Additiional Information on State Surveys

OMB No. 2127-0646

Supplemental Information Required For Approval of State/Demonstration Surveys

Background

On June 1, 2006, OMB approved NHTSA’s information collection request entitled “Evaluation Surveys for Impaired Driving and Safety Belt Interventions” using the above approval number and setting an expiration date of 6/30/2008. Under the terms of clearance, OMB gave Generic Clearance to State Surveys and Demonstration Surveys subsumed under 2127-0646 and wrote that NHTSA was to obtain approval of each specific State and Demonstration Survey. The terms of clearance further stated that OMB will review the complete specification of each State/Demonstration survey instance within 15 business days. NHTSA was instructed to provide OMB the following information for that review: (1) the exact questionnaire to be used; (2) designated survey population and sampling design; (3) whether or not the sample will be stratified, and if so, how; (4) specific sample sizes; (5) the scheduling of each survey instance; (6) the goals of each survey instance; and (7) how the data will be analyzed.

This document formally requests OMB approval for conducting State information collections subsumed under 2127-0646 to support the project entitled “Click It or Ticket: Next Generation,” which will work with Iowa, Pennsylvania, and Virginia to conduct multiple high visibility enforcement campaigns annually.

The required information follows:

1. Questionnaire

The questionnaires are included as Attachment A; one for each of the three project Sites. They will average 10 minutes in duration. The only difference between the questionnaires is that Q9 and Q31 contain the name of the different States (Iowa, Pennsylvania, and Virginia)

2. Survey Population and Sampling Design

NHTSA will survey the general licensed driver population age 18 and older by media markets in three States. (1) In Iowa, the media markets are Cedar Rapids, Davenport, Des Moines, and Sioux City; (2) in Pennsylvania, the media markets are Philadelphia, Allegheny County (Pittsburgh), Harrisburg, and Erie & Johnstown/Altoona; and (3) in Virginia, the media markets are Roanoke, Norfolk, Richmond, and Washington D.C. (Northern VA).

Beginning with the May 2007 National Click It or Ticket Mobilization, Iowa, Pennsylvania, and Virginia will participate in additional quarterly seat belt enforcement interventions in select media markets: Iowa will implement two additional seat belt interventions in the Cedar Rapids, Des Moines, and Sioux City media markets during July and November 2007. Pennsylvania will implement two additional seat belt enforcement interventions in Philadelphia, Allegheny County, and Harrisburg during September and November 2007. Virginia will implement three additional seat belt interventions in Roanoke, Richmond, and Norfolk media markets during July and November 2007 and March 2008.

NHTSA will purchase media with a safety belt message at different levels of intensity within select media markets in Pennsylvania, Iowa and Virginia. Then, NHTSA will conduct pre and post survey waves in the media markets where the seat belt media and enforcement interventions occurred, as well as in a control media market for each State. The result will be six survey waves per media market in Iowa and Pennsylvania, and eight survey waves per media market in Virginia.

Data collection methods will be the same across survey waves and across Sites. A probability sample of households from each media market using list-assisted RDD methodology will be employed. The sample will be purchased from the Marketing Systems Group, Inc. (MSG), one of the premier sample generation companies in the nation. The initial sample released will be proportional to the intended number of completed interviews. For example, in order to obtain a sample size of 250 completed surveys, we would select an initial release of 2000 telephone numbers. Based on about half that number being rejected for being business numbers, disconnects, or fax/computer tones, that would project a completion rate of about 25%. The sample purchased will be divided into replicates of 100 numbers to facilitate management of the sample to achieve an acceptable response rate. Additional replicates, if any, will be added to the sample as it becomes clear that the initial sample will not provide enough completed interviews. After the initial sample, any subsequent replicates will be added as soon as the need is recognized, so that there is maximum time for pursuing the numbers in the sample for consistent use of all target numbers and maximum completion rates.

The contractor will utilize MSG’s GENESYS Sampling System for the RDD sample. To generate sample, the GENESYS System employs a random digit dialing list-assisted methodology. List-assisted refers to the use of commercial lists of directory-listed telephone numbers to increase the likelihood of dialing household residences. The system utilizes a database consisting of all residential telephone exchanges, working bank information, and various geographic service parameters such as State, county, primary ZIP code, etc. In addition, the database provides working bank information at the two-digit level – each of the 100 banks (i.e., first two digits of the four-digit suffix) in each exchange will be defined as “working” if it contains one or more listed telephone households.

Within each State, there are four media markets that will be surveyed – three that are the targets of specialized enforcement campaigns and a control area. Each will receive an equal fraction of the telephone surveys within the state. Within each media market, telephone numbers will be distributed between the counties in the market according to county population.

Initial calls will be made between 9 a.m. and 9 p.m., with approximately 80% after 5 p.m. and on weekends to maximize the likelihood of reaching a suitable respondent. Within a household, in order to obtain a random selection of drivers, we will ask for the licensed driver age 18 or older whose birthday is closest to the call date. If the target is not available, we will try to schedule a specific callback. Three to five calls will be made to each number to make contact, and up to 10 calls total in order to reach a suitable target.

Surveys will be conducted utilizing Computer Assisted Telephone Interviewing (CATI) stations manned by interviewers.

3. Sample Stratification

The samples will not be stratified.

4. Sample Sizes

The total sample size for this study is 16,000 interviews, which includes 3 waves of pre and post surveys in 8 media markets in Iowa and Pennsylvania and 4 waves of pre and post surveys in 4 media markets in Virginia. For the first and last pre and post survey waves, the sample in each media market in each State will be composed of 250 respondents, while the intermittent pre and post survey wavers (2 each in PA and IA , and 4 in VA) will be 125.. Assuming a simple random sample, within-State pre-post differences for each intervention will need to exceed 7.1 percent to be statistically significant (pre-post, 3 Sites, 125 observations per Site) or 5.0 percent with 250 observations per Site. Sensitivity will be greater for evaluations across waves or across States. That is an adequate level of precision for this study given past evaluation experience with safety belt mobilizations.

Iowa		Intervention			Control
		Cedar Rapids	Des Moines	Sioux City	Davenport/ Omaha
Wave 1: CIOT, May 2007	Pre	250	250	250	250
	Intvn?	CIOT	CIOT	CIOT	CIOT
	Post	250	250	250	250
Wave 2 June-July 2007	Pre	125	125	125	125
	Intvn?	yes	yes	yes	no
	Post	125	125	125	125
Wave 3 Nov 2007	Pre	250	250	250	250
	Intvn?	yes	yes	yes	no
	Post	250	250	250	250

Pennsylvania		Intervention			Control
		Philadelphia	Allegheny County	Harrisburg	Erie/ Johnstown- Altoona
Wave 1: CIOT, May 2007	Pre	250	250	250	250
	Intvn?	CIOT	CIOT	CIOT	CIOT
	Post	250	250	250	250
Wave 2 Sept 2007	Pre	125	125	125	125
	Intvn?	yes	yes	yes	no
	Post	125	125	125	125
Wave 3 Nov 2007	Pre	250	250	250	250
	Intvn?	yes	yes	yes	no
	Post	250	250	250	250

Virginia		Intervention			Control
		Norfolk	Richmond	Roanoke	Northern VA (DC area)
Wave 1: CIOT, May 2007	Pre	250	250	250	250
	Intvn?	CIOT	CIOT	CIOT	CIOT
	Post	250	250	250	250
Wave 2 July 2007	Pre	125	125	125	125
	Intvn?	yes	yes	yes	no
	Post	125	125	125	125
Wave 3 Nov 2007	Pre	125	125	125	125
	Intvn?	yes	yes	yes	no
	Post	125	125	125	125
Wave 4 Mar 2008	Pre	250	250	250	250
	Intvn?	yes	yes	yes	no
	Post	250	250	250	250

5. Scheduling of Each Survey Instance

At each Site, the survey waves will immediately surround intervention activity occurring during periodically during 2007 and early 2008; the exact schedules vary from State to State. The table below outlines the schedule for the pre and post surveys as well as the media and enforcement for each intervention. For the May 2007 Click It or Ticket Mobilization, the enforcement intervention will consist of one week of media followed by two weeks of combined media and enforcement activity. The other enforcement waves are shorter in duration and differ in the length of media and enforcement. The pre-intervention survey wave will be conducted during the two weeks prior to onset of media. The post-intervention survey wave will start at the conclusion of the intervention and will proceed for two weeks.

		Iowa	Pennsylvania	Virginia
Wave 1	Pre-Survey	Apr 30 – May 13, 2007	April 30 – May 13, 2007	April 30 – May 13, 2007
	Media	May 14-28, 2007	May 14-28, 2007	May 14-28, 2007
	Enforcement	May 21-31, 2007	May 14 – June 3, 2007	May 13-June 3, 2007
	Post-Survey	June 1-14, 2007	June 4 – 18, 2007	June 4 – 18, 2007
Wave2	Pre-Survey	SKIP	Sept 5-18, 2007	SKIP
	Media	June 28- July 4, 2007	Sept 19-26, 2007	July 22-29, 2007
	Enforcement	July 2 – 6, 2007	Sept 24 - 30, 2007	July 30 – Aug 4, 2007
	Post-Survey	July 7-21, 2007	Oct 1-14, 2007	Aug 5-19, 2007
Wave 3	Pre-Survey	Nov 1-14, 2007	Oct 31-Nov 13, 2007	Oct 27- Nov 10, 2007
	Media	Nov 15 – 21, 2007	Nov 14-21, 2007	Nov 11-17, 2007
	Enforcement	Nov 19 – 25, 2007	Nov 19 – 25, 2007	Nov 18-25, 2007
	Post-Survey	Nov 26- Dec 10, 2007	Nov 26- Dec 10, 2007	Nov 26- Dec 10, 2007

		Iowa	Pennsylvania	Virginia
Wave 4	Pre-Survey			Feb 24- Mar 10, 2008
	Media			Mar 11-17, 2008
	Enforcement			Mar 18-24, 2008
	Post-Survey			Mar 5- Mar 19, 2008

6. Goals of Each Survey Instance

The goal of this Project is to evaluate the effectiveness of different levels of media and enforcement and any cumulative impact over time. This information collection is one part of the evaluation plan, which will also include observation surveys of belt use, brief written surveys conducted by the States in DMV offices, and process information provided by the Sites (e.g., media activity, citation data). The telephone survey is designed to assess the penetration of the communications portion of the intervention as well as assess public perceptions of the intervention and any changes in relevant perceptions and attitudes associated with the intervention. Each individual intervention will be assessed using a pre/post survey design.

7. How Data Will Be Analyzed

For each intervention wave, pre- and post-survey data will be compared to determine if:

• Awareness of project-related messaging increased from pre- to post-survey waves;

• Observation of enforcement activity increased from pre- to post-survey waves;

• Attitudes relating to enforcement of safety belt laws changed from pre- to post-survey waves.

For each intervention wave at each Site, pre/post responses to survey questions will be compared and the difference tested for statistical significance. Sites will be compared in how often they experienced statistically significant change and the magnitude of those changes. There will also be within-Site comparison of pre/post differences across survey waves for selected questionnaire items in order to assess the impact of Site intervention adjustments. There will also be evaluations of overall changes across waves to assess cumulative benefits of the repeated treatments.