequency, characteristics and consequences of criminal victimization in the United States. These
data are used to track trends in crime over time, as well as characterize the experience of victims.
We are writing you to see if [CITY] is interested in collaborating with us on a study that seeks to
improve the NCVS methodology.
As part of the study, we will attempt to conduct a victimization survey with a random sample of
households in particular areas in [CITY]. We would also like to include a sample of addresses where
the resident(s) had reported a crime to the police. Crime victimization is relatively rare in the
general population. By including addresses from which a crime has been reported, it will be possible
to test the new methodology with a significant number of persons who have experienced a
victimization.
I have included an attachment that provides a short summary of the study, including what we would
request from your department and the proposed methodology. In the next few days, someone from
Westat, a contractor that is carrying out this project for BJS, will be contacting you to discuss this
study with you. If you would like to talk with someone at BJS, please contact Michael Rand, Senior
Statistician at BJS at (202) 616-3494. If there is someone you would prefer for us to contact in lieu
of yourself, please let us know by contacting Pat Dean Brick at 301517-4196 or by e-mail at
[email protected].

Sincerely,

Michael Rand
Senior Statistician
Bureau of Justice Statistics

Study Summary
The purpose of this study is to test the feasibility of using Interactive Voice Response (IVR) for the
National Crime Victimization Survey (NCVS). IVR is a methodology that has a computer read the
survey questions over the telephone and the respondent provide answers by either using their keypad
or by speaking. IVR has the potential to reduce the costs of conducting the NCVS and could be
used as a way to follow-up households within the NCVS sample and/or as a way to conduct local
area surveys. It also has the potential to improve the measurement of sensitive crimes that are better
reported using a self-administered mode of interviewing.

The study will sample addresses in 3 cities. At each address, an adult will be randomly selected to
report about any crimes against the household (e.g., burglary, household theft, motor vehicle theft)
and against the respondent (e.g. robbery, assault). Three different methods of interviewing will be
compared: (1) IVR after a telephone interviewer contacts the household, (2) IVR after the household
is contacted by mail and (3) a telephone interview. The telephone interview condition is used to
compare the IVR to current methods being used on the NCVS. The study will evaluate the extent
respondents co-operate with the different methods and the extent they report victimizations.

The sample size for the study is 13,000 addresses across the 3 cities. Within each city, the sampled
addresses will be drawn from two different sources. One source will be a random sample of
addresses taken from a listing provided by the United States Postal Service. The second source will
be a sample of addresses taken from records of crimes reported to the police. The use of the police
records will enable the study to fully test the different methodologies with a significant number of
individuals who report a victimization.

Information Requirements for Police Records
Types of Crime to be covered by police records. Study staff will work with the police
department to determine the mix of crimes that would be included in the sample. Pending
discussions with the departments, the study would like to sample both violent and property crimes.
It would exclude: (1) Crimes involving domestic disputes, (2) Sexual violence and (3) Any crimes
involving ongoing investigations.

Time Frame to be covered by police records. Crimes occurring within 12-months from the
time the respondent is interviewed. This timing is important ensure the survey will be asking
about crimes that occurred during the prior 12-month period.
Information provided in each crime record. For each crime record, the study would need the
address of the person that reported it, the type of crime and victim demographics (if personal crime)
such as age and gender. We will not ask for the name of the victim or the person that reported the
crime to the police.

Study Procedures
Method of contact. Sampled addresses will be sent a letter which explains the study. This letter
will provide basic information about the study and the conditions related to participation. If the
initial contact is by telephone, an interviewer will call the household who is a randomly select adult.
If the initial contact is by mail, the notification letter will ask that the adult with the next birthday call
the IVR system to take the survey.1

Informed consent. In the introductory material, the respondent will be told that:
1. The study is being conducted for the US Bureau of Justice Statistics.
2. The study is voluntary.
3. The study obtained the sample from two sources: (a) a random sample of addresses in the
community and (b) police records of individuals reporting a crime to the police.
4. The study will not collect the name of the respondent.
5. The study will maintain the confidentiality of the data. With respect to
confidentiality, all data collected by BJS are protected by federal law 42 USC 3789g.

1 This is the standard method for a respondent to be selected in telephone surveys.