Attachment_J.6_Summary of Study Design and Analyses.pdf

RFP: 693JJ924R000021

Title: Older Novice Driver Naturalistic Driving Study

Attachment J.6

Summary of Study Design and Analyses

Materials related to the study design and analyses were developed under a prior project. The full set of materials developed under the prior study will be shared with the Contractor at award.

However, many details about the study design and analyses are available in the prior project’s Final Report (https://rosap.ntl.bts.gov/view/dot/67045). Additionally, a summary of the study design and analyses is below.

The study developed under the prior project (https://rosap.ntl.bts.gov/view/dot/67045) has the following characteristics:

• Naturalistic driving data will be collected for a period of twelve (12) months from 500 novices ages 15.5 – 16.5 (depending on the minimum licensing age of the State in which the study is conducted) and 500 novices ages 18 – 20.

• All participants will have naturalistic driving data collected using a smartphone-based DAS (smartphone app) capable of collecting data necessary to answer the research questions in Section C.2. A subset of 84 younger and 84 older novices (168 total) will also have naturalistic driving data collected using a vehicle-based DAS (i.e., with camera, accelerometer, gyroscope, GPS, etc.).

• Naturalistic driving data collection will begin when participants receive the first license that allows them to drive unsupervised.

• Potential participants will be quota sampled to ensure that the study sample aligns with what is known about the sociodemographic characteristics of novice drivers in the U.S.

(e.g., 12). Currently, the sampling plan includes considering sex, race and ethnicity, and socioeconomic characteristics (e.g., family affluence) when recruiting novice drivers.

• The study will be conducted in a State that does not currently apply GDL provisions to novices 18 and older; the State should also have a minimum age for unsupervised driving between 15.5 and 16.5 (i.e., the minimum age in the majority of States).

• To recruit participants for the study, the research team will need to partner with a State licensing agency. The State licensing agency should have sufficient numbers of younger and older novices applying for licensure to reach recruitment targets.

• Potential participants will be identified using information available in the State licensing agency records (e.g., age, sex, race and ethnicity, licensure status). Potential participants will additionally be administered screening questions to further determine eligibility (e.g., ownership of an Android or iOS smartphone, information about vehicle access necessary to determine eligibility for the in-vehicle DAS subgroup) at the time they schedule their on-road/behind-the-wheel licensing exams.

• Based in part on the experience of Thomas et al., 2016,1 the current study plan estimates that a high-volume State licensing agency could yield 3 to 18 participants (including 1 to 6 older novices) per day. Assuming this rate, if 3 individual licensing agency offices participated in the study, it would take a maximum of 167 office days (approximately 8

1 Thomas, F. D., Rilea, S. L., Blomberg, R. D., Peck. R. C., & Korbelak, K. T. (2016, January). Evaluation of the safety benefits of the risk awareness and perception training program for novice teen drivers (Report No. DOT HS 812 235). National Highway Traffic Safety Administration https://rosap.ntl.bts.gov/view/dot/67045

RFP: 693JJ924R000021

Title: Older Novice Driver Naturalistic Driving Study

Attachment J.6 mos.) to complete recruitment. With 6 offices participating, it would take a maximum of 84 office days (approximately 4 mos.) to complete recruitment.

• The current study plan estimates compensation for novices in the “smartphone only” data collection group to be $475 by the end of the study. The plan estimates the “smartphone plus in-vehicle DAS” data collection group to be compensated $675 by the end of the study.

• Participants will complete three web-based, self-administered questionnaires at the beginning, middle, and end of the study that do not take more than 30 minutes to complete. The baseline questionnaire will contain questions about sociodemographic information and other information that prior studies suggest will predict driving performance and/or exposure, including but not limited to: measures of personality, sensation-seeking, attitudes about traffic safety, self-reported driving behavior, parental involvement, experience with driving and driver education, and vehicle access. Follow-up questionnaires will be shorter and only include measures that may change over time and/or with driving experience, e.g., attitudes about traffic safety, vehicle access.

• Naturalistic driving and questionnaire responses will be used to derive variables necessary to answer the research questions in Section C.2, including outcome variables like crashes or near-crashes and kinematic risky driving (i.e., elevated g-force) events.

Other variables will include information about the amount and type of exposure (e.g., road type, trip duration, trip time of day, weather, posted speed limit, route familiarity, complexity of driving environment), information about risky behaviors (e.g., speeding, cell phone use), and information from the questionnaires.

• The current plan includes a variety of statistical techniques for data analysis, including approaches that discover groups of participants with similar trajectories of risky driving over the study period (e.g., group-based trajectory models, latent class growth analysis using growth mixture modeling), as well as approaches that allow group-based comparisons between younger and older novices, or higher- and lower-risk novices (e.g., mixed-effects generalized linear models.

• Potential challenges for conducting the study include: finding a State licensing agency willing to participate and maintain its commitment during the approval process, recruiting a sample representative of the younger (15.5 – 16.5) and older (18 – 20) novice driver populations; recruiting the desired number of participants within the specified amount of time; and obtaining accurate measures of performance and exposure from the smartphone app (e.g., distinguishing travel modes; distinguishing passenger from driver).