RAND Supporting Statement B rev 4-24-2013 - Clean Version

Interviews with State Officials and Experts to Inform Cost Estimates of Implementing Policy and Behavioral Interventions to Reduce Motor Vehicle Injuries

Supporting Statement – Section B

Submitted by:

Department of Health and Human Services

Center for Disease Control and Prevention

National Center for Injury Prevention and Control

Division of Unintentional Injury Prevention

4770 Buford Highway, NE, Mailstop F62
Atlanta, GA  30341-3717

Program Official/Project Officer

Angela Salazar

Public Health Advisor

Phone: 770.488.3949

Fax: 770.488.1317

Email: aos9@cdc.gov

4-26-2013

Section B – Data Collection Procedures

1. Respondent Universe and Sampling Methods

Based on the information that we need to collect, we have stratified our respondent universe into six categories: public safety advocates, DWI/ DUI defense attorneys, Court case managers, State parole agency staff, State department of public safety staff, and law enforcement agency personnel. We are conducting qualitative interviews and only need a few respondents to represent each stratum.

1. Public safety advocates – 4 respondents

2. DWI/ DUI defense attorneys – 4 respondents

3. Court case managers – 4 respondents

4. State parole agency staff – 2 respondents

5. State department of public safety staff – 6 respondents

6. Law enforcement agency personnel – 4 respondents

The information compiled on the states that have implemented interventions are reflected in the sources listed in (Section 12) of the Supporting Statement A.

Information has been collected by RAND through the implementation of online searches and literature review (i.e. through public sources which does not require a burden to anyone) about which interventions are currently used in each state, and many of these data sources include information about the date of implementation. (A list of these data sources is provided in Supporting Statement Part A, Section 12.)

We do not expect a 100% response rate and expect to sample with replacement. The expectation is to use snowball sampling as necessary. If a selected respondent declines to participate in the interview, they will be asked to identify one or two other individuals that they believe would be appropriate respondents based on their knowledge and expertise of the topic. Using this technique, we anticipate a response rate of 80 percent.

2. Procedures for the Collection of Information

Because non-probability sampling methods are being used, issues such as statistical methods for stratification and sample selection, estimation procedure, and degree of accuracy, etc., are not relevant.

The twenty-four qualitative subject matter expert telephone interviews will be conducted by RAND researchers, Dr. Andres Villaveces and Liisa Ecola. These two seasoned researchers are well-versed and experienced in conducting interviews with subject matter experts.

We have identified five states to target for interviews: California, New Mexico, Maryland, Oregon, and Tennessee. Our criteria for selection included the number of relevant laws in the state and the desire to get a diverse distribution of states in terms of geography, total population, and population groups. For each state, we will begin by identifying public officials such as parole agency staff, public safety staff, and law enforcement staff via online searches both on state web sites as well as other sources (e.g., conference programs). We will ask them to provide suggestions of knowledgeable experts in the other topic areas. RAND’s Safety and Justice program has extensive contacts in law enforcement and we will enlist their assistance in identifying specific types of personnel to target. These searches will be done by Dr. Andres Villaveces and Liisa Ecola. Attachment 12 provides the draft email invitation that will be sent to prospective respondents. This email invitation will also include a listing of the questions to be covered during the interview.

For the online Delphi panel, we will select 8 experts to participate based on our knowledge of the persons with the required expertise. These persons will be employed by academia or a public agency (e.g. Centers for Disease Control and Prevention or the National Highway Traffic Safety Administration. We will provide the panel with the questions for which we have not obtained data from the state interviews (see Attachments 11 and 13).

3. Methods to Maximize Response Rates / Deal with Nonresponse

The data collection plan assumes that we will conduct the number of interviews within each category of subject matter expert as indicated in section 1. The team will use snowball-sampling techniques to replace non-responders as necessary. Because this is a non-probability based sampling approach, items such as accuracy and reliability are not relevant. RAND anticipates being able to contact and review 80 percent of the desired respondents. The first 2-3 interviews will serve as a pretest of contact procedures and interview questions.

If a subject matter expert replies to our email invitation by declining the request to be interviewed, we will follow-up with another email asking for names and contact information of individuals who would be likely replacement subject matter experts. We will then extend the email invitation to those named individuals. This procedure will continue until the desired number of qualitative interviews has been completed. For the online Delphi panel, the selected 8 individuals will be called by the principal investigators and asked to participate.

4. Test of Procedures or Methods to be Undertaken

Because this data collection plan is comprised of a small number of qualitative interviews and an online Delphi panel as indicated in (Attachments 5-13), a pilot test is not proposed.

5. Individuals Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data

The person who designed the data collection is Dr. Johanna Zmud, RAND Corporation. The persons who will collect the data are Dr. Andres Villaveces and Liisa Ecola, RAND Corporation. The person who will analyze the data is Kathryn Connor, RAND Corporation.